Medical Image Analysis

Journal Name	Impact Factor	CiteScore	Launched Year	First Decision (median)	APC
Applied Sciences applsci	2.5	5.3	2011	17.8 Days	CHF 2400
Journal of Imaging jimaging	2.7	5.9	2015	20.9 Days	CHF 1800
Electronics electronics	2.6	5.3	2012	16.8 Days	CHF 2400
Diagnostics diagnostics	3.0	4.7	2011	20.5 Days	CHF 2600
Biomedicines biomedicines	3.9	5.2	2013	15.3 Days	CHF 2600

23 pages, 6559 KiB

Open AccessArticle

Predicting the Tumour Response to Radiation by Modelling the Five Rs of Radiotherapy Using PET Images

by Rihab Hami, Sena Apeke, Pascal Redou, Laurent Gaubert, Ludwig J. Dubois, Philippe Lambin, Dimitris Visvikis and Nicolas Boussion

J. Imaging 2023, 9(6), 124; https://doi.org/10.3390/jimaging9060124 - 20 Jun 2023

Cited by 1 | Viewed by 2748

Abstract

Despite the intensive use of radiotherapy in clinical practice, its effectiveness depends on several factors. Several studies showed that the tumour response to radiation differs from one patient to another. The non-uniform response of the tumour is mainly caused by multiple interactions between [...] Read more.

Despite the intensive use of radiotherapy in clinical practice, its effectiveness depends on several factors. Several studies showed that the tumour response to radiation differs from one patient to another. The non-uniform response of the tumour is mainly caused by multiple interactions between the tumour microenvironment and healthy cells. To understand these interactions, five major biologic concepts called the “5 Rs” have emerged. These concepts include reoxygenation, DNA damage repair, cell cycle redistribution, cellular radiosensitivity and cellular repopulation. In this study, we used a multi-scale model, which included the five Rs of radiotherapy, to predict the effects of radiation on tumour growth. In this model, the oxygen level was varied in both time and space. When radiotherapy was given, the sensitivity of cells depending on their location in the cell cycle was taken in account. This model also considered the repair of cells by giving a different probability of survival after radiation for tumour and normal cells. Here, we developed four fractionation protocol schemes. We used simulated and positron emission tomography (PET) imaging with the hypoxia tracer 18F-flortanidazole (18F-HX4) images as input data of our model. In addition, tumour control probability curves were simulated. The result showed the evolution of tumours and normal cells. The increase in the cell number after radiation was seen in both normal and malignant cells, which proves that repopulation was included in this model. The proposed model predicts the tumour response to radiation and forms the basis for a more patient-specific clinical tool where related biological data will be included. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

12 pages, 797 KiB

Open AccessArticle

A Convolutional Neural Network-Based Connectivity Enhancement Approach for Autism Spectrum Disorder Detection

by Fatima Zahra Benabdallah, Ahmed Drissi El Maliani, Dounia Lotfi and Mohammed El Hassouni

J. Imaging 2023, 9(6), 110; https://doi.org/10.3390/jimaging9060110 - 31 May 2023

Cited by 5 | Viewed by 2208

Abstract

Autism spectrum disorder (ASD) represents an ongoing obstacle facing many researchers to achieving early diagnosis with high accuracy. To advance developments in ASD detection, the corroboration of findings presented in the existing body of autism-based literature is of high importance. Previous works put [...] Read more.

Autism spectrum disorder (ASD) represents an ongoing obstacle facing many researchers to achieving early diagnosis with high accuracy. To advance developments in ASD detection, the corroboration of findings presented in the existing body of autism-based literature is of high importance. Previous works put forward theories of under- and over-connectivity deficits in the autistic brain. An elimination approach based on methods that are theoretically comparable to the aforementioned theories proved the existence of these deficits. Therefore, in this paper, we propose a framework that takes into account the properties of under- and over-connectivity in the autistic brain using an enhancement approach coupled with deep learning through convolutional neural networks (CNN). In this approach, image-alike connectivity matrices are created, and then connections related to connectivity alterations are enhanced. The overall objective is the facilitation of early diagnosis of this disorder. After conducting tests using information from the large multi-site Autism Brain Imaging Data Exchange (ABIDE I) dataset, the results show that this approach provides an accurate prediction value reaching up to 96%. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

9 pages, 1317 KiB

Open AccessArticle

Tomosynthesis-Detected Architectural Distortions: Correlations between Imaging Characteristics and Histopathologic Outcomes

by Giovanna Romanucci, Francesca Fornasa, Andrea Caneva, Claudia Rossati, Marta Mandarà, Oscar Tommasini and Rossella Rella

J. Imaging 2023, 9(5), 103; https://doi.org/10.3390/jimaging9050103 - 19 May 2023

Cited by 3 | Viewed by 2235

Abstract

Objective: to determine the positive predictive value (PPV) of tomosynthesis (DBT)-detected architectural distortions (ADs) and evaluate correlations between AD’s imaging characteristics and histopathologic outcomes. Methods: biopsies performed between 2019 and 2021 on ADs were included. Images were interpreted by dedicated breast imaging radiologists. [...] Read more.

Objective: to determine the positive predictive value (PPV) of tomosynthesis (DBT)-detected architectural distortions (ADs) and evaluate correlations between AD’s imaging characteristics and histopathologic outcomes. Methods: biopsies performed between 2019 and 2021 on ADs were included. Images were interpreted by dedicated breast imaging radiologists. Pathologic results after DBT-vacuum assisted biopsy (DBT-VAB) and core needle biopsy were compared with AD detected by DBT, synthetic2D (synt2D) and ultrasound (US). Results: US was performed to assess a correlation for ADs in all 123 cases and a US correlation was identified in 12/123 (9.7%) cases, which underwent US-guided core needle biopsy (CNB). The remaining 111/123 (90.2%) ADs were biopsied under DBT guidance. Among the 123 ADs included, 33/123 (26.8%) yielded malignant results. The overall PPV for malignancy was 30.1% (37/123). The imaging-specific PPV for malignancy was 19.2% (5/26) for DBT-only ADs, 28.2% (24/85) for ADs visible on DBT and synth2D mammography and 66.7% (8/12) for ADs with a US correlation with a statistically significant difference among the three groups (p = 0.01). Conclusions: DBT-only ADs demonstrated a lower PPV of malignancy when compared with syntD mammography, and DBT detected ADs but not low enough to avoid biopsy. As the presence of a US correlate was found to be related with malignancy, it should increase the radiologist’s level of suspicion, even when CNB returned a B3 result. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

15 pages, 11972 KiB

Open AccessArticle

Functional Magnetic Resonance Urography in Children—Tips and Pitfalls

by Małgorzata Grzywińska, Dominik Świętoń, Agnieszka Sabisz and Maciej Piskunowicz

Diagnostics 2023, 13(10), 1786; https://doi.org/10.3390/diagnostics13101786 - 18 May 2023

Viewed by 1996

Abstract

MR urography can be an alternative to other imaging methods of the urinary tract in children. However, this examination may present technical problems influencing further results. Special attention must be paid to the parameters of dynamic sequences to obtain valuable data for further [...] Read more.

MR urography can be an alternative to other imaging methods of the urinary tract in children. However, this examination may present technical problems influencing further results. Special attention must be paid to the parameters of dynamic sequences to obtain valuable data for further functional analysis. The analysis of methodology for renal function assessment using 3T magnetic resonance in children. A retrospective analysis of MR urography studies was performed in a group of 91 patients. Particular attention was paid to the acquisition parameters of the 3D-Thrive dynamic with contrast medium administration as a basic urography sequence. The authors have evaluated images qualitatively and compared contrast-to-noise ratio (CNR), curves smoothness, and quality of baseline (evaluation signal noise ratio) in every dynamic in each patient in every protocol used in our institution. Quality analysis of the image (ICC = 0.877, p < 0.001) was improved so that we have a statistically significant difference in image quality between protocols (χ²(3) = 20.134, p < 0.001). The results obtained for SNR in the medulla and cortex show that there was a statistically significant difference in SNR in the cortex (χ²(3) = 9.060, p = 0.029). Therefore, the obtained results show that with the newer protocol, we obtain lower values of standard deviation for TTP in the aorta (in ChopfMRU: first protocol SD = 14.560 vs. fourth protocol SD = 5.599; in IntelliSpace Portal: first protocol SD = 15.241 vs. fourth protocol SD = 5.506). Magnetic resonance urography is a promising technique with a few challenges that arise and need to be overcome. New technical opportunities should be introduced for everyday practice to improve MRU results. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

15 pages, 37806 KiB

Open AccessReview

Mammography Datasets for Neural Networks—Survey

by Adam Mračko, Lucia Vanovčanová and Ivan Cimrák

J. Imaging 2023, 9(5), 95; https://doi.org/10.3390/jimaging9050095 - 10 May 2023

Cited by 2 | Viewed by 8476

Abstract

Deep neural networks have gained popularity in the field of mammography. Data play an integral role in training these models, as training algorithms requires a large amount of data to capture the general relationship between the model’s input and output. Open-access databases are [...] Read more.

Deep neural networks have gained popularity in the field of mammography. Data play an integral role in training these models, as training algorithms requires a large amount of data to capture the general relationship between the model’s input and output. Open-access databases are the most accessible source of mammography data for training neural networks. Our work focuses on conducting a comprehensive survey of mammography databases that contain images with defined abnormal areas of interest. The survey includes databases such as INbreast, the Curated Breast Imaging Subset of Digital Database for Screening Mammography (CBIS-DDSM), the OPTIMAM Medical Image Database (OMI-DB), and The Mammographic Image Analysis Society Digital Mammogram Database (MIAS). Additionally, we surveyed recent studies that have utilized these databases in conjunction with neural networks and the results they have achieved. From these databases, it is possible to obtain at least 3801 unique images with 4125 described findings from approximately 1842 patients. The number of patients with important findings can be increased to approximately 14,474, depending on the type of agreement with the OPTIMAM team. Furthermore, we provide a description of the annotation process for mammography images to enhance the understanding of the information gained from these datasets. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

13 pages, 6064 KiB

Open AccessArticle

Self-Paced Dual-Axis Attention Fusion Network for Retinal Vessel Segmentation

by Yueting Shi, Weijiang Wang, Minzhi Yuan and Xiaohua Wang

Electronics 2023, 12(9), 2107; https://doi.org/10.3390/electronics12092107 - 5 May 2023

Viewed by 1407

Abstract

The segmentation of retinal vessels plays an essential role in the early recognition of ophthalmic diseases in clinics. Increasingly, approaches based on deep learning have been pushing vessel segmentation performance, yet it is still a challenging problem due to the complex structure of [...] Read more.

The segmentation of retinal vessels plays an essential role in the early recognition of ophthalmic diseases in clinics. Increasingly, approaches based on deep learning have been pushing vessel segmentation performance, yet it is still a challenging problem due to the complex structure of retinal vessels and the lack of precisely labeled samples. In this paper, we propose a self-paced dual-axis attention fusion network (SPDAA-Net). Firstly, a self-paced learning mechanism using a query-by-committee algorithm is designed to guide the model to learn from easy to hard, which makes model training more intelligent. Secondly, during fusing of multi-scale features, a dual-axis attention mechanism composed of height and width attention is developed to perceive the object, which brings in long-range dependencies while reducing computation complexity. Furthermore, CutMix data augmentation is applied to increase the generalization of the model, enhance the recognition ability of global and local features, and ultimately boost accuracy. We implement comprehensive experiments validating that our SPDAA-Net obtains remarkable performance on both the public DRIVE and CHASE-DB1 datasets. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

15 pages, 1926 KiB

Open AccessArticle

A Channel Correction and Spatial Attention Framework for Anterior Cruciate Ligament Tear with Ordinal Loss

by Weilun Lin and Kehua Miao

Appl. Sci. 2023, 13(8), 5005; https://doi.org/10.3390/app13085005 - 16 Apr 2023

Cited by 2 | Viewed by 2537

Abstract

The anterior cruciate ligament (ACL) is critical for controlling the motion of the knee joint, but it is prone to injury during sports activities and physical work. If left untreated, ACL injuries can lead to various pathologies such as meniscal damage and osteoarthritis. [...] Read more.

The anterior cruciate ligament (ACL) is critical for controlling the motion of the knee joint, but it is prone to injury during sports activities and physical work. If left untreated, ACL injuries can lead to various pathologies such as meniscal damage and osteoarthritis. While previous studies have used deep learning to diagnose ACL tears, there has been a lack of standardization in human unit classification, leading to mismatches between their findings and actual clinical diagnoses. To address this, we perform a triple classification task based on various tear classes using an ordinal loss on the KneeMRI dataset. We utilize a channel correction module to address image distribution issues across multiple patients, along with a spatial attention module, and test its effectiveness with various backbone networks. Our results show that the modules are effective on various backbone networks, achieving an accuracy of 83.3% on ResNet-18, a 6.65% improvement compared to the baseline. Additionally, we carry out an ablation experiment to verify the effectiveness of the three modules and present our findings with figures and tables. Overall, our study demonstrates the potential of deep learning in diagnosing ACL tear and provides insights into improving the accuracy and standardization of such diagnoses. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

28 pages, 2684 KiB

Open AccessReview

Deep Learning Approaches for Data Augmentation in Medical Imaging: A Review

by Aghiles Kebaili, Jérôme Lapuyade-Lahorgue and Su Ruan

J. Imaging 2023, 9(4), 81; https://doi.org/10.3390/jimaging9040081 - 13 Apr 2023

Cited by 66 | Viewed by 17939

Abstract

Deep learning has become a popular tool for medical image analysis, but the limited availability of training data remains a major challenge, particularly in the medical field where data acquisition can be costly and subject to privacy regulations. Data augmentation techniques offer a [...] Read more.

Deep learning has become a popular tool for medical image analysis, but the limited availability of training data remains a major challenge, particularly in the medical field where data acquisition can be costly and subject to privacy regulations. Data augmentation techniques offer a solution by artificially increasing the number of training samples, but these techniques often produce limited and unconvincing results. To address this issue, a growing number of studies have proposed the use of deep generative models to generate more realistic and diverse data that conform to the true distribution of the data. In this review, we focus on three types of deep generative models for medical image augmentation: variational autoencoders, generative adversarial networks, and diffusion models. We provide an overview of the current state of the art in each of these models and discuss their potential for use in different downstream tasks in medical imaging, including classification, segmentation, and cross-modal translation. We also evaluate the strengths and limitations of each model and suggest directions for future research in this field. Our goal is to provide a comprehensive review about the use of deep generative models for medical image augmentation and to highlight the potential of these models for improving the performance of deep learning algorithms in medical image analysis. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

21 pages, 1642 KiB

Open AccessArticle

On The Potential of Image Moments for Medical Diagnosis

by Cecilia Di Ruberto, Andrea Loddo and Lorenzo Putzu

J. Imaging 2023, 9(3), 70; https://doi.org/10.3390/jimaging9030070 - 17 Mar 2023

Cited by 2 | Viewed by 1992

Abstract

Medical imaging is widely used for diagnosis and postoperative or post-therapy monitoring. The ever-increasing number of images produced has encouraged the introduction of automated methods to assist doctors or pathologists. In recent years, especially after the advent of convolutional neural networks, many researchers [...] Read more.

Medical imaging is widely used for diagnosis and postoperative or post-therapy monitoring. The ever-increasing number of images produced has encouraged the introduction of automated methods to assist doctors or pathologists. In recent years, especially after the advent of convolutional neural networks, many researchers have focused on this approach, considering it to be the only method for diagnosis since it can perform a direct classification of images. However, many diagnostic systems still rely on handcrafted features to improve interpretability and limit resource consumption. In this work, we focused our efforts on orthogonal moments, first by providing an overview and taxonomy of their macrocategories and then by analysing their classification performance on very different medical tasks represented by four public benchmark data sets. The results confirmed that convolutional neural networks achieved excellent performance on all tasks. Despite being composed of much fewer features than those extracted by the networks, orthogonal moments proved to be competitive with them, showing comparable and, in some cases, better performance. In addition, Cartesian and harmonic categories provided a very low standard deviation, proving their robustness in medical diagnostic tasks. We strongly believe that the integration of the studied orthogonal moments can lead to more robust and reliable diagnostic systems, considering the performance obtained and the low variation of the results. Finally, since they have been shown to be effective on both magnetic resonance and computed tomography images, they can be easily extended to other imaging techniques. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

14 pages, 6932 KiB

Open AccessArticle

Using Mean Arterial Pressure in Hypertension Diagnosis versus Using Either Systolic or Diastolic Blood Pressure Measurements

by Heba Kandil, Ahmed Soliman, Norah Saleh Alghamdi, J. Richard Jennings and Ayman El-Baz

Biomedicines 2023, 11(3), 849; https://doi.org/10.3390/biomedicines11030849 - 10 Mar 2023

Cited by 12 | Viewed by 5041

Abstract

Hypertension is a severe and highly prevalent disease. It is considered a leading contributor to mortality worldwide. Diagnosis guidelines for hypertension use systolic and diastolic blood pressure (BP) together. Mean arterial pressure (MAP), which refers to the average of the arterial blood pressure [...] Read more.

Hypertension is a severe and highly prevalent disease. It is considered a leading contributor to mortality worldwide. Diagnosis guidelines for hypertension use systolic and diastolic blood pressure (BP) together. Mean arterial pressure (MAP), which refers to the average of the arterial blood pressure through a single cardiac cycle, can be an alternative index that may capture the overall exposure of the person to a heightened pressure. A clinical hypothesis, however, suggests that in patients over 50 years old in age, systolic BP may be more predictive of adverse events, while in patients under 50 years old, diastolic BP may be slightly more predictive. In this study, we investigated the correlation between cerebrovascular changes, (impacted by hypertension), and MAP, systolic BP, and diastolic BP separately. Several experiments were conducted using real and synthetic magnetic resonance angiography (MRA) data, along with corresponding BP measurements. Each experiment employs the following methodology: First, MRA data were processed to remove noise, bias, or inhomogeneity. Second, the cerebrovasculature was delineated for MRA subjects using a 3D adaptive region growing connected components algorithm. Third, vascular features (changes in blood vessel’s diameters and tortuosity) that describe cerebrovascular alterations that occur prior to and during the development of hypertension were extracted. Finally, feature vectors were constructed, and data were classified using different classifiers, such as SVM, KNN, linear discriminant, and logistic regression, into either normotensives or hypertensives according to the cerebral vascular alterations and the BP measurements. The initial results showed that MAP would be more beneficial and accurate in identifying the cerebrovascular impact of hypertension (accuracy up to

95.2 %

) than just using either systolic BP (accuracy up to

89.3 %

) or diastolic BP (accuracy up to

88.9 %

). This result emphasizes the pathophysiological significance of MAP and supports prior views that this simple measure may be a superior index for the definition of hypertension and research on hypertension. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

21 pages, 50786 KiB

Open AccessArticle

Impact of Training Data, Ground Truth and Shape Variability in the Deep Learning-Based Semantic Segmentation of HeLa Cells Observed with Electron Microscopy

by Cefa Karabağ, Mauricio Alberto Ortega-Ruíz and Constantino Carlos Reyes-Aldasoro

J. Imaging 2023, 9(3), 59; https://doi.org/10.3390/jimaging9030059 - 1 Mar 2023

Cited by 5 | Viewed by 4130

Abstract

This paper investigates the impact of the amount of training data and the shape variability on the segmentation provided by the deep learning architecture U-Net. Further, the correctness of ground truth (GT) was also evaluated. The input data consisted of a three-dimensional set [...] Read more.

This paper investigates the impact of the amount of training data and the shape variability on the segmentation provided by the deep learning architecture U-Net. Further, the correctness of ground truth (GT) was also evaluated. The input data consisted of a three-dimensional set of images of HeLa cells observed with an electron microscope with dimensions

8192 \times 8192 \times 517

. From there, a smaller region of interest (ROI) of

2000 \times 2000 \times 300

was cropped and manually delineated to obtain the ground truth necessary for a quantitative evaluation. A qualitative evaluation was performed on the

8192 \times 8192

slices due to the lack of ground truth. Pairs of patches of data and labels for the classes nucleus, nuclear envelope, cell and background were generated to train U-Net architectures from scratch. Several training strategies were followed, and the results were compared against a traditional image processing algorithm. The correctness of GT, that is, the inclusion of one or more nuclei within the region of interest was also evaluated. The impact of the extent of training data was evaluated by comparing results from 36,000 pairs of data and label patches extracted from the odd slices in the central region, to 135,000 patches obtained from every other slice in the set. Then, 135,000 patches from several cells from the

8192 \times 8192

slices were generated automatically using the image processing algorithm. Finally, the two sets of 135,000 pairs were combined to train once more with 270,000 pairs. As would be expected, the accuracy and Jaccard similarity index improved as the number of pairs increased for the ROI. This was also observed qualitatively for the

8192 \times 8192

slices. When the

8192 \times 8192

slices were segmented with U-Nets trained with 135,000 pairs, the architecture trained with automatically generated pairs provided better results than the architecture trained with the pairs from the manually segmented ground truths. This suggests that the pairs that were extracted automatically from many cells provided a better representation of the four classes of the various cells in the

8192 \times 8192

slice than those pairs that were manually segmented from a single cell. Finally, the two sets of 135,000 pairs were combined, and the U-Net trained with these provided the best results. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

12 pages, 3425 KiB

Open AccessBrief Report

Prostate Ultrasound Image Segmentation Based on DSU-Net

by Xinyu Wang, Zhengqi Chang, Qingfang Zhang, Cheng Li, Fei Miao and Gang Gao

Biomedicines 2023, 11(3), 646; https://doi.org/10.3390/biomedicines11030646 - 21 Feb 2023

Cited by 8 | Viewed by 2725

Abstract

In recent years, the incidence of prostate cancer in the male population has been increasing year by year. Transrectal ultrasound (TRUS) is an important means of prostate cancer diagnosis. The accurate segmentation of the prostate in TRUS images can assist doctors in needle [...] Read more.

In recent years, the incidence of prostate cancer in the male population has been increasing year by year. Transrectal ultrasound (TRUS) is an important means of prostate cancer diagnosis. The accurate segmentation of the prostate in TRUS images can assist doctors in needle biopsy and surgery and is also the basis for the accurate identification of prostate cancer. Due to the asymmetric shape and blurred boundary line of the prostate in TRUS images, it is difficult to obtain accurate segmentation results with existing segmentation methods. Therefore, a prostate segmentation method called DSU-Net is proposed in this paper. This proposed method replaces the basic convolution in the U-Net model with the improved convolution combining shear transformation and deformable convolution, making the network more sensitive to border features and more suitable for prostate segmentation tasks. Experiments show that DSU-Net has higher accuracy than other existing traditional segmentation methods. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

12 pages, 1029 KiB

Open AccessArticle

Palatine Tonsil Measurements and Echogenicity during Tonsillitis Using Ultrasonography: A Case–Control Study

by Zohida A. Abdelgabar, Mahasin G. Hassan, Tasneem S. A. Elmahdi, Shanoo Sheikh and Wireen Leila T. Dator

Diagnostics 2023, 13(4), 742; https://doi.org/10.3390/diagnostics13040742 - 15 Feb 2023

Cited by 3 | Viewed by 3015

Abstract

This case–control study aimed to assess the size and echogenicity of inflamed tonsils using ultrasonography. It was carried out at different hospitals, nurseries, and primary schools in Khartoum state. About 131 Sudanese volunteers between 1 and 24 years old were recruited. The sample [...] Read more.

This case–control study aimed to assess the size and echogenicity of inflamed tonsils using ultrasonography. It was carried out at different hospitals, nurseries, and primary schools in Khartoum state. About 131 Sudanese volunteers between 1 and 24 years old were recruited. The sample included 79 volunteers with normal tonsils and 52 with tonsillitis according to hematological investigations. The sample was divided into groups according to age—1–5 years old, 6–10 years old, and more than ten years. Measurements in centimeters of height (AP) and width (transverse) of both tonsils (right and left) were taken. Echogenicity was assessed according to normal and abnormal appearances. A data collection sheet containing all the study variables was used. The independent samples test (t-test) showed an insignificant height difference between normal controls and cases with tonsillitis. The transverse diameter increased significantly with inflammation (p-value < 0.05) for both tonsils in all groups. Echogenicity can differentiate between normal and abnormal tonsils (p-value < 0.05 using the chi-square test) for samples from 1–5 years and 6–10 years. The study concluded that measurements and appearance are reliable indicators of tonsillitis, which can be confirmed with the use of ultrasonography, helping physicians to make the correct diagnosis and decisions. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

17 pages, 4584 KiB

Open AccessArticle

Reverse-Net: Few-Shot Learning with Reverse Teaching for Deformable Medical Image Registration

by Xin Zhang, Tiejun Yang, Xiang Zhao and Aolin Yang

Appl. Sci. 2023, 13(2), 1040; https://doi.org/10.3390/app13021040 - 12 Jan 2023

Cited by 1 | Viewed by 2303

Abstract

Multimodal medical image registration has an important role in monitoring tumor growth, radiotherapy, and disease diagnosis. Deep-learning-based methods have made great progress in the past few years. However, its success depends on large training datasets, and the performance of the model decreases due [...] Read more.

Multimodal medical image registration has an important role in monitoring tumor growth, radiotherapy, and disease diagnosis. Deep-learning-based methods have made great progress in the past few years. However, its success depends on large training datasets, and the performance of the model decreases due to overfitting and poor generalization when only limited data are available. In this paper, a multimodal medical image registration framework based on few-shot learning is proposed, named reverse-net, which can improve the accuracy and generalization ability of the network by using a few segmentation labels. Firstly, we used the border enhancement network to enhance the ROI (region of interest) boundaries of T1 images to provide high-quality data for the subsequent pixel alignment stage. Secondly, through a coarse registration network, the T1 image and T2 image were roughly aligned. Then, the pixel alignment network generated more smooth deformation fields. Finally, the reverse teaching network used the warped T1 segmentation labels and warped images generated by the deformation field to teach the border enhancement network more structural knowledge. The performance and generalizability of our model have been evaluated on publicly available brain datasets including the MRBrainS13DataNii-Pro, SRI24, CIT168, and OASIS datasets. Compared with VoxelMorph, the reverse-net obtained performance improvements of 4.36% in DSC on the publicly available MRBrainS13DataNii-Pro dataset. On the unseen dataset OASIS, the reverse-net obtained performance improvements of 4.2% in DSC compared with VoxelMorph, which shows that the model can obtain better generalizability. The promising performance on dataset CIT168 indicates that the model is practicable. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

21 pages, 3129 KiB

Open AccessArticle

The Correlation between the Vascular Calcification Score of the Coronary Artery and the Abdominal Aorta in Patients with Psoriasis

by Trang Nguyen-Mai Huynh, Fumikazu Yamazaki, Izumi Kishimoto, Akihiro Tanaka, Yonsu Son, Yoshio Ozaki, Kazuya Takehana and Hideaki Tanizaki

Diagnostics 2023, 13(2), 274; https://doi.org/10.3390/diagnostics13020274 - 11 Jan 2023

Cited by 1 | Viewed by 2652

Abstract

Psoriasis is known as an independent risk factor for cardiovascular disease due to its chronic inflammation. Studies have been conducted to evaluate the progress of atherosclerotic plaques in psoriasis. However, inadequate efforts have been made to clarify the relationship between atherosclerosis progress in [...] Read more.

Psoriasis is known as an independent risk factor for cardiovascular disease due to its chronic inflammation. Studies have been conducted to evaluate the progress of atherosclerotic plaques in psoriasis. However, inadequate efforts have been made to clarify the relationship between atherosclerosis progress in coronary arteries and other important blood vessels. For that reason, we investigated the correlation and development of the coronary artery calcification score (CACS) and the abdominal aortic calcification score (AACS) during a follow-up examination. Eighty-three patients with psoriasis underwent coronary computed tomography angiography (CCTA) for total CACS and abdominal computed tomography (AbCT) for total AACS. PASI score, other clinical features, and blood samples were collected at the same time. The patients’ medical histories were also retrieved for further analysis. Linear regression was used to analyze the CACS and AACS associations. There was a moderate correlation between CACS and AACS, while both calcification scores relatively reflected the coronary plaque number, coronary stenosis number, and stenosis severity observed with CCTA. Both calcification scores were independent of the PASI score. However, a significantly higher CACS was found in psoriatic arthritis, whereas no similar phenomenon was recorded for AACS. To conclude, both CACS and AACS might be potential alternative tests to predict the presence of coronary lesions as confirmed by CCTA. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

22 pages, 5342 KiB

Open AccessArticle

A Feature Extraction Using Probabilistic Neural Network and BTFSC-Net Model with Deep Learning for Brain Tumor Classification

by Arun Singh Yadav, Surendra Kumar, Girija Rani Karetla, Juan Carlos Cotrina-Aliaga, José Luis Arias-Gonzáles, Vinod Kumar, Satyajee Srivastava, Reena Gupta, Sufyan Ibrahim, Rahul Paul, Nithesh Naik, Babita Singla and Nisha S. Tatkar

J. Imaging 2023, 9(1), 10; https://doi.org/10.3390/jimaging9010010 - 31 Dec 2022

Cited by 19 | Viewed by 3920

Abstract

Background and Objectives: Brain Tumor Fusion-based Segments and Classification-Non-enhancing tumor (BTFSC-Net) is a hybrid system for classifying brain tumors that combine medical image fusion, segmentation, feature extraction, and classification procedures. Materials and Methods: to reduce noise from medical images, the hybrid probabilistic wiener [...] Read more.

Background and Objectives: Brain Tumor Fusion-based Segments and Classification-Non-enhancing tumor (BTFSC-Net) is a hybrid system for classifying brain tumors that combine medical image fusion, segmentation, feature extraction, and classification procedures. Materials and Methods: to reduce noise from medical images, the hybrid probabilistic wiener filter (HPWF) is first applied as a preprocessing step. Then, to combine robust edge analysis (REA) properties in magnetic resonance imaging (MRI) and computed tomography (CT) medical images, a fusion network based on deep learning convolutional neural networks (DLCNN) is developed. Here, the brain images’ slopes and borders are detected using REA. To separate the sick region from the color image, adaptive fuzzy c-means integrated k-means (HFCMIK) clustering is then implemented. To extract hybrid features from the fused image, low-level features based on the redundant discrete wavelet transform (RDWT), empirical color features, and texture characteristics based on the gray-level cooccurrence matrix (GLCM) are also used. Finally, to distinguish between benign and malignant tumors, a deep learning probabilistic neural network (DLPNN) is deployed. Results: according to the findings, the suggested BTFSC-Net model performed better than more traditional preprocessing, fusion, segmentation, and classification techniques. Additionally, 99.21% segmentation accuracy and 99.46% classification accuracy were reached using the proposed BTFSC-Net model. Conclusions: earlier approaches have not performed as well as our presented method for image fusion, segmentation, feature extraction, classification operations, and brain tumor classification. These results illustrate that the designed approach performed more effectively in terms of enhanced quantitative evaluation with better accuracy as well as visual performance. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

14 pages, 1685 KiB

Open AccessArticle

Cascaded Hough Transform-Based Hair Mask Generation and Harmonic Inpainting for Automated Hair Removal from Dermoscopy Images

by Amira S. Ashour, Basant S. Abd El-Wahab, Maram A. Wahba, Diaa-Eldin A. Mansour, Abeer Abd Elhakam Hodeib, Rasha Abd El-Ghany Khedr and Ghada F. R. Hassan

Diagnostics 2022, 12(12), 3040; https://doi.org/10.3390/diagnostics12123040 - 4 Dec 2022

Cited by 2 | Viewed by 2051

Abstract

Restoring information obstructed by hair is one of the main issues for the accurate analysis and segmentation of skin images. For retrieving pixels obstructed by hair, the proposed system converts dermoscopy images into the L*a*b* color space, then principal component analysis (PCA) is [...] Read more.

Restoring information obstructed by hair is one of the main issues for the accurate analysis and segmentation of skin images. For retrieving pixels obstructed by hair, the proposed system converts dermoscopy images into the L*a*b* color space, then principal component analysis (PCA) is applied to produce grayscale images. Afterward, the contrast-limited adaptive histogram equalization (CLAHE) and the average filter are implemented to enhance the grayscale image. Subsequently, the binary image is generated using the iterative thresholding method. After that, the Hough transform (HT) is applied to each image block to generate the hair mask. Finally, the hair pixels are removed by harmonic inpainting. The performance of the proposed automated hair removal was evaluated by applying the proposed system to the International Skin Imaging Collaboration (ISIC) dermoscopy dataset as well as to clinical images. Six performance evaluation metrics were measured, namely the mean squared error (MSE), the peak signal-to-noise ratio (PSNR), the signal-to-noise ratio (SNR), the structural similarity index (SSIM), the universal quality image index (UQI), and the correlation (C). Using the clinical dataset, the system achieved MSE, PSNR, SNR, SSIM, UQI, and C values of 34.7957, 66.98, 42.39, 0.9813, 0.9801, and 0.9985, respectively. The results demonstrated that the proposed system could satisfy the medical diagnostic requirements and achieve the best performance compared to the state-of-art. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

14 pages, 3628 KiB

Open AccessArticle

Image Decomposition Technique Based on Near-Infrared Transmission

by Toto Aminoto, Purnomo Sidi Priambodo and Harry Sudibyo

J. Imaging 2022, 8(12), 322; https://doi.org/10.3390/jimaging8120322 - 3 Dec 2022

Cited by 1 | Viewed by 1746

Abstract

One way to diagnose a disease is to examine pictures of tissue thought to be affected by the disease. Near-infrared properties are subdivided into nonionizing, noninvasive, and nonradiative properties. Near-infrared also has selectivity properties for the objects it passes through. With this selectivity, [...] Read more.

One way to diagnose a disease is to examine pictures of tissue thought to be affected by the disease. Near-infrared properties are subdivided into nonionizing, noninvasive, and nonradiative properties. Near-infrared also has selectivity properties for the objects it passes through. With this selectivity, the resulting attenuation coefficient value will differ depending on the type of material or wavelength. By measuring the output and input intensity values, as well as the attenuation coefficient, the thickness of a material can be measured. The thickness value can then be used to display a reconstructed image. In this study, the object studied was a phantom consisting of silicon rubber, margarine, and gelatin. The results showed that margarine materials could be decomposed from other ingredients with a wavelength of 980 nm. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

26 pages, 27891 KiB

Open AccessArticle

How Well Do Self-Supervised Models Transfer to Medical Imaging?

by Jonah Anton, Liam Castelli, Mun Fai Chan, Mathilde Outters, Wan Hee Tang, Venus Cheung, Pancham Shukla, Rahee Walambe and Ketan Kotecha

J. Imaging 2022, 8(12), 320; https://doi.org/10.3390/jimaging8120320 - 1 Dec 2022

Cited by 2 | Viewed by 5120

Abstract

Self-supervised learning approaches have seen success transferring between similar medical imaging datasets, however there has been no large scale attempt to compare the transferability of self-supervised models against each other on medical images. In this study, we compare the generalisability of seven self-supervised [...] Read more.

Self-supervised learning approaches have seen success transferring between similar medical imaging datasets, however there has been no large scale attempt to compare the transferability of self-supervised models against each other on medical images. In this study, we compare the generalisability of seven self-supervised models, two of which were trained in-domain, against supervised baselines across eight different medical datasets. We find that ImageNet pretrained self-supervised models are more generalisable than their supervised counterparts, scoring up to 10% better on medical classification tasks. The two in-domain pretrained models outperformed other models by over 20% on in-domain tasks, however they suffered significant loss of accuracy on all other tasks. Our investigation of the feature representations suggests that this trend may be due to the models learning to focus too heavily on specific areas. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

21 pages, 5018 KiB

Open AccessArticle

MULTforAD: Multimodal MRI Neuroimaging for Alzheimer’s Disease Detection Based on a 3D Convolution Model

by Walaa N. Ismail, Fathimathul Rajeena P.P and Mona A. S. Ali

Electronics 2022, 11(23), 3893; https://doi.org/10.3390/electronics11233893 - 24 Nov 2022

Cited by 16 | Viewed by 3348

Abstract

Alzheimer’s disease (AD) is a neurological disease that affects numerous people. The condition causes brain atrophy, which leads to memory loss, cognitive impairment, and death. In its early stages, Alzheimer’s disease is tricky to predict. Therefore, treatment provided at an early stage of [...] Read more.

Alzheimer’s disease (AD) is a neurological disease that affects numerous people. The condition causes brain atrophy, which leads to memory loss, cognitive impairment, and death. In its early stages, Alzheimer’s disease is tricky to predict. Therefore, treatment provided at an early stage of AD is more effective and causes less damage than treatment at a later stage. Although AD is a common brain condition, it is difficult to recognize, and its classification requires a discriminative feature representation to separate similar brain patterns. Multimodal neuroimage information that combines multiple medical images can classify and diagnose AD more accurately and comprehensively. Magnetic resonance imaging (MRI) has been used for decades to assist physicians in diagnosing Alzheimer’s disease. Deep models have detected AD with high accuracy in computing-assisted imaging and diagnosis by minimizing the need for hand-crafted feature extraction from MRI images. This study proposes a multimodal image fusion method to fuse MRI neuroimages with a modular set of image preprocessing procedures to automatically fuse and convert Alzheimer’s disease neuroimaging initiative (ADNI) into the BIDS standard for classifying different MRI data of Alzheimer’s subjects from normal controls. Furthermore, a 3D convolutional neural network is used to learn generic features by capturing AlD biomarkers in the fused images, resulting in richer multimodal feature information. Finally, a conventional CNN with three classifiers, including Softmax, SVM, and RF, forecasts and classifies the extracted Alzheimer’s brain multimodal traits from a normal healthy brain. The findings reveal that the proposed method can efficiently predict AD progression by combining high-dimensional MRI characteristics from different public sources with an accuracy range from 88.7% to 99% and outperforming baseline models when applied to MRI-derived voxel features. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

10 pages, 2001 KiB

Open AccessArticle

Putamen Atrophy Is a Possible Clinical Evaluation Index for Parkinson’s Disease Using Human Brain Magnetic Resonance Imaging

by Keisuke Kinoshita, Takehito Kuge, Yoshie Hara and Kojiro Mekata

J. Imaging 2022, 8(11), 299; https://doi.org/10.3390/jimaging8110299 - 2 Nov 2022

Cited by 3 | Viewed by 4334

Abstract

Parkinson’s disease is characterized by motor dysfunction caused by functional deterioration of the substantia nigra. Lower putamen volume (i.e., putamen atrophy) may be an important clinical indicator of motor dysfunction and neurological symptoms, such as autonomic dysfunction, in patients with Parkinson’s disease. We [...] Read more.

Parkinson’s disease is characterized by motor dysfunction caused by functional deterioration of the substantia nigra. Lower putamen volume (i.e., putamen atrophy) may be an important clinical indicator of motor dysfunction and neurological symptoms, such as autonomic dysfunction, in patients with Parkinson’s disease. We proposed and applied a new evaluation method for putamen volume measurement on 31 high-resolution T2-weighted magnetic resonance images from 16 patients with Parkinson’s disease (age, 80.3 ± 7.30 years; seven men, nine women) and 30 such images from 19 control participants (age, 75.1 ± 7.85 years; eleven men, eight women). Putamen atrophy was expressed using a ratio based on the thalamus. The obtained values were used to assess differences between the groups using the Wilcoxon rank-sum test. The intraclass correlation coefficient showed sufficient intra-rater reliability and validity of this method. The Parkinson’s disease group had a significantly lower mean change ratio in the putamen (0.633) than the control group (0.719), suggesting that putamen atrophy may be identified using two-dimensional images. The evaluation method presented in this study may indicate the appearance of motor dysfunction and cognitive decline and could serve as a clinical evaluation index for Parkinson’s disease. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

11 pages, 2168 KiB

Open AccessArticle

Iodine-123 β-methyl-P-iodophenyl-pentadecanoic Acid (¹²³I-BMIPP) Myocardial Scintigraphy for Breast Cancer Patients and Possible Early Signs of Cancer-Therapeutics-Related Cardiac Dysfunction (CTRCD)

by Yuko Harada, Kyosuke Shimada, Satoshi John Harada, Tomomi Sato, Yukino Kubota and Miyoko Yamashita

J. Imaging 2022, 8(11), 296; https://doi.org/10.3390/jimaging8110296 - 29 Oct 2022

Cited by 3 | Viewed by 2223

Abstract

(1) Background: The mortality of breast cancer has decreased due to the advancement of cancer therapies. However, more patients are suffering from cancer-therapeutics-related cardiac dysfunction (CTRCD). Diagnostic and treatment guidelines for CTRCD have not been fully established yet. Ultrasound cardiogram (UCG) is the [...] Read more.

(1) Background: The mortality of breast cancer has decreased due to the advancement of cancer therapies. However, more patients are suffering from cancer-therapeutics-related cardiac dysfunction (CTRCD). Diagnostic and treatment guidelines for CTRCD have not been fully established yet. Ultrasound cardiogram (UCG) is the gold standard for diagnosis of CTRCD, but many breast cancer patients cannot undergo UCG due to the surgery wounds or anatomical reasons. The purpose of the study is to evaluate the usefulness of myocardial scintigraphy using Iodine-123 β-methyl-P-iodophenyl-pentadecanoic acid (¹²³I-BMIPP) in comparison with UCG. (2) Methods: 100 breast cancer patients who received chemotherapy within 3 years underwent Thallium (²⁰¹Tl) and ²³I-BMIPP myocardial perfusion and metabolism scintigraphy. The images were visually evaluated by doctors and radiological technologists, and the grade of uptake reduction was scored by Heart Risk View-S software (Nihon Medi-Physics). The scores were deployed in a 17-segment model of the heart. The distribution of the scores were analyzed. (3) Results: Nine patients (9%) could not undergo UCG. No correlation was found between left ventricular ejection fraction (LVEF) and Heart Risk View-S scores of ²⁰¹Tl myocardial perfusion scintigraphy nor those of BMIPP myocardial metabolism scintigraphy. In a 17-segment model of the heart, the scores of the middle rings were higher than for the basal ring. (4) Conclusions: Evaluation by UCG is not possible for some patients. Myocardial scintigraphy cannot serve as a perfect alternative to UCG. However, it will become the preferable second-choice screening test, as it could point out the early stage of CTRCD. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

18 pages, 2997 KiB

Open AccessArticle

An Intelligent Tongue Diagnosis System via Deep Learning on the Android Platform

by Zibin Yang, Yuping Zhao, Jiarui Yu, Xiaobo Mao, Huaxing Xu and Luqi Huang

Diagnostics 2022, 12(10), 2451; https://doi.org/10.3390/diagnostics12102451 - 10 Oct 2022

Cited by 7 | Viewed by 5146

Abstract

To quickly and accurately identify the pathological features of the tongue, we developed an intelligent tongue diagnosis system that uses deep learning on a mobile terminal. We also propose an efficient and accurate tongue image processing algorithm framework to infer the category of [...] Read more.

To quickly and accurately identify the pathological features of the tongue, we developed an intelligent tongue diagnosis system that uses deep learning on a mobile terminal. We also propose an efficient and accurate tongue image processing algorithm framework to infer the category of the tongue. First, a software system integrating registration, login, account management, tongue image recognition, and doctor–patient dialogue was developed based on the Android platform. Then, the deep learning models, based on the official benchmark models, were trained by using the tongue image datasets. The tongue diagnosis algorithm framework includes the YOLOv5s6, U-Net, and MobileNetV3 networks, which are employed for tongue recognition, tongue region segmentation, and tongue feature classification (tooth marks, spots, and fissures), respectively. The experimental results demonstrate that the performance of the tongue diagnosis model was satisfying, and the accuracy of the final classification of tooth marks, spots, and fissures was 93.33%, 89.60%, and 97.67%, respectively. The construction of this system has a certain reference value for the objectification and intelligence of tongue diagnosis. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

14 pages, 3185 KiB

Open AccessArticle

Using an Ultrasound Tissue Phantom Model for Hybrid Training of Deep Learning Models for Shrapnel Detection

by Sofia I. Hernandez-Torres, Emily N. Boice and Eric J. Snider

J. Imaging 2022, 8(10), 270; https://doi.org/10.3390/jimaging8100270 - 2 Oct 2022

Cited by 7 | Viewed by 2901

Abstract

Tissue phantoms are important for medical research to reduce the use of animal or human tissue when testing or troubleshooting new devices or technology. Development of machine-learning detection tools that rely on large ultrasound imaging data sets can potentially be streamlined with high [...] Read more.

Tissue phantoms are important for medical research to reduce the use of animal or human tissue when testing or troubleshooting new devices or technology. Development of machine-learning detection tools that rely on large ultrasound imaging data sets can potentially be streamlined with high quality phantoms that closely mimic important features of biological tissue. Here, we demonstrate how an ultrasound-compliant tissue phantom comprised of multiple layers of gelatin to mimic bone, fat, and muscle tissue types can be used for machine-learning training. This tissue phantom has a heterogeneous composition to introduce tissue level complexity and subject variability in the tissue phantom. Various shrapnel types were inserted into the phantom for ultrasound imaging to supplement swine shrapnel image sets captured for applications such as deep learning algorithms. With a previously developed shrapnel detection algorithm, blind swine test image accuracy reached more than 95% accuracy when training was comprised of 75% tissue phantom images, with the rest being swine images. For comparison, a conventional MobileNetv2 deep learning model was trained with the same training image set and achieved over 90% accuracy in swine predictions. Overall, the tissue phantom demonstrated high performance for developing deep learning models for ultrasound image classification. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

12 pages, 2387 KiB

Open AccessArticle

The Feasibility of Shadowed Image Restoration Using the Synthetic Aperture Focusing Technique

by Kuo-Yuan Huang, Chih-Hsiung Chang, Young-Fo Chang, Jia-Wei Liu and Jer-Wei Chang

Appl. Sci. 2022, 12(18), 9297; https://doi.org/10.3390/app12189297 - 16 Sep 2022

Cited by 1 | Viewed by 1756

Abstract

The phenomenon of acoustic shadowing on ultrasonography is characterized by an echo signal void behind structures that strongly absorb or reflect ultrasonic energy. In medical ultrasonography, once the ultrasound energy is shielded, acoustic shadowing makes it difficult to create an image, leading to [...] Read more.

The phenomenon of acoustic shadowing on ultrasonography is characterized by an echo signal void behind structures that strongly absorb or reflect ultrasonic energy. In medical ultrasonography, once the ultrasound energy is shielded, acoustic shadowing makes it difficult to create an image, leading to misinterpretations and obscure diagnoses. Hence, instead of dealing with the defocused problem encountered in an ultrasound scan (US), this current research focuses on revealing the existence of an acoustically shadowed target (or a potential lesion) using a well-known restoration algorithm, i.e., the synthetic aperture focusing technique (SAFT). To demonstrate the effects of an acoustic shadow on an ultrasound scan (US), a forward model study is carried out. In laboratory manipulations, a purposely designed physical model is created and then scanned using B-mode and pitch/catch arrangements to carry out shadowed and shadow-free scans in a water tank. Thereafter, making use of a delay-and-sum (DAS) operation, the echo signals are processed by the synthetic aperture focusing technique (SAFT) to perform image restoration. The results of the restoration process show that the SAFT algorithm performs well with respect to directional shadowing. Once the target or lesion is positioned in a total anechoic zone, or even in a multi-channel scan, it will fail. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

33 pages, 5637 KiB

Open AccessArticle

CLAIRE—Parallelized Diffeomorphic Image Registration for Large-Scale Biomedical Imaging Applications

by Naveen Himthani, Malte Brunn, Jae-Youn Kim, Miriam Schulte, Andreas Mang and George Biros

J. Imaging 2022, 8(9), 251; https://doi.org/10.3390/jimaging8090251 - 16 Sep 2022

Cited by 3 | Viewed by 2739

Abstract

We study the performance of CLAIRE—a diffeomorphic multi-node, multi-GPU image-registration algorithm and software—in large-scale biomedical imaging applications with billions of voxels. At such resolutions, most existing software packages for diffeomorphic image registration are prohibitively expensive. As a result, practitioners first significantly downsample the [...] Read more.

We study the performance of CLAIRE—a diffeomorphic multi-node, multi-GPU image-registration algorithm and software—in large-scale biomedical imaging applications with billions of voxels. At such resolutions, most existing software packages for diffeomorphic image registration are prohibitively expensive. As a result, practitioners first significantly downsample the original images and then register them using existing tools. Our main contribution is an extensive analysis of the impact of downsampling on registration performance. We study this impact by comparing full-resolution registrations obtained with CLAIRE to lower resolution registrations for synthetic and real-world imaging datasets. Our results suggest that registration at full resolution can yield a superior registration quality—but not always. For example, downsampling a synthetic image from

1024^{3}

to

256^{3}

decreases the Dice coefficient from 92% to 79%. However, the differences are less pronounced for noisy or low contrast high resolution images. CLAIRE allows us not only to register images of clinically relevant size in a few seconds but also to register images at unprecedented resolution in reasonable time. The highest resolution considered are CLARITY images of size

2816 \times 3016 \times 1162

. To the best of our knowledge, this is the first study on image registration quality at such resolutions. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

3 pages, 736 KiB

Open AccessInteresting Images

Chronic Headache Attributed to Vertebrobasilar Insufficiency

by Sang Woo Ha, Young Seo Kim, Eun Joo Yoon and Hyun Goo Kang

Diagnostics 2022, 12(9), 2038; https://doi.org/10.3390/diagnostics12092038 - 23 Aug 2022

Cited by 1 | Viewed by 2064

Abstract

Vertebrobasilar insufficiency, a condition characterized by poor blood flow to the posterior portion of the brain, can cause headaches. However, the exact underlying mechanism is not yet fully understood. The patient enrolled in our study reported experiencing intermittent headaches radiating from the left [...] Read more.

Vertebrobasilar insufficiency, a condition characterized by poor blood flow to the posterior portion of the brain, can cause headaches. However, the exact underlying mechanism is not yet fully understood. The patient enrolled in our study reported experiencing intermittent headaches radiating from the left shoulder, similar to chronic tension-type headaches. His aggravated headache and severe left vertebral artery stenosis were detected by brain computed tomography angiography. Stent insertion successfully expanded the patient’s narrowed left vertebral artery orifice. Subsequently, the patient’s headaches improved without recurrence during the one-year follow-up period. In summary, chronic headaches attributed to vertebrobasilar insufficiency in this study, improved after stent insertion to reverse severe left vertebral artery stenosis. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

17 pages, 6774 KiB

Open AccessArticle

Deep Segmentation Networks for Segmenting Kidneys and Detecting Kidney Stones in Unenhanced Abdominal CT Images

by Dan Li, Chuda Xiao, Yang Liu, Zhuo Chen, Haseeb Hassan, Liyilei Su, Jun Liu, Haoyu Li, Weiguo Xie, Wen Zhong and Bingding Huang

Diagnostics 2022, 12(8), 1788; https://doi.org/10.3390/diagnostics12081788 - 23 Jul 2022

Cited by 35 | Viewed by 11290

Abstract

Recent breakthroughs of deep learning algorithms in medical imaging, automated detection, and segmentation techniques for renal (kidney) in abdominal computed tomography (CT) images have been limited. Radiomics and machine learning analyses of renal diseases rely on the automatic segmentation of kidneys in CT [...] Read more.

Recent breakthroughs of deep learning algorithms in medical imaging, automated detection, and segmentation techniques for renal (kidney) in abdominal computed tomography (CT) images have been limited. Radiomics and machine learning analyses of renal diseases rely on the automatic segmentation of kidneys in CT images. Inspired by this, our primary aim is to utilize deep semantic segmentation learning models with a proposed training scheme to achieve precise and accurate segmentation outcomes. Moreover, this work aims to provide the community with an open-source, unenhanced abdominal CT dataset for training and testing the deep learning segmentation networks to segment kidneys and detect kidney stones. Five variations of deep segmentation networks are trained and tested both dependently (based on the proposed training scheme) and independently. Upon comparison, the models trained with the proposed training scheme enable the highly accurate 2D and 3D segmentation of kidneys and kidney stones. We believe this work is a fundamental step toward AI-driven diagnostic strategies, which can be an essential component of personalized patient care and improved decision-making in treating kidney diseases. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

19 pages, 1505 KiB

Open AccessArticle

Lung Volume Calculation in Preclinical MicroCT: A Fast Geometrical Approach

by Juan Antonio Camara, Anna Pujol, Juan Jose Jimenez, Jaime Donate, Marina Ferrer and Greetje Vande Velde

J. Imaging 2022, 8(8), 204; https://doi.org/10.3390/jimaging8080204 - 22 Jul 2022

Cited by 1 | Viewed by 2815

Abstract

In this study, we present a time-efficient protocol for thoracic volume calculation as a proxy for total lung volume. We hypothesize that lung volume can be calculated indirectly from this thoracic volume. We compared the measured thoracic volume with manually segmented and automatically [...] Read more.

In this study, we present a time-efficient protocol for thoracic volume calculation as a proxy for total lung volume. We hypothesize that lung volume can be calculated indirectly from this thoracic volume. We compared the measured thoracic volume with manually segmented and automatically thresholded lung volumes, with manual segmentation as the gold standard. A linear regression formula was obtained and used for calculating the theoretical lung volume. This volume was compared with the gold standard volumes. In healthy animals, thoracic volume was 887.45 mm³, manually delineated lung volume 554.33 mm³ and thresholded aerated lung volume 495.38 mm³ on average. Theoretical lung volume was 554.30 mm³. Finally, the protocol was applied to three animal models of lung pathology (lung metastasis and transgenic primary lung tumor and fungal infection). In confirmed pathologic animals, thoracic volumes were: 893.20 mm³, 860.12 and 1027.28 mm³. Manually delineated volumes were 640.58, 503.91 and 882.42 mm³, respectively. Thresholded lung volumes were 315.92 mm³, 408.72 and 236 mm³, respectively. Theoretical lung volume resulted in 635.28, 524.30 and 863.10.42 mm³. No significant differences were observed between volumes. This confirmed the potential use of this protocol for lung volume calculation in pathologic models. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

11 pages, 1073 KiB

Open AccessArticle

Influence of Prior Imaging Information on Diagnostic Accuracy for Focal Skeletal Processes—A Retrospective Analysis of the Consistency between Biopsy-Verified Imaging Diagnoses

by Mine Benedicte Lange, Lars J. Petersen, Mads Lausen, Niels Henrik Bruun, Michael Bachmann Nielsen and Helle D. Zacho

Diagnostics 2022, 12(7), 1735; https://doi.org/10.3390/diagnostics12071735 - 17 Jul 2022

Cited by 1 | Viewed by 2233

Abstract

Introduction: Comparing imaging examinations with those previously obtained is considered mandatory in imaging guidelines. To our knowledge, no studies are available on neither the influence, nor the sequence, of prior imaging and reports on diagnostic accuracy using biopsy as the reference standard. Such [...] Read more.

Introduction: Comparing imaging examinations with those previously obtained is considered mandatory in imaging guidelines. To our knowledge, no studies are available on neither the influence, nor the sequence, of prior imaging and reports on diagnostic accuracy using biopsy as the reference standard. Such data are important to minimize diagnostic errors and to improve the preparation of diagnostic imaging guidelines. The aim of our study was to provide such data. Materials and methods: A retrospective cohort of 216 consecutive skeletal biopsies from patients with at least 2 different imaging modalities (X-ray, CT and MRI) performed within 6 months of biopsy was identified. The diagnostic accuracy of the individual imaging modality was assessed. Finally, the possible influence of the sequence of imaging modalities was investigated. Results: No significant difference in the accuracy of the imaging modalities was shown, being preceded by another imaging modality or not. However, the sequence analyses indicate sequential biases, particularly if MRI was the first imaging modality. Conclusion: The sequence of the imaging modalities seems to influence the diagnostic accuracy against a pathology reference standard. Further studies are needed to establish evidence-based guidelines for the strategy of using previous imaging and reports to improve diagnostic accuracy. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

12 pages, 1320 KiB

Open AccessArticle

The Necessity of Magnetic Resonance Imaging in Congenital Diaphragmatic Hernia

by Erick George Neștianu, Cristina Guramba Brădeanu, Dragoș Ovidiu Alexandru and Radu Vlădăreanu

Diagnostics 2022, 12(7), 1733; https://doi.org/10.3390/diagnostics12071733 - 17 Jul 2022

Viewed by 1913

Abstract

This is a retrospective study investigating the relationship between ultrasound and magnetic resonance imaging (MRI) examinations in congenital diaphragmatic hernia (CDH). CDH is a rare cause of pulmonary hypoplasia that increases the mortality and morbidity of patients. Inclusion criteria were: patients diagnosed with [...] Read more.

This is a retrospective study investigating the relationship between ultrasound and magnetic resonance imaging (MRI) examinations in congenital diaphragmatic hernia (CDH). CDH is a rare cause of pulmonary hypoplasia that increases the mortality and morbidity of patients. Inclusion criteria were: patients diagnosed with CDH who underwent MRI examination after the second-trimester morphology ultrasound confirmed the presence of CDH. The patients came from three university hospitals in Bucharest, Romania. A total of 22 patients were included in the study after applying the exclusion criteria. By analyzing the total lung volume (TLV) using MRI, and the lung to head ratio (LHR) calculated using MRI and ultrasound, we observed that LHR can severely underestimate the severity of the pulmonary hypoplasia, even showing values close to normal in some cases. This also proves to be statistically relevant if we eliminate certain extreme values. We found significant correlations between the LHR percentage and herniated organs, such as the left and right liver lobes and gallbladder. MRI also provided additional insights, indicating the presence of pericarditis or pleurisy. We wish to underline the necessity of MRI follow-up in all cases of CDH, as the accurate measurement of the TLV is important for future treatment and therapeutic strategy. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

14 pages, 13709 KiB

Open AccessArticle

Force Estimation during Cell Migration Using Mathematical Modelling

by Fengwei Yang, Chandrasekhar Venkataraman, Sai Gu, Vanessa Styles and Anotida Madzvamuse

J. Imaging 2022, 8(7), 199; https://doi.org/10.3390/jimaging8070199 - 15 Jul 2022

Cited by 1 | Viewed by 2388

Abstract

Cell migration is essential for physiological, pathological and biomedical processes such as, in embryogenesis, wound healing, immune response, cancer metastasis, tumour invasion and inflammation. In light of this, quantifying mechanical properties during the process of cell migration is of great interest in experimental [...] Read more.

Cell migration is essential for physiological, pathological and biomedical processes such as, in embryogenesis, wound healing, immune response, cancer metastasis, tumour invasion and inflammation. In light of this, quantifying mechanical properties during the process of cell migration is of great interest in experimental sciences, yet few theoretical approaches in this direction have been studied. In this work, we propose a theoretical and computational approach based on the optimal control of geometric partial differential equations to estimate cell membrane forces associated with cell polarisation during migration. Specifically, cell membrane forces are inferred or estimated by fitting a mathematical model to a sequence of images, allowing us to capture dynamics of the cell migration. Our approach offers a robust and accurate framework to compute geometric mechanical membrane forces associated with cell polarisation during migration and also yields geometric information of independent interest, we illustrate one such example that involves quantifying cell proliferation levels which are associated with cell division, cell fusion or cell death. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

20 pages, 5094 KiB

Open AccessArticle

Clinically Inspired Skin Lesion Classification through the Detection of Dermoscopic Criteria for Basal Cell Carcinoma

by Carmen Serrano, Manuel Lazo, Amalia Serrano, Tomás Toledo-Pastrana, Rubén Barros-Tornay and Begoña Acha

J. Imaging 2022, 8(7), 197; https://doi.org/10.3390/jimaging8070197 - 12 Jul 2022

Cited by 15 | Viewed by 3061

Abstract

Background and Objective. Skin cancer is the most common cancer worldwide. One of the most common non-melanoma tumors is basal cell carcinoma (BCC), which accounts for 75% of all skin cancers. There are many benign lesions that can be confused with these [...] Read more.

Background and Objective. Skin cancer is the most common cancer worldwide. One of the most common non-melanoma tumors is basal cell carcinoma (BCC), which accounts for 75% of all skin cancers. There are many benign lesions that can be confused with these types of cancers, leading to unnecessary biopsies. In this paper, a new method to identify the different BCC dermoscopic patterns present in a skin lesion is presented. In addition, this information is applied to classify skin lesions into BCC and non-BCC. Methods. The proposed method combines the information provided by the original dermoscopic image, introduced in a convolutional neural network (CNN), with deep and handcrafted features extracted from color and texture analysis of the image. This color analysis is performed by transforming the image into a uniform color space and into a color appearance model. To demonstrate the validity of the method, a comparison between the classification obtained employing exclusively a CNN with the original image as input and the classification with additional color and texture features is presented. Furthermore, an exhaustive comparison of classification employing different color and texture measures derived from different color spaces is presented. Results. Results show that the classifier with additional color and texture features outperforms a CNN whose input is only the original image. Another important achievement is that a new color cooccurrence matrix, proposed in this paper, improves the results obtained with other texture measures. Finally, sensitivity of 0.99, specificity of 0.94 and accuracy of 0.97 are achieved when lesions are classified into BCC or non-BCC. Conclusions. To the best of our knowledge, this is the first time that a methodology to detect all the possible patterns that can be present in a BCC lesion is proposed. This detection leads to a clinically explainable classification into BCC and non-BCC lesions. In this sense, the classification of the proposed tool is based on the detection of the dermoscopic features that dermatologists employ for their diagnosis. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

18 pages, 3296 KiB

Open AccessArticle

Artificial Intelligence-Based Multimodal Medical Image Fusion Using Hybrid S² Optimal CNN

by Marwah Mohammad Almasri and Abrar Mohammed Alajlan

Electronics 2022, 11(14), 2124; https://doi.org/10.3390/electronics11142124 - 6 Jul 2022

Cited by 11 | Viewed by 3451

Abstract

In medical applications, medical image fusion methods are capable of fusing the medical images from various morphologies to obtain a reliable medical diagnosis. A single modality image cannot provide sufficient information for an exact diagnosis. Hence, an efficient multimodal medical image fusion-based artificial [...] Read more.

In medical applications, medical image fusion methods are capable of fusing the medical images from various morphologies to obtain a reliable medical diagnosis. A single modality image cannot provide sufficient information for an exact diagnosis. Hence, an efficient multimodal medical image fusion-based artificial intelligence model is proposed in this paper. Initially, the multimodal medical images are obtained for an effective fusion process by using a modified discrete wavelet transform (MDWT) thereby attaining an image with high visual clarity. Then, the fused images are classified as malignant or benign using the proposed convolutional neural network-based hybrid optimization dynamic algorithm (CNN-HOD). To enhance the weight function and classification accuracy of the CNN, a hybrid optimization dynamic algorithm (HOD) is proposed. The HOD is the integration of the sailfish optimizer algorithm and seagull optimization algorithm. Here, the seagull optimizer algorithm replaces the migration operation toobtain the optimal location. The experimental analysis is carried out and acquired with standard deviation (58%), average gradient (88%), and fusion factor (73%) compared with the other approaches. The experimental results demonstrate that the proposed approach performs better than other approaches and offers high-quality fused images for an accurate diagnosis. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

15 pages, 17499 KiB

Open AccessArticle

Pneumonia Detection on Chest X-ray Images Using Ensemble of Deep Convolutional Neural Networks

by Alhassan Mabrouk, Rebeca P. Díaz Redondo, Abdelghani Dahou, Mohamed Abd Elaziz and Mohammed Kayed

Appl. Sci. 2022, 12(13), 6448; https://doi.org/10.3390/app12136448 - 25 Jun 2022

Cited by 56 | Viewed by 9085

Abstract

Pneumonia is a life-threatening lung infection resulting from several different viral infections. Identifying and treating pneumonia on chest X-ray images can be difficult due to its similarity to other pulmonary diseases. Thus, the existing methods for predicting pneumonia cannot attain substantial levels of [...] Read more.

Pneumonia is a life-threatening lung infection resulting from several different viral infections. Identifying and treating pneumonia on chest X-ray images can be difficult due to its similarity to other pulmonary diseases. Thus, the existing methods for predicting pneumonia cannot attain substantial levels of accuracy. This paper presents a computer-aided classification of pneumonia, coined Ensemble Learning (EL), to simplify the diagnosis process on chest X-ray images. Our proposal is based on Convolutional Neural Network (CNN) models, which are pretrained CNN models that have been recently employed to enhance the performance of many medical tasks instead of training CNN models from scratch. We propose to use three well-known CNNs (DenseNet169, MobileNetV2, and Vision Transformer) pretrained using the ImageNet database. These models are trained on the chest X-ray data set using fine-tuning. Finally, the results are obtained by combining the extracted features from these three models during the experimental phase. The proposed EL approach outperforms other existing state-of-the-art methods and obtains an accuracy of 93.91% and a F1-score of 93.88% on the testing phase. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

19 pages, 5376 KiB

Open AccessArticle

Combining High-Resolution Hard X-ray Tomography and Histology for Stem Cell-Mediated Distraction Osteogenesis

by Griffin Rodgers, Guido R. Sigron, Christine Tanner, Simone E. Hieber, Felix Beckmann, Georg Schulz, Arnaud Scherberich, Claude Jaquiéry, Christoph Kunz and Bert Müller

Appl. Sci. 2022, 12(12), 6286; https://doi.org/10.3390/app12126286 - 20 Jun 2022

Cited by 2 | Viewed by 2147

Abstract

Distraction osteogenesis is a clinically established technique for lengthening, molding and shaping bone by new bone formation. The experimental evaluation of this expensive and time-consuming treatment is of high impact for better understanding of tissue engineering but mainly relies on a limited number [...] Read more.

Distraction osteogenesis is a clinically established technique for lengthening, molding and shaping bone by new bone formation. The experimental evaluation of this expensive and time-consuming treatment is of high impact for better understanding of tissue engineering but mainly relies on a limited number of histological slices. These tissue slices contain two-dimensional information comprising only about one percent of the volume of interest. In order to analyze the soft and hard tissues of the entire jaw of a single rat in a multimodal assessment, we combined micro computed tomography (µCT) with histology. The µCT data acquired before and after decalcification were registered to determine the impact of decalcification on local tissue shrinkage. Identification of the location of the H&E-stained specimen within the synchrotron radiation-based µCT data collected after decalcification was achieved via non-rigid slice-to-volume registration. The resulting bi- and tri-variate histograms were divided into clusters related to anatomical features from bone and soft tissues, which allowed for a comparison of the approaches and resulted in the hypothesis that the combination of laboratory-based µCT before decalcification, synchrotron radiation-based µCT after decalcification and histology with hematoxylin-and-eosin staining could be used to discriminate between different types of collagen, key components of new bone formation. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Graphical abstract

13 pages, 2493 KiB

Open AccessArticle

Agreement of the Discrepancy Index Obtained Using Digital and Manual Techniques—A Comparative Study

by Nestor A. Burgos-Arcega, Rogelio J. Scougall-Vilchis, Adriana A. Morales-Valenzuela, Wael Hegazy-Hassan, Edith Lara-Carrillo, Víctor H. Toral-Rizo, Ulises Velázquez-Enríquez and Elias N. Salmerón-Valdés

Appl. Sci. 2022, 12(12), 6105; https://doi.org/10.3390/app12126105 - 16 Jun 2022

Cited by 1 | Viewed by 2705

Abstract

The discrepancy index evaluates the complexity of the initial orthodontic diagnosis. The objective is to compare whether there is a difference in the final discrepancy index score of the American Board of Orthodontics (ABO) when obtained using digital and manual techniques. Fifty-six initial [...] Read more.

The discrepancy index evaluates the complexity of the initial orthodontic diagnosis. The objective is to compare whether there is a difference in the final discrepancy index score of the American Board of Orthodontics (ABO) when obtained using digital and manual techniques. Fifty-six initial orthodontic records in a digital and physical format were included (28 each) in 2022 at the Center for Research and Advanced Studies in Dentistry. For the digital measurements, iTero and TRIOS 3 intraoral scanners were used, along with Insignia software and cephalometric tracing with Dolphin Imaging software. Manual measurements were obtained in dental casts using the ruler indicated for the previously mentioned discrepancy index, in addition to conventional cephalometric tracing. Student’s t-test did not show statistically significant differences between the digital and manual techniques, with final discrepancy index scores of 24.61 (13.34) and 24.86 (14.14), respectively (p = 0.769). Cohen’s kappa index showed very good agreement between both categorical measurements (kappa value = 1.00, p = 0.001). The Bland–Altman method demonstrated a good agreement between continuous measurements obtained by both techniques with a bias of 0.2500 (superior limit of agreement =9.0092988, inferior limit of agreement = −8.5092988). Excellent agreement was observed in obtaining the discrepancy index through digital technique (Intraoral scanning and digital records) and manual technique (conventional records). Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

9 pages, 872 KiB

Open AccessArticle

Mobile-PolypNet: Lightweight Colon Polyp Segmentation Network for Low-Resource Settings

by Ranit Karmakar and Saeid Nooshabadi

J. Imaging 2022, 8(6), 169; https://doi.org/10.3390/jimaging8060169 - 14 Jun 2022

Cited by 6 | Viewed by 2787

Abstract

Colon polyps, small clump of cells on the lining of the colon, can lead to colorectal cancer (CRC), one of the leading types of cancer globally. Hence, early detection of these polyps automatically is crucial in the prevention of CRC. The deep learning [...] Read more.

Colon polyps, small clump of cells on the lining of the colon, can lead to colorectal cancer (CRC), one of the leading types of cancer globally. Hence, early detection of these polyps automatically is crucial in the prevention of CRC. The deep learning models proposed for the detection and segmentation of colorectal polyps are resource-consuming. This paper proposes a lightweight deep learning model for colorectal polyp segmentation that achieved state-of-the-art accuracy while significantly reducing the model size and complexity. The proposed deep learning autoencoder model employs a set of state-of-the-art architectural blocks and optimization objective functions to achieve the desired efficiency. The model is trained and tested on five publicly available colorectal polyp segmentation datasets (CVC-ClinicDB, CVC-ColonDB, EndoScene, Kvasir, and ETIS). We also performed ablation testing on the model to test various aspects of the autoencoder architecture. We performed the model evaluation by using most of the common image-segmentation metrics. The backbone model achieved a DICE score of 0.935 on the Kvasir dataset and 0.945 on the CVC-ClinicDB dataset, improving the accuracy by 4.12% and 5.12%, respectively, over the current state-of-the-art network, while using 88 times fewer parameters, 40 times less storage space, and being computationally 17 times more efficient. Our ablation study showed that the addition of ConvSkip in the autoencoder slightly improves the model’s performance but it was not significant (p-value = 0.815). Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

11 pages, 2114 KiB

Open AccessArticle

Low-Dose High-Resolution Photon-Counting CT of the Lung: Radiation Dose and Image Quality in the Clinical Routine

by Matthias Michael Woeltjen, Julius Henning Niehoff, Arwed Elias Michael, Sebastian Horstmeier, Christoph Moenninghoff, Jan Borggrefe and Jan Robert Kroeger

Diagnostics 2022, 12(6), 1441; https://doi.org/10.3390/diagnostics12061441 - 11 Jun 2022

Cited by 37 | Viewed by 3941

Abstract

This study aims to investigate the qualitative and quantitative image quality of low-dose high-resolution (LD-HR) lung CT scans acquired with the first clinical approved photon counting CT (PCCT) scanner. Furthermore, the radiation dose used by the PCCT is compared to a conventional CT [...] Read more.

This study aims to investigate the qualitative and quantitative image quality of low-dose high-resolution (LD-HR) lung CT scans acquired with the first clinical approved photon counting CT (PCCT) scanner. Furthermore, the radiation dose used by the PCCT is compared to a conventional CT scanner with an energy-integrating detector system (EID-CT). Twenty-nine patients who underwent a LD-HR chest CT scan with dual-source PCCT and had previously undergone a LD-HR chest CT with a standard EID-CT scanner were retrospectively included in this study. Images of the whole lung as well as enlarged image sections displaying a specific finding (lesion) were evaluated in terms of overall image quality, image sharpness and image noise by three senior radiologists using a 5-point Likert scale. The PCCT images were reconstructed with and without a quantum iterative reconstruction algorithm (PCCT QIR+/−). Noise and signal-to-noise (SNR) were measured and the effective radiation dose was calculated. Overall, image quality and image sharpness were rated best in PCCT (QIR+) images. A significant difference was seen particularly in image sections of PCCT (QIR+) images compared to EID-CT images (p < 0.005). Image noise of PCCT (QIR+) images was significantly lower compared to EID-CT images in image sections (p = 0.005). In contrast, noise was lowest on EID-CT images (p < 0.001). The PCCT used significantly less radiation dose compared to the EID-CT (p < 0.001). In conclusion, LD-HR PCCT scans of the lung provide better image quality while using significantly less radiation dose compared to EID-CT scans. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

8 pages, 1108 KiB

Open AccessArticle

Altered Transmission of Cardiac Cycles to Ductus Venosus Blood Flow in Fetal Growth Restriction: Why Ductus Venosus Reflects Fetal Circulatory Changes More Precisely

by Naomi Seo, Yasushi Kurihara, Tomoki Suekane, Natsuko Yokoi, Kayoko Nakagawa, Mie Tahara, Akihiro Hamuro, Takuya Misugi, Akemi Nakano, Masayasu Koyama and Daisuke Tachibana

Diagnostics 2022, 12(6), 1393; https://doi.org/10.3390/diagnostics12061393 - 4 Jun 2022

Viewed by 2271

Abstract

We aimed to investigate the relation between the time intervals of the flow velocity waveform of ductus venosus (DV-FVW) and cardiac cycles. We defined Delta A as the difference in the time measurements between DV-FVW and cardiac cycles on the assumption that the [...] Read more.

We aimed to investigate the relation between the time intervals of the flow velocity waveform of ductus venosus (DV-FVW) and cardiac cycles. We defined Delta A as the difference in the time measurements between DV-FVW and cardiac cycles on the assumption that the second peak of ductus venosus (D-wave) starts simultaneously with the opening of the mitral valve (MV). As well, we defined Delta B as the difference of the time measurements between DV-FVW and cardiac cycles on the assumption that the D-wave starts simultaneously with the closure of the aortic valve (AV). We then compared Delta A and Delta B in the control and fetal growth restriction (FGR) groups. In the control group of healthy fetuses, Delta A was strikingly shorter than Delta B. On the other hand, in all FGR cases, no difference was observed. The acceleration of the D-wave is suggested to be generated by the opening of the MV under normal fetal hemodynamics, whereas it precedes the opening of the MV in FGR. Our results indicate that the time interval of DV analysis might be a more informative parameter than the analysis of cardiac cycles. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

25 pages, 16364 KiB

Open AccessArticle

Low-Cost Probabilistic 3D Denoising with Applications for Ultra-Low-Radiation Computed Tomography

by Illia Horenko, Lukáš Pospíšil, Edoardo Vecchi, Steffen Albrecht, Alexander Gerber, Beate Rehbock, Albrecht Stroh and Susanne Gerber

J. Imaging 2022, 8(6), 156; https://doi.org/10.3390/jimaging8060156 - 31 May 2022

Cited by 5 | Viewed by 3233

Abstract

We propose a pipeline for synthetic generation of personalized Computer Tomography (CT) images, with a radiation exposure evaluation and a lifetime attributable risk (LAR) assessment. We perform a patient-specific performance evaluation for a broad range of denoising algorithms (including the most popular deep [...] Read more.

We propose a pipeline for synthetic generation of personalized Computer Tomography (CT) images, with a radiation exposure evaluation and a lifetime attributable risk (LAR) assessment. We perform a patient-specific performance evaluation for a broad range of denoising algorithms (including the most popular deep learning denoising approaches, wavelets-based methods, methods based on Mumford–Shah denoising, etc.), focusing both on accessing the capability to reduce the patient-specific CT-induced LAR and on computational cost scalability. We introduce a parallel Probabilistic Mumford–Shah denoising model (PMS) and show that it markedly-outperforms the compared common denoising methods in denoising quality and cost scaling. In particular, we show that it allows an approximately 22-fold robust patient-specific LAR reduction for infants and a 10-fold LAR reduction for adults. Using a normal laptop, the proposed algorithm for PMS allows cheap and robust (with a multiscale structural similarity index >90%) denoising of very large 2D videos and 3D images (with over

10^{7}

voxels) that are subject to ultra-strong noise (Gaussian and non-Gaussian) for signal-to-noise ratios far below 1.0. The code is provided for open access. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

13 pages, 3376 KiB

Open AccessArticle

Thyroid Biokinetics for Radioactive I-131 in Twelve Thyroid Cancer Patients via the Refined Nine-Compartmental Model

by Lung-Fa Pan, Chao-Yu Chiang, Chao-Chun Huang, Hua-Tsan Kao, Chih-Feng Chen, Bing-Ru Peng and Lung-Kwang Pan

Appl. Sci. 2022, 12(11), 5538; https://doi.org/10.3390/app12115538 - 30 May 2022

Cited by 2 | Viewed by 2365

Abstract

The thyroid biokinetic model of radioactive I-131 was re-evaluated using a refined nine-compartmental model and applied to twelve thyroid cancer patients. In contrast to the simplified four-compartmental model regulated by the ICRP-56 report, the revised model included nine compartments specified in the ICRP-128 [...] Read more.

The thyroid biokinetic model of radioactive I-131 was re-evaluated using a refined nine-compartmental model and applied to twelve thyroid cancer patients. In contrast to the simplified four-compartmental model regulated by the ICRP-56 report, the revised model included nine compartments specified in the ICRP-128 report, namely, oral, stomach, body fluid, thyroid, whole body, liver, kidney, bladder, and remainder (i.e., the whole body minus kidney and bladder). A self-developed program run in MATLAB was designed to solve the nine first-order simultaneous linear differential equations. The model was realized in standard and simplified versions. The latter neglected two feedback paths (body fluid to oral, i₃₁, and kidney to the whole body, i₈₇) to reduce computations. Accordingly, the biological half-lives for the major compartments (thyroid and body fluid + whole body) were 36.00 ± 15.01, 15.04 ± 5.63, 34.33 ± 15.42, and 14.83 ± 5.91 of standard and simplified version. The correlations between theoretical and empirical data for each patient were quantified by the dimensionless AT (agreement) index and, the AT_tot index integrated each individual AT of a specific organ of one patient. Since small AT values indicated a closer correlation, the obtained range of AT_tot (0.048 ± 0.019) proved the standard model’s reliability and high accuracy, while the simplified one yielded slightly higher AT_tot (0.058 ± 0.023). The detailed outcomes among various compartments of twelve patients were calculated and compared with other researchers’ work. The correlation results on radioactive I-131 evolution in thyroid cancer patients’ bodies are instrumental in viewpoint of radioactive protection of patients and radiological personnel. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

11 pages, 4023 KiB

Open AccessArticle

Comparison Study of Myocardial Radiomics Feature Properties on Energy-Integrating and Photon-Counting Detector CT

by Isabelle Ayx, Hishan Tharmaseelan, Alexander Hertel, Dominik Nörenberg, Daniel Overhoff, Lukas T. Rotkopf, Philipp Riffel, Stefan O. Schoenberg and Matthias F. Froelich

Diagnostics 2022, 12(5), 1294; https://doi.org/10.3390/diagnostics12051294 - 23 May 2022

Cited by 19 | Viewed by 2827

Abstract

The implementation of radiomics-based, quantitative imaging parameters is hampered by a lack of stability and standardization. Photon-counting computed tomography (PCCT), compared to energy-integrating computed tomography (EICT), does rely on a novel detector technology, promising better spatial resolution and contrast-to-noise ratio. However, its effect [...] Read more.

The implementation of radiomics-based, quantitative imaging parameters is hampered by a lack of stability and standardization. Photon-counting computed tomography (PCCT), compared to energy-integrating computed tomography (EICT), does rely on a novel detector technology, promising better spatial resolution and contrast-to-noise ratio. However, its effect on radiomics feature properties is unknown. This work investigates this topic in myocardial imaging. In this retrospective, single-center IRB-approved study, the left ventricular myocardium was segmented on CT, and the radiomics features were extracted using pyradiomics. To compare features between scanners, a t-test for non-paired samples and F-test was performed, with a threshold of 0.05 set as a benchmark for significance. Feature correlations were calculated by the Pearson correlation coefficient, and visualization was performed with heatmaps. A total of 50 patients (56% male, mean age 56) were enrolled in this study, with equal proportions of PCCT and EICT. First-order features were, nearly, comparable between both groups. However, higher-order features showed a partially significant difference between PCCT and EICT. While first-order radiomics features of left ventricular myocardium show comparability between PCCT and EICT, detected differences of higher-order features may indicate a possible impact of improved spatial resolution, better detection of lower-energy photons, and a better signal-to-noise ratio on texture analysis on PCCT. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

10 pages, 1436 KiB

Open AccessArticle

Repeatability of Contrast-Enhanced Ultrasound to Determine Renal Cortical Perfusion

by Shatha J. Almushayt, Alisa Pham, Bethan E. Phillips, John P. Williams, Maarten W. Taal and Nicholas M. Selby

Diagnostics 2022, 12(5), 1293; https://doi.org/10.3390/diagnostics12051293 - 23 May 2022

Cited by 5 | Viewed by 2692

Abstract

Alterations in renal perfusion play a major role in the pathogenesis of renal diseases. Renal contrast-enhanced ultrasound (CEUS) is increasingly applied to quantify renal cortical perfusion and to assess its change over time, but comprehensive assessment of the technique’s repeatability is lacking. Ten [...] Read more.

Alterations in renal perfusion play a major role in the pathogenesis of renal diseases. Renal contrast-enhanced ultrasound (CEUS) is increasingly applied to quantify renal cortical perfusion and to assess its change over time, but comprehensive assessment of the technique’s repeatability is lacking. Ten adults attended two renal CEUS scans within 14 days. In each session, five destruction/reperfusion sequences were captured. One-phase association was performed to derive the following parameters: acoustic index (AI), mean transit time (mTT), perfusion index (PI), and wash-in rate (WiR). Intra-individual and inter-operator (image analysis) repeatability for the perfusion variables were assessed using intra-class correlation (ICC), with the agreement assessed using a Bland–Altman analysis. The 10 adults had a median (IQR) age of 39 years (30–46). Good intra-individual repeatability was found for mTT (ICC: 0.71) and PI (ICC: 0.65). Lower repeatability was found for AI (ICC: 0.50) and WiR (ICC: 0.56). The correlation between the two operators was excellent for all variables: the ICCs were 0.99 for PI, 0.98 for AI, 0.87 for mTT, and 0.83 for WiR. The Bland–Altman analysis showed that the mean biases (± SD) between the two operators were 0.03 ± 0.16 for mTT, 0.005 ± 0.09 for PI, 0.04 ± 0.19 for AI, and −0.02 ± 0.11 for WiR. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Graphical abstract

17 pages, 2846 KiB

Open AccessArticle

Effects of Different Scan Duration on Brain Effective Connectivity among Default Mode Network Nodes

by Nor Shafiza Abdul Wahab, Noorazrul Yahya, Ahmad Nazlim Yusoff, Rozman Zakaria, Jegan Thanabalan, Elza Othman, Soon Bee Hong, Ramesh Kumar Athi Kumar and Hanani Abdul Manan

Diagnostics 2022, 12(5), 1277; https://doi.org/10.3390/diagnostics12051277 - 20 May 2022

Cited by 6 | Viewed by 2416

Abstract

Background: Resting-state functional magnetic resonance imaging (rs-fMRI) can evaluate brain functional connectivity without requiring subjects to perform a specific task. This rs-fMRI is very useful in patients with cognitive decline or unable to respond to tasks. However, long scan durations have been suggested [...] Read more.

Background: Resting-state functional magnetic resonance imaging (rs-fMRI) can evaluate brain functional connectivity without requiring subjects to perform a specific task. This rs-fMRI is very useful in patients with cognitive decline or unable to respond to tasks. However, long scan durations have been suggested to measure connectivity between brain areas to produce more reliable results, which are not clinically optimal. Therefore, this study aims to evaluate a shorter scan duration and compare the scan duration of 10 and 15 min using the rs-fMRI approach. Methods: Twenty-one healthy male and female participants (seventeen right-handed and four left-handed), with ages ranging between 21 and 60 years, were recruited. All participants underwent both 10 and 15 min of rs-fMRI scans. The present study evaluated the default mode network (DMN) areas for both scan durations. The areas involved were the posterior cingulate cortex (PCC), medial prefrontal cortex (mPFC), left inferior parietal cortex (LIPC), and right inferior parietal cortex (RIPC). Fifteen causal models were constructed and inverted using spectral dynamic causal modelling (spDCM). The models were compared using Bayesian Model Selection (BMS) for group studies. Result: The BMS results indicated that the fully connected model was the winning model among 15 competing models for both 10 and 15 min scan durations. However, there was no significant difference in effective connectivity among the regions of interest between the 10 and 15 min scans. Conclusion: Scan duration in the range of 10 to 15 min is sufficient to evaluate the effective connectivity within the DMN region. In frail subjects, a shorter scan duration is more favourable. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

32 pages, 13986 KiB

Open AccessArticle

Micro-Computed Tomography Soft Tissue Biological Specimens Image Data Visualization

by Branislav Gaspar, Jana Mrzilkova, Jiri Hozman, Petr Zach, Anastasiya Lahutsina, Alexandra Morozova, Giulia Guarnieri and Jitka Riedlova

Appl. Sci. 2022, 12(10), 4918; https://doi.org/10.3390/app12104918 - 12 May 2022

Cited by 4 | Viewed by 3414

Abstract

Visualization of soft tissues in microCT scanning using X-rays is still a complicated matter. There is no simple tool or methodology on how to set up an optimal look-up-table while respecting the type of soft tissue. A partial solution may be the use [...] Read more.

Visualization of soft tissues in microCT scanning using X-rays is still a complicated matter. There is no simple tool or methodology on how to set up an optimal look-up-table while respecting the type of soft tissue. A partial solution may be the use of a contrast agent. However, this must be accompanied by an appropriate look-up-table setting that respects the relationship between the soft tissue type and the Hounsfield units. The main aim of the study is to determine experimentally derived look-up-tables and relevant values of the Hounsfield units based on the statistical correlation analysis. These values were obtained from the liver and kidneys of 24 mice in solutions of ethanol as the centroid value of the opacity look-up-table area under this graph. Samples and phantom were scanned by a Bruker SkyScan 1275 micro-CT and Phywe XR 4.0 and processed using CTvox and ORS Dragonfly software. To reconstruct the micro-CT projections, NRecon software was used. The main finding of the study is that there is a statistically significant relationship between the centroid of the area under the look-up-table curve and the number of days for which the animal sample was stored in an ethanol solution. H1 of the first hypothesis, i.e. that suggested the Spearman’s correlation coefficient does not equal zero (r₁ ≠ 0) regarding this relationship was confirmed. On the other hand, there is no statistically significant relationship between the centroid of the area under the look-up-table curve and the concentration of the ethanol solution. In this case, H1 of the second hypothesis, i.e. that the Spearman’s correlation coefficient does not equal zero (r₂ ≠ 0) regarding this relationship was not confirmed. Spearman’s correlation coefficients were −0.27 for the concentration and −0.87 for the number of days stored in ethanol solution in the case of the livers of 13 mice and 0.06 for the concentration and 0.94 for the number of days stored in ethanol solution in the case of kidneys of 11 mice. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

10 pages, 1916 KiB

Open AccessArticle

Detection of the Lateral Thermal Spread during Bipolar Vessel Sealing in an Ex Vivo Model—Preliminary Results

by Andreas Kirschbaum, Jan Jonas, Thomas M. Surowiec, Anika Pehl and Nikolas Mirow

Diagnostics 2022, 12(5), 1217; https://doi.org/10.3390/diagnostics12051217 - 12 May 2022

Cited by 1 | Viewed by 2845

Abstract

Background: As an unwanted side effect, lateral thermal expansion in bipolar tissue sealing may lead to collateral tissue damage. Materials and Methods: Our investigations were carried out on an ex vivo model of porcine carotid arteries. Lateral thermal expansion was measured and a [...] Read more.

Background: As an unwanted side effect, lateral thermal expansion in bipolar tissue sealing may lead to collateral tissue damage. Materials and Methods: Our investigations were carried out on an ex vivo model of porcine carotid arteries. Lateral thermal expansion was measured and a calculated index, based on thermographic recording and histologic examination, was designed to describe the risk of tissue damage. Results: For instrument 1, the mean extent of the critical zone > 50 °C was 2315 ± 509.2 µm above and 1700 ± 331.3 µm below the branches. The width of the necrosis zone was 412.5 ± 79.0 µm above and 426.7 ± 100.7µm below the branches. For instrument 2, the mean extent of the zone > 50 °C was 2032 ± 592.4 µm above and 1182 ± 386.9 µm below the branches. The width of the necrosis zone was 642.6 ± 158.2 µm above and 645.3 ± 111.9 µm below the branches. Our risk index indicated a low risk of damage for instrument 1 and a moderate to high risk for instrument 2. Conclusion: Thermography is a suitable method to estimate lateral heat propagation, and a validated risk index may lead to improved surgical handling. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

19 pages, 1702 KiB

Open AccessArticle

Radiomics Profiling Identifies the Value of CT Features for the Preoperative Evaluation of Lymph Node Metastasis in Papillary Thyroid Carcinoma

by Guoqiang Yang, Fan Yang, Fengyan Zhang, Xiaochun Wang, Yan Tan, Ying Qiao and Hui Zhang

Diagnostics 2022, 12(5), 1119; https://doi.org/10.3390/diagnostics12051119 - 29 Apr 2022

Cited by 6 | Viewed by 2432

Abstract

Background: The aim of this study was to identify the increased value of integrating computed tomography (CT) radiomics analysis with the radiologists’ diagnosis and clinical factors to preoperatively diagnose cervical lymph node metastasis (LNM) in papillary thyroid carcinoma (PTC) patients. Methods: A total [...] Read more.

Background: The aim of this study was to identify the increased value of integrating computed tomography (CT) radiomics analysis with the radiologists’ diagnosis and clinical factors to preoperatively diagnose cervical lymph node metastasis (LNM) in papillary thyroid carcinoma (PTC) patients. Methods: A total of 178 PTC patients were randomly divided into a training (n = 125) and a test cohort (n = 53) with a 7:3 ratio. A total of 2553 radiomic features were extracted from noncontrast, arterial contrast-enhanced and venous contrast-enhanced CT images of each patient. Principal component analysis (PCA) and Pearson’s correlation coefficient (PCC) were used for feature selection. Logistic regression was employed to build clinical–radiological, radiomics and combined models. A nomogram was developed by combining the radiomics features, CT-reported lymph node status and clinical factors. Results: The radiomics model showed a predictive performance similar to that of the clinical–radiological model, with similar areas under the curve (AUC) and accuracy (ACC). The combined model showed an optimal predictive performance in both the training (AUC, 0.868; ACC, 86.83%) and test cohorts (AUC, 0.878; ACC, 83.02%). Decision curve analysis demonstrated that the combined model has good clinical application value. Conclusions: Embedding CT radiomics into the clinical diagnostic process improved the diagnostic accuracy. The developed nomogram provides a potential noninvasive tool for LNM evaluation in PTC patients. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

15 pages, 2757 KiB

Open AccessArticle

Applying Taguchi Methodology to Optimize the Brain Image Quality of 128-Sliced CT: A Feasibility Study

by Hsien-Chun Tseng, Hung-Chih Lin, Yu-Che Tsai, Cheng-Hsun Lin, Sheng-Pin Changlai, Yueh-Chun Lee and Chien-Yi Chen

Appl. Sci. 2022, 12(9), 4378; https://doi.org/10.3390/app12094378 - 26 Apr 2022

Cited by 6 | Viewed by 2218

Abstract

Injuries due to traffic accidents have been significant causes of death in Taiwan and traffic accidents have been most common in recent years. Brain computed tomography (CT) examinations can improve imaging quality and increase the value of an imaging diagnosis. The image quality [...] Read more.

Injuries due to traffic accidents have been significant causes of death in Taiwan and traffic accidents have been most common in recent years. Brain computed tomography (CT) examinations can improve imaging quality and increase the value of an imaging diagnosis. The image quality of the brain gray/white matter was optimized using the Taguchi design with an indigenous polymethylmethacrylate (PMMA) slit gauge to imitate the adult brain and solid water phantoms. The two gauges without coating contrast media were located inside the center of a plate to simulate the brain and scanned to obtain images for further analysis. Five major parameters—CT slice thickness, milliampere-seconds, current voltage, filter type, and field of view—were optimized. Analysis of variance was used to determine individual interactions among all control parameters. The optimal experimental acquisition/settings were: slice thickness 2.5 mm, 300 mAs, 140 kVp, smooth filter, and FOV 200 mm². Signal-to-noise was improved by 106% (p < 0.001) over a routine examination. The effective dose (H_E) is approximately 1.33 mSv. Further clinical verification and the image quality of the ACR 464 head phantom is also discussed. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

9 pages, 1146 KiB

Open AccessArticle

Forrest Classification for Bleeding Peptic Ulcer: A New Look at the Old Endoscopic Classification

by Hsu-Heng Yen, Ping-Yu Wu, Tung-Lung Wu, Siou-Ping Huang, Yang-Yuan Chen, Mei-Fen Chen, Wen-Chen Lin, Cheng-Lun Tsai and Kang-Ping Lin

Diagnostics 2022, 12(5), 1066; https://doi.org/10.3390/diagnostics12051066 - 24 Apr 2022

Cited by 8 | Viewed by 7107

Abstract

The management of peptic ulcer bleeding is clinically challenging. For decades, the Forrest classification has been used for risk stratification for nonvariceal ulcer bleeding. The perception and interpretation of the Forrest classification vary among different endoscopists. The relationship between the bleeder and ulcer [...] Read more.

The management of peptic ulcer bleeding is clinically challenging. For decades, the Forrest classification has been used for risk stratification for nonvariceal ulcer bleeding. The perception and interpretation of the Forrest classification vary among different endoscopists. The relationship between the bleeder and ulcer images and the different stages of the Forrest classification has not been studied yet. Endoscopic still images of 276 patients with peptic ulcer bleeding for the past 3 years were retrieved and reviewed. The intra-rater agreement and inter-rater agreement were compared. The obtained endoscopic images were manually drawn to delineate the extent of the ulcer and bleeding area. The areas of the region of interest were compared between the different stages of the Forrest classification. A total of 276 images were first classified by two experienced tutor endoscopists. The images were reviewed by six other endoscopists. A good intra-rater correlation was observed (0.92–0.98). A good inter-rater correlation was observed among the different levels of experience (0.639–0.859). The correlation was higher among tutor and junior endoscopists than among experienced endoscopists. Low-risk Forrest IIC and III lesions show distinct patterns compared to high-risk Forrest I, IIA, or IIB lesions. We found good agreement of the Forrest classification among different endoscopists in a single institution. This is the first study to quantitively analyze the obtained and explain the distinct patterns of bleeding ulcers from endoscopy images. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

10 pages, 2640 KiB

Open AccessArticle

Evaluating the Cisplatin Dose Dependence of Testicular Dysfunction Using Creatine Chemical Exchange Saturation Transfer Imaging

by Reika Sawaya, Sohei Kuribayashi, Junpei Ueda and Shigeyoshi Saito

Diagnostics 2022, 12(5), 1046; https://doi.org/10.3390/diagnostics12051046 - 21 Apr 2022

Cited by 3 | Viewed by 1831

Abstract

Chemical exchange saturation transfer (CEST) imaging is a non-invasive molecular imaging technique for indirectly measuring low-concentration endogenous metabolites. Conventional CEST has low specificity, owing to the effects of spillover, magnetization transfer (MT), and T₁ relaxation, thus necessitating an inverse Z-spectrum analysis. We [...] Read more.

Chemical exchange saturation transfer (CEST) imaging is a non-invasive molecular imaging technique for indirectly measuring low-concentration endogenous metabolites. Conventional CEST has low specificity, owing to the effects of spillover, magnetization transfer (MT), and T₁ relaxation, thus necessitating an inverse Z-spectrum analysis. We aimed to investigate the usefulness of inverse Z-spectrum analysis in creatine (Cr)-CEST in mice, by conducting preclinical 7T-magnetic resonance imaging (MRI) and comparing the conventional analysis metric magnetization transfer ratio (MTR_conv) with the novel metric apparent exchange-dependent relaxation (AREX). We performed Cr-CEST imaging using 7T-MRI on mouse testes, using C57BL/6 mice as the control and a cisplatin-treated model. We prepared different doses of cisplatin to observe its dose dependence effect on testicular function. CEST imaging was obtained using an MT pulse with varying saturation frequencies, ranging from −4.8 ppm to +4.8 ppm. The application of control mouse testes improved the specificity of the CEST effect and image contrast between the testes and testicular epithelium. The cisplatin-treated model revealed impaired testicular function, and the Cr-CEST imaging displayed decreased Cr levels in the testes. There was a significant difference between the low- and high-dose models. The MTR values of Cr-CEST reflected the cisplatin dose dependence of testicular dysfunction. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

15 pages, 4356 KiB

Open AccessArticle

MRI-Based Radiomics Models to Discriminate Hepatocellular Carcinoma and Non-Hepatocellular Carcinoma in LR-M According to LI-RADS Version 2018

by Haiping Zhang, Dajing Guo, Huan Liu, Xiaojing He, Xiaofeng Qiao, Xinjie Liu, Yangyang Liu, Jun Zhou, Zhiming Zhou, Xi Liu and Zheng Fang

Diagnostics 2022, 12(5), 1043; https://doi.org/10.3390/diagnostics12051043 - 21 Apr 2022

Cited by 9 | Viewed by 2384

Abstract

Differentiating hepatocellular carcinoma (HCC) from other primary liver malignancies in the Liver Imaging Reporting and Data System (LI-RADS) M (LR-M) tumours noninvasively is critical for patient treatment options, but visual evaluation based on medical images is a very challenging task. This study aimed [...] Read more.

Differentiating hepatocellular carcinoma (HCC) from other primary liver malignancies in the Liver Imaging Reporting and Data System (LI-RADS) M (LR-M) tumours noninvasively is critical for patient treatment options, but visual evaluation based on medical images is a very challenging task. This study aimed to evaluate whether magnetic resonance imaging (MRI) models based on radiomics features could further improve the ability to classify LR-M tumour subtypes. A total of 102 liver tumours were defined as LR-M by two radiologists based on LI-RADS and were confirmed to be HCC (n = 31) and non-HCC (n = 71) by surgery. A radiomics signature was constructed based on reproducible features using the max-relevance and min-redundancy (mRMR) and least absolute shrinkage and selection operator (LASSO) logistic regression algorithms with tenfold cross-validation. Logistic regression modelling was applied to establish different models based on T2-weighted imaging (T2WI), arterial phase (AP), portal vein phase (PVP), and combined models. These models were verified independently in the validation cohort. The area under the curve (AUC) of the models based on T2WI, AP, PVP, T2WI + AP, T2WI + PVP, AP + PVP, and T2WI + AP + PVP were 0.768, 0.838, 0.778, 0.880, 0.818, 0.832, and 0.884, respectively. The combined model based on T2WI + AP + PVP showed the best performance in the training cohort and validation cohort. The discrimination efficiency of each radiomics model was significantly better than that of junior radiologists’ visual assessment (p < 0.05; Delong). Therefore, the MRI-based radiomics models had a good ability to discriminate between HCC and non-HCC in LR-M tumours, providing more options to improve the accuracy of LI-RADS classification. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

17 pages, 6990 KiB

Open AccessArticle

Detection of Chronic Blast-Related Mild Traumatic Brain Injury with Diffusion Tensor Imaging and Support Vector Machines

by Deborah L. Harrington, Po-Ya Hsu, Rebecca J. Theilmann, Annemarie Angeles-Quinto, Ashley Robb-Swan, Sharon Nichols, Tao Song, Lu Le, Carl Rimmele, Scott Matthews, Kate A. Yurgil, Angela Drake, Zhengwei Ji, Jian Guo, Chung-Kuan Cheng, Roland R. Lee, Dewleen G. Baker and Mingxiong Huang

Diagnostics 2022, 12(4), 987; https://doi.org/10.3390/diagnostics12040987 - 14 Apr 2022

Cited by 10 | Viewed by 3090

Abstract

Blast-related mild traumatic brain injury (bmTBI) often leads to long-term sequalae, but diagnostic approaches are lacking due to insufficient knowledge about the predominant pathophysiology. This study aimed to build a diagnostic model for future verification by applying machine-learning based support vector machine (SVM) [...] Read more.

Blast-related mild traumatic brain injury (bmTBI) often leads to long-term sequalae, but diagnostic approaches are lacking due to insufficient knowledge about the predominant pathophysiology. This study aimed to build a diagnostic model for future verification by applying machine-learning based support vector machine (SVM) modeling to diffusion tensor imaging (DTI) datasets to elucidate white-matter features that distinguish bmTBI from healthy controls (HC). Twenty subacute/chronic bmTBI and 19 HC combat-deployed personnel underwent DTI. Clinically relevant features for modeling were selected using tract-based analyses that identified group differences throughout white-matter tracts in five DTI metrics to elucidate the pathogenesis of injury. These features were then analyzed using SVM modeling with cross validation. Tract-based analyses revealed abnormally decreased radial diffusivity (RD), increased fractional anisotropy (FA) and axial/radial diffusivity ratio (AD/RD) in the bmTBI group, mostly in anterior tracts (29 features). SVM models showed that FA of the anterior/superior corona radiata and AD/RD of the corpus callosum and anterior limbs of the internal capsule (5 features) best distinguished bmTBI from HCs with 89% accuracy. This is the first application of SVM to identify prominent features of bmTBI solely based on DTI metrics in well-defined tracts, which if successfully validated could promote targeted treatment interventions. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

10 pages, 1939 KiB

Open AccessArticle

Sinus Plain Film Can Predict a Risky Distance from the Lacrimal Sac to the Anterior Skull Base: An Anatomic Study of Dacryocystorhinostomy

by Kuan-Chung Fang, Ren-Wen Ho, Sheng-Dean Luo, Wei-Che Lin, Ching-Nung Wu and Wei-Chih Chen

Diagnostics 2022, 12(4), 930; https://doi.org/10.3390/diagnostics12040930 - 8 Apr 2022

Viewed by 2268

Abstract

Background: Removal of the surrounding bone during dacryocystorhinostomy may present a higher risk of skull base injury in patients with frontal sinus aplasia. We used sinus plain films to predict cases with a greater risk of a reduced skull base distance in dacryocystorhinostomy. [...] Read more.

Background: Removal of the surrounding bone during dacryocystorhinostomy may present a higher risk of skull base injury in patients with frontal sinus aplasia. We used sinus plain films to predict cases with a greater risk of a reduced skull base distance in dacryocystorhinostomy. Methods: Sinus plain films and computed tomography data from patients were retrospectively evaluated. The frontal sinus was classified as normal, hypoplastic, or aplastic according to Waters’ view. Correlations of the frontal sinus roof-supraorbital margin (F-O) and the frontal sinus roof-nasion (F-N) distances on plain film with the closest lacrimal sac-anterior skull base (LS-ASB) distance measured on computed tomography images were assessed. Results: We evaluated 110 patients. In total, 16 (11.8%) patients had frontal sinus aplasia, of whom 6 (2.7%) had bilateral and 10 (9.1%) had unilateral aplasia. Sides with frontal sinus aplasia based on Waters’ view had a shorter median LS-ASB distance than normal or hypoplastic sides. The F-O and F-N distances in Waters’ view were significantly positively correlated with the computed tomographic LS-ASB distance. The F-O margin and F-N distance thresholds for predicting an LS-ASB distance < 10 mm, considered a risky distance, were 11.6 and 14.4 mm, respectively, with sensitivities of 100% and 91.7%, and specificities of 76% and 82.7%, respectively. Conclusions: The LS-ASB distance is closer on aplastic frontal sinus sides. Waters’ view on plain sinus films can provide a fast and inexpensive method for evaluating the skull base distance and sinonasal condition during planning for dacryocystorhinostomy. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

13 pages, 7480 KiB

Open AccessArticle

Removal of Specular Reflection Using Angle Adjustment of Linear Polarized Filter in Medical Imaging Diagnosis

by Kicheol Yoon, Jaehwang Seol and Kwang Gi Kim

Diagnostics 2022, 12(4), 863; https://doi.org/10.3390/diagnostics12040863 - 30 Mar 2022

Cited by 4 | Viewed by 2287

Abstract

The biggest problem in imaging medicine is the occurrence of light reflection in the imaging process for lesion diagnosis. The formation of light reflection obscures the diagnostic field of the lesion and interferes with the correct diagnosis of the observer. The existing method [...] Read more.

The biggest problem in imaging medicine is the occurrence of light reflection in the imaging process for lesion diagnosis. The formation of light reflection obscures the diagnostic field of the lesion and interferes with the correct diagnosis of the observer. The existing method has the inconvenience of performing a diagnosis in a state in which light reflection is suppressed by adjusting the direction angle of the camera. This paper proposes a method for rotating a linear polarization filter to remove light reflection in a diagnostic imaging camera. Vertical polarization and horizontal polarization are controlled through the rotation of the filter, and the polarization is adjusted to horizontal polarization. The rotation angle of the filter for horizontal polarization control will be 90°, and the vertical and horizontal polarization waves induce a 90° difference from each other. In this study, light reflection can be effectively removed during the imaging process, and light reflection removal can secure the field of view of the lesion. The removal of light reflection can help the observer’s accurate diagnosis, and these results are expected to be highly reliable and commercialized for direct application in the field of diagnostic medicine. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

12 pages, 479 KiB

Open AccessArticle

Heart Failure and Atrial Fibrillation: Diastolic Function Differences Depending on Left Ventricle Ejection Fraction

by Ruxandra-Nicoleta Horodinschi and Camelia Cristina Diaconu

Diagnostics 2022, 12(4), 839; https://doi.org/10.3390/diagnostics12040839 - 29 Mar 2022

Cited by 4 | Viewed by 3096

Abstract

Background: Heart failure (HF) and atrial fibrillation (AF) are prevalent cardiovascular diseases, and their association is common. Diastolic dysfunction may be present in patients with AF and all types of HF, leading to elevated intracardiac pressures. The objective of this study was [...] Read more.

Background: Heart failure (HF) and atrial fibrillation (AF) are prevalent cardiovascular diseases, and their association is common. Diastolic dysfunction may be present in patients with AF and all types of HF, leading to elevated intracardiac pressures. The objective of this study was to analyze diastolic dysfunction in patients with HF and AF depending on left ventricle ejection fraction (LVEF). Material and methods: This prospective study included 324 patients with chronic HF and AF (paroxysmal, persistent, or permanent) hospitalized between January 2018 and March 2021. The inclusion criteria were age older than 18 years, diagnosis of chronic HF and AF, and available echocardiographic data. The exclusion criteria were a suboptimal echocardiographic view, other cardiac rhythms than AF, congenital heart disease, or coronavirus 2 infection. Patients were divided into three subgroups according to LVEF: subgroup 1 included 203 patients with HF with reduced ejection fraction (HFrEF) and AF (62.65%), subgroup 2 included 42 patients with HF with mildly reduced ejection fraction (HFmrEF) and AF (12.96%), and subgroup 3 included 79 patients with HF with preserved ejection fraction (HFpEF) and AF (24.38%). We performed 2D transthoracic echocardiography in all patients. Statistical analysis was performed using R software. Results: The E/e′ ratio (p = 0.0352, OR 1.9) and left atrial volume index (56.4 mL/m² vs. 53.6 mL/m²) were higher in patients with HFrEF than in those with HFpEF. Conclusions: Patients with HFrEF and AF had more severe diastolic dysfunction and higher left ventricular filling pressures than those with HFpEF and AF. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

10 pages, 961 KiB

Open AccessArticle

Accuracy of Inflow Inversion Recovery (IFIR) for Upper Abdominal Arteries Evaluation: Comparison with Contrast-Enhanced MR and CTA

by Roberto Simonini, Pietro Andrea Bonaffini, Marco Porta, Cesare Maino, Francesco Saverio Carbone, Ludovico Dulcetta, Paolo Brambilla, Paolo Marra and Sandro Sironi

Diagnostics 2022, 12(4), 825; https://doi.org/10.3390/diagnostics12040825 - 28 Mar 2022

Cited by 2 | Viewed by 2512

Abstract

Background: Inflow-sensitive inversion recovery (IFIR) is a recently introduced technique to perform unenhanced magnetic resonance angiography (MRA). The purpose of our study is to determine the accuracy of IFIR-MRA in the evaluation of upper abdominal arteries, compared to standard MRA and computed tomography [...] Read more.

Background: Inflow-sensitive inversion recovery (IFIR) is a recently introduced technique to perform unenhanced magnetic resonance angiography (MRA). The purpose of our study is to determine the accuracy of IFIR-MRA in the evaluation of upper abdominal arteries, compared to standard MRA and computed tomography angiography (CTA). Materials and Methods: Seventy patients undergoing upper abdomen Magnetic Resonance Imaging (MRI) in different clinical settings were enrolled. The MRI protocol included an IFIR-MRA sequence that was intra-individually compared by using a qualitative 4-point scale in the same patients who underwent concomitant or close MRA (n = 65) and/or CTA (n = 44). Celiac trunk (CA), common-proper-left-right hepatic artery (C-P-L-R-HA), left gastric artery (LGA), gastroduodenal artery (GDA), splenic artery (SA), renal arteries (RA) and superior mesenteric artery (SMA) were assessed. Results: IFIR-MRA images were better rated in comparison with MRA. Particularly, all arteries obtained a statistically significant higher qualitative rating value (all p < 0.05). IFIR-MRA and MRA exhibited acceptable intraclass correlation coefficients (ICC) values for CA, C-L-R-HA, and SMA (ICC 0.507, 0.591, 0.615, 0.570, 0.525). IFIR-MRA and CTA showed significant correlations in C-P-L-R-HA (τ = 0.362, 0.261, 0.308, 0.307, respectively; p < 0.05), and in RA (τ = 0.279, p < 0.05). Conclusions: Compared to MRA, IFIR-MRA demonstrated a higher image quality in the majority of upper abdomen arterial vessels assessment. LHA and RHA branches could be better visualized with IFIR sequences, when visualizable. Based on these findings, we suggest to routinely integrate IFIR sequences in upper abdomen MRI studies. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

9 pages, 2227 KiB

Open AccessArticle

Quantifying and Statistically Modeling Residual Pneumoperitoneum after Robotic-Assisted Laparoscopic Prostatectomy: A Prospective, Single-Center, Observational Study

by Venkat M. Ramakrishnan, Tilo Niemann, Philipp Maletzki, Edward Guenther, Teodora Bujaroska, Olanrewaju Labulo, Zhufeng Li, Juliette Slieker, Rahel A. Kubik-Huch, Kurt Lehmann, Antonio Nocito and Lukas J. Hefermehl

Diagnostics 2022, 12(4), 785; https://doi.org/10.3390/diagnostics12040785 - 23 Mar 2022

Viewed by 2119

Abstract

Background: Laparoscopic surgery (LS) requires CO₂ insufflation to establish the operative field. Patients with worsening pain post-operatively often undergo computed tomography (CT). CT is highly sensitive in detecting free air—the hallmark sign of a bowel injury. Yet, the clinical significance of free [...] Read more.

Background: Laparoscopic surgery (LS) requires CO₂ insufflation to establish the operative field. Patients with worsening pain post-operatively often undergo computed tomography (CT). CT is highly sensitive in detecting free air—the hallmark sign of a bowel injury. Yet, the clinical significance of free air is often confounded by residual CO₂ and is not usually due to a visceral injury. The aim of this study was to attempt to quantify the residual pneumoperitoneum (RPP) after a robotic-assisted laparoscopic prostatectomy (RALP). Methods: We prospectively enrolled patients who underwent RALP between August 2018 and January 2020. CT scans were performed on postoperative days (POD) 3, 5, and 7. To investigate potential factors influencing the quantity of RPP, correlation plots were made against common variables. Results: In total, 31 patients with a mean age of 66 years (median 67, IQR 62–70.5) and mean BMI 26.59 (median 25.99, IQR: 24.06–29.24) underwent RALP during the study period. All patients had a relatively unremarkable post-operative course (30/31 with Clavien–Dindo class 0; 1/31 with class 2). After 3, 5, and 7 days, 3.2%, 6.4%, and 32.3% were completely without RPP, respectively. The mean RPP at 3 days was 37.6 mL (median 9.58 mL, max 247 mL, IQR 3.92–31.82 mL), whereas the mean RPP at 5 days was 19.85 mL (median 1.36 mL, max 220.77 mL, IQR 0.19–5.61 mL), and 7 days was 10.08 mL (median 0.09 mL, max 112.42 mL, IQR 0–1.5 mL). There was a significant correlation between RPP and obesity (p = 0.04665), in which higher BMIs resulted in lower initial insufflation volumes and lower RPP. Conclusions: This is the first study to systematically assess RPP after a standardized laparoscopic procedure using CT. Larger patients tend to have smaller residuals. Our data may help surgeons interpreting post-operative CTs in similar patient populations. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

2 pages, 9972 KiB

Open AccessInteresting Images

Left Nonrecurrent Laryngeal Nerve with Situs Inversus Totalis

by Yin-Yang Chen, Chi-You Liao and Chung-Chin Yao

Diagnostics 2022, 12(3), 730; https://doi.org/10.3390/diagnostics12030730 - 17 Mar 2022

Cited by 2 | Viewed by 2609

Abstract

The recurrent laryngeal nerve (RLN), a branch of the vagus nerve, supplies the motor and sensation function of the larynx. Generally, RLN detours around the right subclavian artery on the right side and the aortic arch on the left side. In a rare [...] Read more.

The recurrent laryngeal nerve (RLN), a branch of the vagus nerve, supplies the motor and sensation function of the larynx. Generally, RLN detours around the right subclavian artery on the right side and the aortic arch on the left side. In a rare anatomical variant, called nonrecurrent laryngeal nerve (NRLN), the nerve takes an aberrant path rather than descending into the thorax as usual. First reported in 1823, NRLN is a rare anomaly arising almost exclusively on the right side, reported in 0.3–0.8% of people, and associated with vascular anomalies of embryonic aortic arch development. The atypical vascular pattern of aberrant subclavian artery (arteria lusoria) running behind the trachea and esophagus allows the vagus nerve to pass freely, which then directly branches out as NRLN at the level of the larynx. On the other hand, cases of left NRLN, only reported in 0.004% of people, are all accompanied by significant pathologies such as situs inversus totalis with opposite vascular pattern of left aberrant subclavian artery. This rare anatomical variation is clinically important, as NLRN is a major risk factor for iatrogenic injury during thyroidectomy, parathyroidectomy, and other invasive procedures in the head and neck region. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

15 pages, 1786 KiB

Open AccessArticle

A Predictive Model for the Risk of Posterior Circulation Stroke in Patients with Intracranial Atherosclerosis Based on High Resolution MRI

by Zhenxing Liu, Feiyang Zhong, Yu Xie, Xuanzhen Lu, Botong Hou, Keni Ouyang, Jiabin Fang, Meiyan Liao and Yumin Liu

Diagnostics 2022, 12(4), 812; https://doi.org/10.3390/diagnostics12040812 - 15 Mar 2022

Cited by 5 | Viewed by 2747 | Correction

Abstract

Intracranial vertebrobasilar atherosclerosis is the main cause of posterior circulation ischemic stroke. We aimed to construct a predictive model for the risk of posterior circulation ischemic stroke in patients with posterior circulation atherosclerosis based on high-resolution MRI (HR-MRI). A total of 208 consecutive [...] Read more.

Intracranial vertebrobasilar atherosclerosis is the main cause of posterior circulation ischemic stroke. We aimed to construct a predictive model for the risk of posterior circulation ischemic stroke in patients with posterior circulation atherosclerosis based on high-resolution MRI (HR-MRI). A total of 208 consecutive patients with posterior circulation atherosclerosis confirmed by HR-MRI, from January 2020 to July 2021, were retrospectively assessed. They were assigned to the posterior circulation stroke (59 patients) and non-posterior circulation stroke group (149 patients) based on clinical presentation and diffusion-weighted imaging (DWI). Demographic data, risk factors of atherosclerosis, laboratory findings, and imaging characteristics were extracted from electronic health records. Plaque features were investigated by HR-MRI. Fifty-three clinical or imaging features were used to derive the model. Multivariable logistic regression analysis was employed to construct the prediction model. The nomogram was evaluated for calibration, differentiation, and clinical usefulness. Plaque enhancement, plaque irregular surface morphology, artery location of plaque, and dorsal quadrant of plaque location were significant predictors for posterior circulation stroke in patients with intracranial atherosclerosis. Subsequently, these variables were selected to establish a nomogram. The model showed good distinction (C-index 0.830, 95% CI 0.766-0.895). The calibration curve also showed excellent consistency between the prediction of the nomogram and the observed curve. Decision curve analysis further demonstrated that the nomogram conferred significantly high clinical net benefit. The nomogram calculated from plaque characteristics in HR-MRI may accurately predict the posterior circulation stroke occurrence and be of great help for stratification of stroke decision making. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

12 pages, 14457 KiB

Open AccessArticle

An Empirical Evaluation of Convolutional Networks for Malaria Diagnosis

by Andrea Loddo, Corrado Fadda and Cecilia Di Ruberto

J. Imaging 2022, 8(3), 66; https://doi.org/10.3390/jimaging8030066 - 7 Mar 2022

Cited by 24 | Viewed by 3766

Abstract

Malaria is a globally widespread disease caused by parasitic protozoa transmitted to humans by infected female mosquitoes of Anopheles. It is caused in humans only by the parasite Plasmodium, further classified into four different species. Identifying malaria parasites is possible by analysing digital [...] Read more.

Malaria is a globally widespread disease caused by parasitic protozoa transmitted to humans by infected female mosquitoes of Anopheles. It is caused in humans only by the parasite Plasmodium, further classified into four different species. Identifying malaria parasites is possible by analysing digital microscopic blood smears, which is tedious, time-consuming and error prone. So, automation of the process has assumed great importance as it helps the laborious manual process of review and diagnosis. This work focuses on deep learning-based models, by comparing off-the-shelf architectures for classifying healthy and parasite-affected cells, by investigating the four-class classification on the Plasmodium falciparum stages of life and, finally, by evaluating the robustness of the models with cross-dataset experiments on two different datasets. The main contributions to the research in this field can be resumed as follows: (i) comparing off-the-shelf architectures in the task of classifying healthy and parasite-affected cells, (ii) investigating the four-class classification on the P. falciparum stages of life and (iii) evaluating the robustness of the models with cross-dataset experiments. Eleven well-known convolutional neural networks on two public datasets have been exploited. The results show that the networks have great accuracy in binary classification, even though they lack few samples per class. Moreover, the cross-dataset experiments exhibit the need for some further regulations. In particular, ResNet-18 achieved up to 97.68% accuracy in the binary classification, while DenseNet-201 reached 99.40% accuracy on the multiclass classification. The cross-dataset experiments exhibit the limitations of deep learning approaches in such a scenario, even though combining the two datasets permitted DenseNet-201 to reach 97.45% accuracy. Naturally, this needs further investigation to improve the robustness. In general, DenseNet-201 seems to offer the most stable and robust performance, offering as a crucial candidate to further developments and modifications. Moreover, the mobile-oriented architectures showed promising and satisfactory performance in the classification of malaria parasites. The obtained results enable extensive improvements, specifically oriented to the application of object detectors for type and stage of life recognition, even in mobile environments. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

11 pages, 4114 KiB

Open AccessArticle

Quantification of SPECT Concentric Ring Artifacts by Radiomics and Radial Features

by Emilio Mezzenga, Anna Sarnelli, Giovanni Bellomo, Frank P. DiFilippo, Christopher J. Palestro and Kenneth J. Nichols

Appl. Sci. 2022, 12(5), 2726; https://doi.org/10.3390/app12052726 - 6 Mar 2022

Cited by 2 | Viewed by 3771

Abstract

(1) Background: Concentric ring artifacts in reconstructed SPECT images indicate the presence of detector non-uniformity in gamma camera systems. The identification of these artifacts is generally visual and not quantitative. The aim of our study was to evaluate observer assessments of the presence [...] Read more.

(1) Background: Concentric ring artifacts in reconstructed SPECT images indicate the presence of detector non-uniformity in gamma camera systems. The identification of these artifacts is generally visual and not quantitative. The aim of our study was to evaluate observer assessments of the presence of concentric rings in reconstructed SPECT phantom images and to verify whether quantitative texture analysis can detect such artifacts, which are detrimental to accurate tumor detection. (2) Methods: Test data were acquired as part of the quarterly quality assurance program using a standardized SPECT phantom containing solid spheres, solid rods, and a water solution of ^99mTc. Forty separate SPECT acquisitions were analyzed to assess the presence of ring artifacts. Two experienced medical physicists independently reviewed transaxial images and graded the severity of artifacts on a five-point scale. Quantitative radiomic features were computed for volumes of interest located in the uniform phantom section. In addition to these, radial contrast (

R_{C o n t r a s t}

) and radial root-mean-square contrast (

R_{R M S C}

) were also calculated and derived from the radial profile of summed slices transformed into polar coordinates. (3) Results: Artifacts were considered sufficiently severe to warrant camera re-tuning in 10 rod sections, 17 sphere sections, and 16 uniform sections. In the uniform sections, there was “good agreement” for inter-observer and intra-rater assessments (κ = 0.66, Fisher exact p < 0.0001 and κ = 0.61, and Fisher exact p = 0.001, respectively). The two radial features agreed significantly (p < 0.001) with visual severity judgment of ring artifacts in uniform sections and were selected as informative about the presence of ring artifacts by LASSO approach. The increased magnitude of

R_{C o n t r a s t}

and

R_{R M S C}

correlated significantly with increasingly severe artifact scores (ρ = 0.65–0.66, p < 0.0001). (4) Conclusions: There was good agreement between the physicists with respect to the presence of circular ring artifacts in uniform sections of SPECT quality assurance scans, with the artifacts accurately detected by radial contrast and noise-to-signal ratio measurements. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

13 pages, 2408 KiB

Open AccessArticle

Semi-Automatic MRI Feature Assessment in Small- and Medium-Volume Benign Prostatic Hyperplasia after Prostatic Artery Embolization

by Vanessa F. Schmidt, Mirjam Schirren, Maurice M. Heimer, Philipp M. Kazmierczak, Clemens C. Cyran, Moritz Wildgruber, Max Seidensticker, Jens Ricke and Olga Solyanik

Diagnostics 2022, 12(3), 585; https://doi.org/10.3390/diagnostics12030585 - 25 Feb 2022

Cited by 2 | Viewed by 2266

Abstract

(1) Background: To assess the treatment response of benign prostatic syndrome (BPS) following prostatic artery embolization (PAE) using a semi-automatic software analysis of magnetic resonance imaging (MRI) features and clinical indexes. (2) Methods: Prospective, monocenter study of MRI and clinical data of n [...] Read more.

(1) Background: To assess the treatment response of benign prostatic syndrome (BPS) following prostatic artery embolization (PAE) using a semi-automatic software analysis of magnetic resonance imaging (MRI) features and clinical indexes. (2) Methods: Prospective, monocenter study of MRI and clinical data of n = 27 patients with symptomatic BPS before and (1, 6, 12 months) after PAE. MRI analysis was performed using a dedicated semi-automatic software for segmentation of the central and the total gland (CG, TG), respectively; signal intensities (SIs) of T1-weighted (T1w), T2-weighted (T2w), and diffusion-weighted images (DWI), as well as intravesical prostatic protrusion (IPP) and prostatic volumes (CGV, TGV), were evaluated at each time point. The semi-automatic assessed TGV was compared to conventional TGV by an ellipse formula. International prostate symptom score (IPSS) and international consultation on incontinence questionnaire–urinary incontinence short form (ICIQ-UI SF) questionnaires were used as clinical indexes. Statistical testing in the form of ANOVA, pairwise comparisons using Bonferroni correction, and multiple linear correlations, were conducted using SPSS. (3) Results: TGV was significantly reduced one, six, and 12 months after PAE as assessed by the semi-automatic approach and conventional ellipse formula (p = 0.005; p = 0.025). CGV significantly decreased after one month (p = 0.038), but showed no significant differences six and 12 months after PAE (p = 0.191; p = 0.283). IPP at baseline was demonstrated by 25/27 patients (92.6%) with a significant decrease one, six, and 12 months after treatment (p = 0.028; p = 0.010; p = 0.008). Significant improvement in IPSS and ICIQ-UI SF (p = 0.002; p = 0.016) after one month correlated moderately with TGV reduction (p = 0.031; p = 0.05, correlation coefficients 0.52; 0.69). Apparent diffusion coefficient (ADC) values of CG significantly decreased one month after embolization (p < 0.001), while there were no significant differences in T1w and T2w SIs before and after treatment at each time point. (4) Conclusions: The semi-automatic approach is appropriate for the assessment of volumetric and morphological changes in prostate MRI following PAE, able to identify significantly different ADC values post-treatment without the need for manual identification of infarct areas. Semi-automatic measured TGV reduction is significant and comparable to the TGV calculated by the conventional ellipse formula, confirming the clinical response after PAE. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

13 pages, 6980 KiB

Open AccessArticle

Considerations on Baseline Generation for Imaging AI Studies Illustrated on the CT-Based Prediction of Empyema and Outcome Assessment

by Raphael Sexauer, Bram Stieltjes, Jens Bremerich, Tugba Akinci D’Antonoli and Noemi Schmidt

J. Imaging 2022, 8(3), 50; https://doi.org/10.3390/jimaging8030050 - 22 Feb 2022

Cited by 1 | Viewed by 3383

Abstract

For AI-based classification tasks in computed tomography (CT), a reference standard for evaluating the clinical diagnostic accuracy of individual classes is essential. To enable the implementation of an AI tool in clinical practice, the raw data should be drawn from clinical routine data [...] Read more.

For AI-based classification tasks in computed tomography (CT), a reference standard for evaluating the clinical diagnostic accuracy of individual classes is essential. To enable the implementation of an AI tool in clinical practice, the raw data should be drawn from clinical routine data using state-of-the-art scanners, evaluated in a blinded manner and verified with a reference test. Three hundred and thirty-five consecutive CTs, performed between 1 January 2016 and 1 January 2021 with reported pleural effusion and pathology reports from thoracocentesis or biopsy within 7 days of the CT were retrospectively included. Two radiologists (4 and 10 PGY) blindly assessed the chest CTs for pleural CT features. If needed, consensus was achieved using an experienced radiologist’s opinion (29 PGY). In addition, diagnoses were extracted from written radiological reports. We analyzed these findings for a possible correlation with the following patient outcomes: mortality and median hospital stay. For AI prediction, we used an approach consisting of nnU-Net segmentation, PyRadiomics features and a random forest model. Specificity and sensitivity for CT-based detection of empyema (n = 81 of n = 335 patients) were 90.94 (95%-CI: 86.55–94.05) and 72.84 (95%-CI: 61.63–81.85%) in all effusions, with moderate to almost perfect interrater agreement for all pleural findings associated with empyema (Cohen’s kappa = 0.41–0.82). Highest accuracies were found for pleural enhancement or thickening with 87.02% and 81.49%, respectively. For empyema prediction, AI achieved a specificity and sensitivity of 74.41% (95% CI: 68.50–79.57) and 77.78% (95% CI: 66.91–85.96), respectively. Empyema was associated with a longer hospital stay (median = 20 versus 14 days), and findings consistent with pleural carcinomatosis impacted mortality. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

11 pages, 2184 KiB

Open AccessArticle

Terahertz Imaging for Formalin Fixed Malignant Liver Tumors Using Two-Band Beamline at the Accelerator Facility of Nihon University

by Yusuke Kawashima, Suemitsu Masaaki, Kayo Kuyama, Takeshi Sakai, Yasushi Hayakawa, Takashi Kaneda and Norihiro Sei

Appl. Sci. 2022, 12(4), 2229; https://doi.org/10.3390/app12042229 - 21 Feb 2022

Cited by 5 | Viewed by 2543

Abstract

We investigated the transmission characteristics of formalin fixed human liver samples in which normal liver tissue and malignant liver tumor were mixed using terahertz (THz) coherent synchrotron radiation at an infrared free-electron laser (FEL) facility at Nihon University. Infrared-FEL imaging has indicated that [...] Read more.

We investigated the transmission characteristics of formalin fixed human liver samples in which normal liver tissue and malignant liver tumor were mixed using terahertz (THz) coherent synchrotron radiation at an infrared free-electron laser (FEL) facility at Nihon University. Infrared-FEL imaging has indicated that the amount of water molecules in the tumor tissue is not different from that in the normal tissue. However, the transmission of the incipient tumor tissue was lower than that of the normal tissue in THz imaging because the tumor tissue contained more water molecular clusters than the normal tissue. The tumor tissue became more permeable owing to the development of fibrous tissue around it. THz imaging will be more useful for discriminating liver tissues by increasing the spatial resolution. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

19 pages, 1473 KiB

Open AccessSystematic Review

Evaluation of Diagnostic Performance of Automatic Breast Volume Scanner Compared to Handheld Ultrasound on Different Breast Lesions: A Systematic Review

by Shahad A. Ibraheem, Rozi Mahmud, Suraini Mohamad Saini, Hasyma Abu Hassan, Aysar Sabah Keiteb and Ahmed M. Dirie

Diagnostics 2022, 12(2), 541; https://doi.org/10.3390/diagnostics12020541 - 19 Feb 2022

Cited by 7 | Viewed by 2668

Abstract

Objective: To compare the diagnostic performance of the automatic breast volume scanner (ABVS) against the handheld ultrasound (HHUS) in the differential diagnosis of benign and malignant breast lesions. Methods: A systematic search and review of studies involving ABVS and HHUS for breast cancer [...] Read more.

Objective: To compare the diagnostic performance of the automatic breast volume scanner (ABVS) against the handheld ultrasound (HHUS) in the differential diagnosis of benign and malignant breast lesions. Methods: A systematic search and review of studies involving ABVS and HHUS for breast cancer screening were performed. The search involved the data taken from Scopus, PubMed, and science direct databases and was conducted between the year 2011 to 2020. The prospective method was used in determining the inclusion and exclusion criteria while the evidence level was determined using the BI-RADS categories for diagnostic studies. In addition, the parameters of specificity, mean age, sensitivity, tumor number, and diagnostic accuracy of the ABVS and HHUS were summarized. Results: No systematic review or randomized controlled trial were identified in the systematic search while one cross-sectional study, eight retrospective studies, and 10 prospective studies were found. Sufficient follow-up of the subjects with benign and malignant findings were made only in 10 studies, in which only two had used ABVS and HHUS after performing mammographic screening and MRI. Analysis was made of 21 studies, which included 5448 lesions (4074 benign and 1374 malignant) taken from 6009 patients. The range of sensitivity was (0.72–1.0) for ABVS and (0.62–1.0) for HHUS; the specificity range was (0.52–0.98)% for ABVS and (0.49–0.99)% for HHUS. The accuracy range among the 11 studies was (80–99)% and (59–98)% for the HHUS and ABVS, respectively. The identified tumors had a mean size of 2.1 cm, and the detected cancers had a mean percentage of 94% (81–100)% in comparison to the non-cancer in all studies. Conclusions: The evidence available in the literature points to the fact that the diagnostic performance of both ABVS and HHUS are similar with reference to the differentiation of malignant and benign breast lesions. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

15 pages, 8865 KiB

Open AccessArticle

Automated Structural Analysis and Quantitative Characterization of Scar Tissue Using Machine Learning

by Luluil Maknuna, Hyeonsoo Kim, Yeachan Lee, Yoonjin Choi, Hyunjung Kim, Myunggi Yi and Hyun Wook Kang

Diagnostics 2022, 12(2), 534; https://doi.org/10.3390/diagnostics12020534 - 19 Feb 2022

Cited by 7 | Viewed by 4816

Abstract

An analysis of scar tissue is necessary to understand the pathological tissue conditions during or after the wound healing process. Hematoxylin and eosin (HE) staining has conventionally been applied to understand the morphology of scar tissue. However, the scar lesions cannot be analyzed [...] Read more.

An analysis of scar tissue is necessary to understand the pathological tissue conditions during or after the wound healing process. Hematoxylin and eosin (HE) staining has conventionally been applied to understand the morphology of scar tissue. However, the scar lesions cannot be analyzed from a whole slide image. The current study aimed to develop a method for the rapid and automatic characterization of scar lesions in HE-stained scar tissues using a supervised and unsupervised learning algorithm. The supervised learning used a Mask region-based convolutional neural network (RCNN) to train a pattern from a data representation using MMDetection tools. The K-means algorithm characterized the HE-stained tissue and extracted the main features, such as the collagen density and directional variance of the collagen. The Mask RCNN model effectively predicted scar images using various backbone networks (e.g., ResNet50, ResNet101, ResNeSt50, and ResNeSt101) with high accuracy. The K-means clustering method successfully characterized the HE-stained tissue by separating the main features in terms of the collagen fiber and dermal mature components, namely, the glands, hair follicles, and nuclei. A quantitative analysis of the scar tissue in terms of the collagen density and directional variance of the collagen confirmed 50% differences between the normal and scar tissues. The proposed methods were utilized to characterize the pathological features of scar tissue for an objective histological analysis. The trained model is time-efficient when used for detection in place of a manual analysis. Machine learning-assisted analysis is expected to aid in understanding scar conditions, and to help establish an optimal treatment plan. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

11 pages, 2111 KiB

Open AccessArticle

Relationship between Apparent Diffusion Coefficient Distribution and Cancer Grade in Prostate Cancer and Benign Prostatic Hyperplasia

by Shigeyoshi Saito, Yoshihiro Koyama, Junpei Ueda and Takashi Hashido

Diagnostics 2022, 12(2), 525; https://doi.org/10.3390/diagnostics12020525 - 18 Feb 2022

Cited by 7 | Viewed by 2143

Abstract

The aim of this paper was to assess the associations between prostate cancer aggressiveness and histogram-derived apparent diffusion coefficient (ADC) parameters and determine which ADC parameters may help distinguish among stromal hyperplasia (SH), glandular hyperplasia (GH), and low-grade, intermediate-grade, and high-grade prostate cancers. [...] Read more.

The aim of this paper was to assess the associations between prostate cancer aggressiveness and histogram-derived apparent diffusion coefficient (ADC) parameters and determine which ADC parameters may help distinguish among stromal hyperplasia (SH), glandular hyperplasia (GH), and low-grade, intermediate-grade, and high-grade prostate cancers. The mean, median, minimum, maximum, and 10th and 25th percentile ADC values were determined from the ADC histogram and compared among two benign prostate hyperplasia (BPH) groups and three Gleason score (GS) groups. Seventy lesions were identified in 58 patients who had undergone proctectomy. Thirty-nine lesions were prostate cancers (GS 6 = 7 lesions, GS 7 = 19 lesions, GS 8 = 11 lesions, GS 9 = 2 lesions), and thirty-one lesions were BPH (SH = 15 lesions, GH = 16 lesions). There were statistically significant differences in 10th percentile and 25th percentile ADC values when comparing GS 6 to GS 7 (p < 0.05). The 10th percentile ADC values yielded the highest area under the curve (AUC). Tenth and 25th percentile ADCs can be used to more accurately differentiate lesions with GS 6 from those with GS 7 than other ADC parameters. Our data indicate that the major challenge with ADC mapping is to differentiate between SH and GS 6, and SH and GS 7. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

17 pages, 956 KiB

Open AccessArticle

COMMA: Propagating Complementary Multi-Level Aggregation Network for Polyp Segmentation

by Wooseok Shin, Min Seok Lee and Sung Won Han

Appl. Sci. 2022, 12(4), 2114; https://doi.org/10.3390/app12042114 - 17 Feb 2022

Cited by 3 | Viewed by 2356

Abstract

Colonoscopy is an effective method for detecting polyps to prevent colon cancer. Existing studies have achieved satisfactory polyp detection performance by aggregating low-level boundary and high-level region information in convolutional neural networks (CNNs) for precise polyp segmentation in colonoscopy images. However, multi-level aggregation [...] Read more.

Colonoscopy is an effective method for detecting polyps to prevent colon cancer. Existing studies have achieved satisfactory polyp detection performance by aggregating low-level boundary and high-level region information in convolutional neural networks (CNNs) for precise polyp segmentation in colonoscopy images. However, multi-level aggregation provides limited polyp segmentation owing to the distribution discrepancy that occurs when integrating different layer representations. To address this problem, previous studies have employed complementary low- and high- level representations. In contrast to existing methods, we focus on propagating complementary information such that the complementary low-level explicit boundary with abstracted high-level representations diminishes the discrepancy. This study proposes COMMA, which propagates complementary multi-level aggregation to reduce distribution discrepancies. COMMA comprises a complementary masking module (CMM) and a boundary propagation module (BPM) as a multi-decoder. The CMM masks the low-level boundary noises through the abstracted high-level representation and leverages the masked information at both levels. Similarly, the BPM incorporates the lowest- and highest-level representations to obtain explicit boundary information and propagates the boundary to the CMMs to improve polyp detection. CMMs can discriminate polyps more elaborately than prior CMMs based on boundary and complementary representations. Moreover, we propose a hybrid loss function to mitigate class imbalance and noisy annotations in polyp segmentation. To evaluate the COMMA performance, we conducted experiments on five benchmark datasets using five metrics. The results proved that the proposed network outperforms state-of-the-art methods in terms of all datasets. Specifically, COMMA improved mIoU performance by 0.043 on average for all datasets compared to the existing state-of-the-art methods. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

13 pages, 1359 KiB

Open AccessArticle

Non-Contrast-Enhanced and Contrast-Enhanced Magnetic Resonance Angiography in Living Donor Liver Vascular Anatomy

by Chien-Chang Liao, Meng-Hsiang Chen, Chun-Yen Yu, Leung-Chit Leo Tsang, Chao-Long Chen, Hsien-Wen Hsu, Wei-Xiong Lim, Yi-Hsuan Chuang, Po-Hsun Huang, Yu-Fan Cheng and Hsin-You Ou

Diagnostics 2022, 12(2), 498; https://doi.org/10.3390/diagnostics12020498 - 15 Feb 2022

Cited by 5 | Viewed by 4251

Abstract

Background: Since the advent of a new generation of inflow-sensitive inversion recovery (IFIR) technology, three-dimensional non-contrast-enhanced magnetic resonance angiography is being used to obtain hepatic vessel images without applying gadolinium contrast agent. The purpose of this study was to explore the diagnostic efficacy [...] Read more.

Background: Since the advent of a new generation of inflow-sensitive inversion recovery (IFIR) technology, three-dimensional non-contrast-enhanced magnetic resonance angiography is being used to obtain hepatic vessel images without applying gadolinium contrast agent. The purpose of this study was to explore the diagnostic efficacy of non-contrast-enhanced magnetic resonance angiography (non-CE MRA), contrast-enhanced magnetic resonance angiography (CMRA), and computed tomography angiography (CTA) in the preoperative evaluation of living liver donors. Methods: A total of 43 liver donor candidates who were evaluated for living donor liver transplantation completed examinations. Donors’ age, gender, renal function (eGFR), and previous CTA and imaging were recorded before non-CE MRA and CMRA. CTA images were used as the standard. Results: Five different classifications of hepatic artery patterns (types I, III, V, VI, VIII) and three different classifications of portal vein patterns (types I, II, and III) were identified among 43 candidates. The pretransplant vascular anatomy was well identified using combined non-CE MRA and CMRA of hepatic arteries (100%), PVs (98%), and hepatic veins (100%) compared with CTA images. Non-CE MRA images had significantly stronger contrast signal intensity of portal veins (p < 0.01) and hepatic veins (p < 0.01) than CMRA. No differences were found in signal intensity of the hepatic artery between non-CE MRA and CMRA. Conclusion: Combined non-CE MRA and CMRA demonstrate comparable diagnostic ability to CTA and provide enhanced biliary anatomy information that assures optimum donor safety. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

13 pages, 1320 KiB

Open AccessSystematic Review

Lanthanum Carbonate Opacities—A Systematic Review

by Jan Kampmann, Nina Pirschel Hansen, Anders Nikolai Ørsted Schultz, Andreas Hjelm Brandt and Frans Brandt

Diagnostics 2022, 12(2), 464; https://doi.org/10.3390/diagnostics12020464 - 11 Feb 2022

Cited by 6 | Viewed by 2617

Abstract

Background: Lanthanum carbonate is a phosphate binder used in advanced kidney disease. Its radiopaque appearance has been described in many case studies and case series. Misinterpretation of this phenomenon leads to unnecessary diagnostic tests and procedures. The objectives of this study were to [...] Read more.

Background: Lanthanum carbonate is a phosphate binder used in advanced kidney disease. Its radiopaque appearance has been described in many case studies and case series. Misinterpretation of this phenomenon leads to unnecessary diagnostic tests and procedures. The objectives of this study were to summarize the literature on lanthanum carbonate opacities and present a visual overview. Methods: A systematic search was conducted using MEDLINE, Embase, and Web of Science. We included all types of studies, including case reports/studies, describing radiological findings of lanthanum carbonate opacities in patients with chronic kidney disease. No filter for time was set. Results: A total of 36 articles were eligible for data extraction, and 33 articles were included in the narrative synthesis. Lanthanum carbonate opacities were most commonly reported in the intestines (26 studies, 73%), stomach (8 studies, 21%), and the aerodigestive tract (2 studies, 6%). The opacities in the intestine were most frequently described as multiple, scattered radiopaque densities, compared with the aerodigestive tract, where the opacities were described as a single, round foreign body. Suspicion of contrast medium or foreign bodies was the most common differential diagnosis. LC opacities in patients with CKD are commonly misinterpreted as foreign bodies or suspect contrast media. Conclusions: CKD patients treated with LC may have opacities throughout the digestive tract that can vary in appearance. Stopping LC treatment or changing to an alternative phosphate binder prior to planned image studies can avoid diagnostic confusion. If this is not an option, knowledge of the presentation of LC opacities is important. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

27 pages, 6237 KiB

Open AccessArticle

Eye Tracking-Based Diagnosis and Early Detection of Autism Spectrum Disorder Using Machine Learning and Deep Learning Techniques

by Ibrahim Abdulrab Ahmed, Ebrahim Mohammed Senan, Taha H. Rassem, Mohammed A. H. Ali, Hamzeh Salameh Ahmad Shatnawi, Salwa Mutahar Alwazer and Mohammed Alshahrani

Electronics 2022, 11(4), 530; https://doi.org/10.3390/electronics11040530 - 10 Feb 2022

Cited by 93 | Viewed by 14248

Abstract

Eye tracking is a useful technique for detecting autism spectrum disorder (ASD). One of the most important aspects of good learning is the ability to have atypical visual attention. The eye-tracking technique provides useful information about children’s visual behaviour for early and accurate [...] Read more.

Eye tracking is a useful technique for detecting autism spectrum disorder (ASD). One of the most important aspects of good learning is the ability to have atypical visual attention. The eye-tracking technique provides useful information about children’s visual behaviour for early and accurate diagnosis. It works by scanning the paths of the eyes to extract a sequence of eye projection points on the image to analyse the behaviour of children with autism. In this study, three artificial-intelligence techniques were developed, namely, machine learning, deep learning, and a hybrid technique between them, for early diagnosis of autism. The first technique, neural networks [feedforward neural networks (FFNNs) and artificial neural networks (ANNs)], is based on feature classification extracted by a hybrid method between local binary pattern (LBP) and grey level co-occurrence matrix (GLCM) algorithms. This technique achieved a high accuracy of 99.8% for FFNNs and ANNs. The second technique used a pre-trained convolutional neural network (CNN) model, such as GoogleNet and ResNet-18, on the basis of deep feature map extraction. The GoogleNet and ResNet-18 models achieved high performances of 93.6% and 97.6%, respectively. The third technique used the hybrid method between deep learning (GoogleNet and ResNet-18) and machine learning (SVM), called GoogleNet + SVM and ResNet-18 + SVM. This technique depends on two blocks. The first block used CNN to extract deep feature maps, whilst the second block used SVM to classify the features extracted from the first block. This technique proved its high diagnostic ability, achieving accuracies of 95.5% and 94.5% for GoogleNet + SVM and ResNet-18 + SVM, respectively. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

3 pages, 846 KiB

Open AccessInteresting Images

Hemorrhagic Transformation after Intravenous Tissue Plasminogen Activator Administration in Acute Distal Middle Cerebral Artery Occlusion

by Chan-Hyuk Lee, Sang Hak Yi, Byoung-Soo Shin and Hyun Goo Kang

Diagnostics 2022, 12(2), 398; https://doi.org/10.3390/diagnostics12020398 - 3 Feb 2022

Viewed by 2408

Abstract

Atrial fibrillation and cerebral embolism are known to increase the risk of hemorrhagic transformation (HT). In addition, a sufficient number of collateral vessels in acute ischemic stroke can maintain the ischemic penumbra and prevent progression to the ischemic core, while an insufficient number [...] Read more.

Atrial fibrillation and cerebral embolism are known to increase the risk of hemorrhagic transformation (HT). In addition, a sufficient number of collateral vessels in acute ischemic stroke can maintain the ischemic penumbra and prevent progression to the ischemic core, while an insufficient number of collateral vessels increase the HT risk after therapeutic recanalization. In this case, when the middle cerebral artery is recanalized, reperfusion injury may occur in the basal ganglia due to insufficient collateral vessels. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

8 pages, 1382 KiB

Open AccessArticle

Impact of Different Metal Artifact Reduction Techniques on Attenuation Correction of Normal Organs in ¹⁸F-FDG-PET/CT

by Janna Morawitz, Ole Martin, Johannes Boos, Lino M. Sawicki, Katrin Wingendorf, Martin Sedlmair, Eduards Mamlins, Christina Antke, Gerald Antoch and Benedikt M. Schaarschmidt

Diagnostics 2022, 12(2), 375; https://doi.org/10.3390/diagnostics12020375 - 1 Feb 2022

Cited by 1 | Viewed by 2005

Abstract

Purpose: To evaluate the impact of different metal artifact reduction algorithms on Hounsfield units (HU) and the standardized uptake value (SUV) in normal organs in patients with different metal implants. Methods: This study prospectively included 66 patients (mean age of 66.02 ± 13.1 [...] Read more.

Purpose: To evaluate the impact of different metal artifact reduction algorithms on Hounsfield units (HU) and the standardized uptake value (SUV) in normal organs in patients with different metal implants. Methods: This study prospectively included 66 patients (mean age of 66.02 ± 13.1 years) with 87 different metal implants. CT image reconstructions were performed using weighted filtered back projection (WFBP) as the standard method, metal artifact reduction in image space (MARIS), and an iterative metal artifacts reduction (iMAR) algorithm for large implants. These datasets were used for PET attenuation correction. HU and SUV measurements were performed in nine predefined anatomical locations: liver, lower lung lobes, descending aorta, thoracic vertebral body, autochthonous back muscles, pectoral muscles, and internal jugular vein. Differences between HU and SUV measurements were compared using paired t-tests. The significance level was determined as p = 0.017 using Bonferroni correction. Results: No significant differences were observed between reconstructed images using iMAR and WFBP concerning HU and SUV measurements in liver (HU: p = 0.055; SUVmax: p = 0.586), lung (HU: p = 0.276; SUVmax: p = 1.0 for the right side and HU: p = 0.630; SUVmax: p = 0.109 for the left side), descending aorta (HU: p = 0.333; SUVmax: p = 0.083), thoracic vertebral body (HU: p = 0.725; SUVmax: p = 0.392), autochthonous back muscles (HU: p = 0.281; SUVmax: p = 0.839), pectoral muscles (HU: p = 0.481; SUVmax: p = 0.277 for the right side and HU: p = 0.313; SUVmax: p = 0.859 for the left side), or the internal jugular vein (HU: p = 0.343; SUVmax: p = 0.194). Conclusion: Metal artifact reduction algorithms such as iMAR do not alter the data information of normal organs not affected by artifacts. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

10 pages, 1201 KiB

Open AccessArticle

Classification of the Confocal Microscopy Images of Colorectal Tumor and Inflammatory Colitis Mucosa Tissue Using Deep Learning

by Jaehoon Jeong, Seung Taek Hong, Ihsan Ullah, Eun Sun Kim and Sang Hyun Park

Diagnostics 2022, 12(2), 288; https://doi.org/10.3390/diagnostics12020288 - 24 Jan 2022

Cited by 2 | Viewed by 3482

Abstract

Confocal microscopy image analysis is a useful method for neoplasm diagnosis. Many ambiguous cases are difficult to distinguish with the naked eye, thus leading to high inter-observer variability and significant time investments for learning this method. We aimed to develop a deep learning-based [...] Read more.

Confocal microscopy image analysis is a useful method for neoplasm diagnosis. Many ambiguous cases are difficult to distinguish with the naked eye, thus leading to high inter-observer variability and significant time investments for learning this method. We aimed to develop a deep learning-based neoplasm classification model that classifies confocal microscopy images of 10× magnified colon tissues into three classes: neoplasm, inflammation, and normal tissue. ResNet50 with data augmentation and transfer learning approaches was used to efficiently train the model with limited training data. A class activation map was generated by using global average pooling to confirm which areas had a major effect on the classification. The proposed method achieved an accuracy of 81%, which was 14.05% more accurate than three machine learning-based methods and 22.6% better than the predictions made by four endoscopists. ResNet50 with data augmentation and transfer learning can be utilized to effectively identify neoplasm, inflammation, and normal tissue in confocal microscopy images. The proposed method outperformed three machine learning-based methods and identified the area that had a major influence on the results. Inter-observer variability and the time required for learning can be reduced if the proposed model is used with confocal microscopy image analysis for diagnosis. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

13 pages, 2253 KiB

Open AccessReview

Magnetic Resonance Imaging as a Prognostic Disability Marker in Clinically Isolated Syndrome and Multiple Sclerosis: A Systematic Review and Meta-Analysis

by Amjad I. AlTokhis, Abrar AlAmrani, Abdulmajeed Alotaibi, Anna Podlasek and Cris S. Constantinescu

Diagnostics 2022, 12(2), 270; https://doi.org/10.3390/diagnostics12020270 - 21 Jan 2022

Cited by 6 | Viewed by 3602

Abstract

To date, there are no definite imaging predictors for long-term disability in multiple sclerosis (MS). Magnetic resonance imaging (MRI) is the key prognostic tool for MS, primarily at the early stage of the disease. Recent findings showed that white matter lesion (WML) counts [...] Read more.

To date, there are no definite imaging predictors for long-term disability in multiple sclerosis (MS). Magnetic resonance imaging (MRI) is the key prognostic tool for MS, primarily at the early stage of the disease. Recent findings showed that white matter lesion (WML) counts and volumes could predict long-term disability for MS. However, the prognostic value of MRI in the early stage of the disease and its link to long-term physical disability have not been assessed systematically and quantitatively. A meta-analysis was conducted using studies from four databases to assess whether MS lesion counts and volumes at baseline MRI scans could predict long-term disability, assessed by the expanded disability status scale (EDSS). Fifteen studies were eligible for the qualitative analysis and three studies for meta-analysis. T2 brain lesion counts and volumes after the disease onset were associated with disability progression after 10 years. Four or more lesions at baseline showed a highly significant association with EDSS 3 and EDSS 6, with a pooled OR of 4.10 and 4.3, respectively. The risk increased when more than 10 lesions were present. This review and meta-analysis confirmed that lesion counts and volumes could be associated with disability and might offer additional valid guidance in treatment decision making. Future work is essential to determine whether these prognostic markers have high predictive potential. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

11 pages, 2211 KiB

Open AccessArticle

Image-Quality Assessment of Polyenergetic and Virtual Monoenergetic Reconstructions of Unenhanced CT Scans of the Head: Initial Experiences with the First Photon-Counting CT Approved for Clinical Use

by Arwed Elias Michael, Jan Boriesosdick, Denise Schoenbeck, Matthias Michael Woeltjen, Saher Saeed, Jan Robert Kroeger, Sebastian Horstmeier, Simon Lennartz, Jan Borggrefe and Julius Henning Niehoff

Diagnostics 2022, 12(2), 265; https://doi.org/10.3390/diagnostics12020265 - 21 Jan 2022

Cited by 24 | Viewed by 4720

Abstract

In 2021, the first clinical photon-counting CT (PCCT) was introduced. The purpose of this study is to evaluate the image quality of polyenergetic and virtual monoenergetic reconstructions in unenhanced PCCTs of the head. A total of 49 consecutive patients with unenhanced PCCTs of [...] Read more.

In 2021, the first clinical photon-counting CT (PCCT) was introduced. The purpose of this study is to evaluate the image quality of polyenergetic and virtual monoenergetic reconstructions in unenhanced PCCTs of the head. A total of 49 consecutive patients with unenhanced PCCTs of the head were retrospectively included. The signals ± standard deviations of the gray and white matter were measured at three different locations in axial slices, and a measure of the artifacts below the cranial calvaria and in the posterior fossa between the petrous bones was also obtained. The signal-to-noise ratios (SNRs) and contrast-to-noise ratios (CNRs) were calculated for all reconstructions. In terms of the SNRs and CNRs, the polyenergetic reconstruction is superior to all virtual monoenergetic reconstructions (p < 0.001). In the MERs, the highest SNR is found in the 70 keV MER, and the highest CNR is in the 65 keV MER. In terms of artifacts below the cranial calvaria and in the posterior fossa, certain MERs are superior to polyenergetic reconstruction (p < 0.001). The PCCT provided excellent image contrast and low-noise profiles for the differentiation of the grey and white matter. Only the artifacts below the calvarium and in the posterior fossa still underperform, which is attributable to the lack of an artifact reduction algorithm in image postprocessing. It is conceivable that the usual improvements in image postprocessing, especially with regard to glaring artifacts, will lead to further improvements in image quality. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

13 pages, 2852 KiB

Open AccessArticle

Ultrasound Radiomics Nomogram Integrating Three-Dimensional Features Based on Carotid Plaques to Evaluate Coronary Artery Disease

by Xiaoting Wang, Peng Luo, Huaan Du, Shiyu Li, Yi Wang, Xun Guo, Li Wan, Binyi Zhao and Jianli Ren

Diagnostics 2022, 12(2), 256; https://doi.org/10.3390/diagnostics12020256 - 20 Jan 2022

Cited by 6 | Viewed by 2636

Abstract

This study aimed to explore the feasibility of ultrasound radiomics analysis before invasive coronary angiography (ICA) for evaluating the severity of coronary artery disease (CAD) quantified by the SYNTAX score (SS). This study included 105 carotid plaques from 105 patients (64 low-SS patients, [...] Read more.

This study aimed to explore the feasibility of ultrasound radiomics analysis before invasive coronary angiography (ICA) for evaluating the severity of coronary artery disease (CAD) quantified by the SYNTAX score (SS). This study included 105 carotid plaques from 105 patients (64 low-SS patients, 41 intermediate-high-SS patients). The clinical characteristics and three-dimensional ultrasound (3D-US) features before ICA were assessed. Ultrasound images of carotid plaques were used for radiomics analysis. Least absolute shrinkage and selection operator (LASSO) regression, which generated several nonzero coefficients, was used to select features that could predict intermediate-high SS. Based on those coefficients, the radiomics score (Rad-score) was calculated. The selected clinical characteristics, 3D-US features, and Rad-score were finally integrated into a radiomics nomogram. Among the clinical characteristics and 3D-US features, high-density lipoprotein (HDL), apolipoprotein B (Apo B), and plaque volume were identified as predictors for distinguishing between low SS and intermediate-high SS. During the radiomics process, 8 optimal radiomics features most capable of identifying intermediate-high SS were selected from 851 candidate radiomics features. The differences in Rad-score between the training and the validation set were significant (p = 0.016 and 0.006). The radiomics nomogram integrating HDL, Apo B, plaque volume, and Rad-score showed excellent results in the training set (AUC, 0.741 (95% confidence interval (CI): 0.646–0.835)) and validation set (AUC, 0.939 (95% CI: 0.860–1.000)), with good calibration (mean absolute errors of 0.028 and 0.059 in training and validation sets, respectively). Decision curve analysis showed that the radiomics nomogram could identify patients who could obtain the most benefit. We concluded that the radiomics nomogram based on carotid plaque ultrasound has favorable value for the noninvasive prediction of intermediate-high SS. This radiomics nomogram has potential value for the risk stratification of CAD before ICA and provides clinicians with a noninvasive diagnostic tool. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

7 pages, 626 KiB

Open AccessArticle

Volumetric Measurements in Lung Cancer Screening Reduces Unnecessary Low-Dose Computed Tomography Scans: Results from a Single-Center Prospective Trial on 4119 Subjects

by Gianluca Milanese, Federica Sabia, Roberta Eufrasia Ledda, Stefano Sestini, Alfonso Vittorio Marchianò, Nicola Sverzellati and Ugo Pastorino

Diagnostics 2022, 12(2), 229; https://doi.org/10.3390/diagnostics12020229 - 18 Jan 2022

Cited by 3 | Viewed by 1837

Abstract

This study aims to compare the low-dose computed tomography (LDCT) outcome and volume-doubling time (VDT) derived from the measured volume (MV) and estimated volume (EV) of pulmonary nodules (PNs) detected in a single-center lung cancer screening trial. MV, EV and VDT were obtained [...] Read more.

This study aims to compare the low-dose computed tomography (LDCT) outcome and volume-doubling time (VDT) derived from the measured volume (MV) and estimated volume (EV) of pulmonary nodules (PNs) detected in a single-center lung cancer screening trial. MV, EV and VDT were obtained for prevalent pulmonary nodules detected at the baseline round of the bioMILD trial. The LDCT outcome (based on bioMILD thresholds) and VDT categories were simulated on PN- and screenee-based analyses. A weighted Cohen’s kappa test was used to assess the agreement between diagnostic categories as per MV and EV, and 1583 screenees displayed 2715 pulmonary nodules. In the PN-based analysis, 40.1% PNs were included in different LDCT categories when measured by MV or EV. The agreements between MV and EV were moderate (κ = 0.49) and fair (κ = 0.37) for the LDCT outcome and VDT categories, respectively. In the screenee-based analysis, 46% pulmonary nodules were included in different LDCT categories when measured by MV or EV. The agreements between MV and EV were moderate (κ = 0.52) and fair (κ = 0.34) for the LDCT outcome and VDT categories, respectively. Within a simulated lung cancer screening based on a recommendation by estimated volumetry, the number of LDCTs performed for the evaluation of pulmonary nodules was higher compared with in prospective volumetric management. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

13 pages, 2454 KiB

Open AccessArticle

Dual-Energy CT Pulmonary Angiography for the Assessment of Surgical Accessibility in Patients with Chronic Thromboembolic Pulmonary Hypertension

by Matthias Eberhard, Micheal McInnis, Marc de Perrot, Mona Lichtblau, Silvia Ulrich, Ilhan Inci, Isabelle Opitz and Thomas Frauenfelder

Diagnostics 2022, 12(2), 228; https://doi.org/10.3390/diagnostics12020228 - 18 Jan 2022

Cited by 6 | Viewed by 2570

Abstract

We assessed the value of dual-energy CT pulmonary angiography (CTPA) for classification of the level of disease in chronic thromboembolic pulmonary hypertension (CTEPH) patients compared to the surgical Jamieson classification and prediction of hemodynamic changes after pulmonary endarterectomy. Forty-three CTEPH patients (mean age, [...] Read more.

We assessed the value of dual-energy CT pulmonary angiography (CTPA) for classification of the level of disease in chronic thromboembolic pulmonary hypertension (CTEPH) patients compared to the surgical Jamieson classification and prediction of hemodynamic changes after pulmonary endarterectomy. Forty-three CTEPH patients (mean age, 57 ± 16 years; 18 females) undergoing CTPA prior to surgery were retrospectively included. “Proximal” and “distal disease” were defined as L1 and 2a (main and lobar pulmonary artery [PA]) and L2b-4 (lower lobe basal trunk to subsegmental PA), respectively. Three radiologists had a moderate interobserver agreement for the radiological classification of disease (k = 0.55). Sensitivity was 92–100% and specificity was 24–53% to predict proximal disease according to the Jamieson classification. A median of 9 segments/patient had CTPA perfusion defects (range, 2–18 segments). L1 disease had a greater decrease in the mean pulmonary artery pressure (p = 0.029) and pulmonary vascular resistance (p = 0.011) after surgery compared to patients with L2a to L3 disease. The extent of perfusion defects was not associated with the level of disease or hemodynamic changes after surgery (p > 0.05 for all). CTPA is highly sensitive for predicting the level of disease in CTEPH patients with a moderate interobserver agreement. The radiological level of disease is associated with hemodynamic improvement after surgery. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

15 pages, 3922 KiB

Open AccessArticle

Local Structure Awareness-Based Retinal Microaneurysm Detection with Multi-Feature Combination

by Jiakun Deng, Puying Tang, Xuegong Zhao, Tian Pu, Chao Qu and Zhenming Peng

Biomedicines 2022, 10(1), 124; https://doi.org/10.3390/biomedicines10010124 - 7 Jan 2022

Cited by 13 | Viewed by 2259

Abstract

Retinal microaneurysm (MA) is the initial symptom of diabetic retinopathy (DR). The automatic detection of MA is helpful to assist doctors in diagnosis and treatment. Previous algorithms focused on the features of the target itself; however, the local structural features of the target [...] Read more.

Retinal microaneurysm (MA) is the initial symptom of diabetic retinopathy (DR). The automatic detection of MA is helpful to assist doctors in diagnosis and treatment. Previous algorithms focused on the features of the target itself; however, the local structural features of the target and background are also worth exploring. To achieve MA detection, an efficient local structure awareness-based retinal MA detection with the multi-feature combination (LSAMFC) is proposed in this paper. We propose a novel local structure feature called a ring gradient descriptor (RGD) to describe the structural differences between an object and its surrounding area. Then, a combination of RGD with the salience and texture features is used by a Gradient Boosting Decision Tree (GBDT) for candidate classification. We evaluate our algorithm on two public datasets, i.e., the e-ophtha MA dataset and retinopathy online challenge (ROC) dataset. The experimental results show that the performance of the trained model significantly improved after combining traditional features with RGD, and the area under the receiver operating characteristic curve (AUC) values in the test results of the datasets e-ophtha MA and ROC increased from 0.9615 to 0.9751 and from 0.9066 to 0.9409, respectively. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

10 pages, 1628 KiB

Open AccessArticle

Blood Flow Analysis of the Great Saphenous Vein in the Su-Pine Position in Clinical Manifestations of Varicose Veins of Different Severities: Application of Phase-Contrast Magnetic Resonance Imaging Data

by Yuan-Hsi Tseng, Chien-Wei Chen, Min-Yi Wong, Teng-Yao Yang, Yu-Hui Lin, Bor-Shyh Lin and Yao-Kuang Huang

Diagnostics 2022, 12(1), 118; https://doi.org/10.3390/diagnostics12010118 - 5 Jan 2022

Cited by 2 | Viewed by 2684

Abstract

The progression of clinical manifestations of lower-limb varicose veins remains unclear. This study investigated changes in lower-limb venous blood flow using phase-contrast magnetic resonance angiography. Data were collected on veins from 141 legs. We compared legs with and without varicose veins and related [...] Read more.

The progression of clinical manifestations of lower-limb varicose veins remains unclear. This study investigated changes in lower-limb venous blood flow using phase-contrast magnetic resonance angiography. Data were collected on veins from 141 legs. We compared legs with and without varicose veins and related symptoms and examined varying levels of varicose vein symptom severity. Legs without varicose veins exhibited a lower absolute stroke volume (ASV, p < 0.01) and mean flux (MF, p = 0.03) for the great saphenous vein (GSV) compared with legs with symptomatic varicose veins. Legs with asymptomatic varicose veins exhibited lower MF for the GSV (p = 0.02) compared with legs with symptomatic varicose veins. Among legs with varicose veins, asymptomatic legs exhibited lower ASV (p = 0.03) and MF (p = 0.046) for the GSV compared with legs that exhibited skin changes or ulcers; however, no significant differences were observed between legs presenting with discomfort or edema and legs with skin changes or ulcers, and between legs presenting with discomfort or edema and asymptomatic legs. In conclusion, in the supine position, increased blood flow rate and blood flow volume in the GSV were associated with symptomatic varicose veins and increased symptom severity. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

20 pages, 1971 KiB

Open AccessSystematic Review

Diagnostic Accuracy of Imaging Findings in Pleural Empyema: Systematic Review and Meta-Analysis

by Desiree Zettinig, Tugba Akinci D’Antonoli, Adrian Wilder-Smith, Jens Bremerich, Jan A. Roth and Raphael Sexauer

J. Imaging 2022, 8(1), 3; https://doi.org/10.3390/jimaging8010003 - 28 Dec 2021

Cited by 6 | Viewed by 4762

Abstract

Computed tomography (CT) diagnosis of empyema is challenging because current literature features multiple overlapping pleural findings. We aimed to identify informative findings for structured reporting. The screening according to inclusion criteria (P: Pleural empyema, I: CT C: culture/gram-stain/pathology/pus, O: Diagnostic accuracy measures), data [...] Read more.

Computed tomography (CT) diagnosis of empyema is challenging because current literature features multiple overlapping pleural findings. We aimed to identify informative findings for structured reporting. The screening according to inclusion criteria (P: Pleural empyema, I: CT C: culture/gram-stain/pathology/pus, O: Diagnostic accuracy measures), data extraction, and risk of bias assessment of studies published between 01-1980 and 10-2021 on Pubmed, Embase, and Web of Science (WOS) were performed independently by two reviewers. CT findings with pooled diagnostic odds ratios (DOR) with 95% confidence intervals, not including 1, were considered as informative. Summary estimates of diagnostic accuracy for CT findings were calculated by using a bivariate random-effects model and heterogeneity sources were evaluated. Ten studies with a total of 252 patients with and 846 without empyema were included. From 119 overlapping descriptors, five informative CT findings were identified: Pleural enhancement, thickening, loculation, fat thickening, and fat stranding with an AUC of 0.80 (hierarchical summary receiver operating characteristic, HSROC). Potential sources of heterogeneity were different thresholds, empyema prevalence, and study year. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

20 pages, 4195 KiB

Open AccessArticle

Effectiveness of Learning Systems from Common Image File Types to Detect Osteosarcoma Based on Convolutional Neural Networks (CNNs) Models

by Chanunya Loraksa, Sirima Mongkolsomlit, Nitikarn Nimsuk, Meenut Uscharapong and Piya Kiatisevi

J. Imaging 2022, 8(1), 2; https://doi.org/10.3390/jimaging8010002 - 27 Dec 2021

Cited by 13 | Viewed by 3345

Abstract

Osteosarcoma is a rare bone cancer which is more common in children than in adults and has a high chance of metastasizing to the patient’s lungs. Due to initiated cases, it is difficult to diagnose and hard to detect the nodule in a [...] Read more.

Osteosarcoma is a rare bone cancer which is more common in children than in adults and has a high chance of metastasizing to the patient’s lungs. Due to initiated cases, it is difficult to diagnose and hard to detect the nodule in a lung at the early state. Convolutional Neural Networks (CNNs) are effectively applied for early state detection by considering CT-scanned images. Transferring patients from small hospitals to the cancer specialized hospital, Lerdsin Hospital, poses difficulties in information sharing because of the privacy and safety regulations. CD-ROM media was allowed for transferring patients’ data to Lerdsin Hospital. Digital Imaging and Communications in Medicine (DICOM) files cannot be stored on a CD-ROM. DICOM must be converted into other common image formats, such as BMP, JPG and PNG formats. Quality of images can affect the accuracy of the CNN models. In this research, the effect of different image formats is studied and experimented. Three popular medical CNN models, VGG-16, ResNet-50 and MobileNet-V2, are considered and used for osteosarcoma detection. The positive and negative class images are corrected from Lerdsin Hospital, and 80% of all images are used as a training dataset, while the rest are used to validate the trained models. Limited training images are simulated by reducing images in the training dataset. Each model is trained and validated by three different image formats, resulting in 54 testing cases. F1-Score and accuracy are calculated and compared for the models’ performance. VGG-16 is the most robust of all the formats. PNG format is the most preferred image format, followed by BMP and JPG formats, respectively. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

25 pages, 4045 KiB

Open AccessArticle

Efficient Pre-Processing and Segmentation for Lung Cancer Detection Using Fused CT Images

by Imran Nazir, Ihsan Ul Haq, Muhammad Mohsin Khan, Muhammad Bilal Qureshi, Hayat Ullah and Sharjeel Butt

Electronics 2022, 11(1), 34; https://doi.org/10.3390/electronics11010034 - 23 Dec 2021

Cited by 23 | Viewed by 5633

Abstract

Over the last two decades, radiologists have been using multi-view images to detect tumors. Computer Tomography (CT) imaging is considered as one of the reliable imaging techniques. Many medical-image-processing techniques have been developed to diagnoses lung cancer at early or later stages through [...] Read more.

Over the last two decades, radiologists have been using multi-view images to detect tumors. Computer Tomography (CT) imaging is considered as one of the reliable imaging techniques. Many medical-image-processing techniques have been developed to diagnoses lung cancer at early or later stages through CT images; however, it is still a big challenge to improve the accuracy and sensitivity of the algorithms. In this paper, we propose an algorithm based on image fusion for lung segmentation to optimize lung cancer diagnosis. The image fusion technique was developed through Laplacian Pyramid (LP) decomposition along with Adaptive Sparse Representation (ASR). The suggested fusion technique fragments medical images into different sizes using the LP. After that, the LP is used to fuse the four decomposed layers. For the evaluation purposes of the proposed technique, the Lungs Image Database Consortium and Image Database Resource Initiative (LIDC-IDRI) was used. The results showed that the Dice Similarity Coefficient (DSC) index of our proposed method was 0.9929, which is better than recently published results. Furthermore, the values of other evaluation parameters such as the sensitivity, specificity, and accuracy were 89%, 98% and 99%, respectively, which are also competitive with the recently published results. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

10 pages, 1442 KiB

Open AccessArticle

Cumulative Effective Dose from Medical Imaging in Inflammatory Bowel Disease

by Agata Łukawska, Dominika Ślósarz, Aneta Zimoch, Karol Serafin, Elżbieta Poniewierka and Radosław Kempiński

Diagnostics 2021, 11(12), 2387; https://doi.org/10.3390/diagnostics11122387 - 18 Dec 2021

Cited by 2 | Viewed by 2440

Abstract

Inflammatory bowel diseases (IBD) are chronic and relapsing disorders usually requiring numerous medical imaging. IBD patients might be exposed to a large dose of radiation. As a cumulative effective dose (CED) ≥ 50 mSv is considered significant for stochastic risks of cancer, it [...] Read more.

Inflammatory bowel diseases (IBD) are chronic and relapsing disorders usually requiring numerous medical imaging. IBD patients might be exposed to a large dose of radiation. As a cumulative effective dose (CED) ≥ 50 mSv is considered significant for stochastic risks of cancer, it is important to monitor the radiation exposure of IBD patients. In the present work, we aimed to quantify the mean CED in IBD patients and identify factors associated with exposure to high doses of diagnostic radiation. A retrospective chart view of patients with IBD hospitalized between 2015 and 2019 was performed. A total of 65 patients with Crohn’s disease (CD) and 98 patients with ulcerative colitis (UC) were selected. Of all imaging studies performed, 73% were with doses of ionizing radiation. Mean CED (SD) amounted to 19.20 (15.64) millisieverts (mSv) and 6.66 (12.39) mSv, respectively, in patients with CD and UC (p < 0.00001). Only 1.84% of the patients received CED ≥ 50 mSv. We identified three factors associated with CED in the IBD patients: number of surgical procedures, and number and length of hospitalization. CD patients with strictures or penetrating disease and UC patients with extensive colitis were more likely to receive higher radiation doses. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

11 pages, 1217 KiB

Open AccessArticle

An Algorithm for Individual Dosage in Cadmium–Zinc–Telluride SPECT-Gated Radionuclide Angiography

by Maria Normand Hansen, Christian Haarmark, Bent Kristensen and Bo Zerahn

Diagnostics 2021, 11(12), 2268; https://doi.org/10.3390/diagnostics11122268 - 4 Dec 2021

Cited by 1 | Viewed by 1705

Abstract

The aim of the present study was to test an individualised dose without compromising the ease of analysing data when performing equilibrium radionuclide angiography (ERNA) using cadmium–zinc–telluride (CZT) SPECT. From March 2018 to January 2019, 1650 patients referred for ERNA received either an [...] Read more.

The aim of the present study was to test an individualised dose without compromising the ease of analysing data when performing equilibrium radionuclide angiography (ERNA) using cadmium–zinc–telluride (CZT) SPECT. From March 2018 to January 2019, 1650 patients referred for ERNA received either an individualised dose of ^99mTc-labeled human serum albumin (HSA) according to their age, sex, height, and weight (n = 1567), or a standard dose of 550 MBq (n = 83). The target count rate (CR_T) was reduced every two months from 2.7 to 1.0 kcps. A final test with a CR_T of 1.7 kcps was run for three months to test whether an agreement within 2% points for the determination of LVEF, on the basis of only two analyses, was obtainable in at least 95% of acquisitions. All the included ERNAs were performed on a dedicated cardiac CZT SPECT camera. When using the algorithm for an individualised dose, we found that agreement between the measured and predicted count rate was 80%. With a CR_T of 1.7 kcps, the need for more than two analyses to obtain sufficient agreement for LVEF was 4.9%. Furthermore, this resulted in a mean dose reduction from 550 to 258 MBq. Patients’ weight, height, sex, and age can, therefore, be used for individualising a tracer dose while reducing the mean dose. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

12 pages, 5987 KiB

Open AccessArticle

Unraveling Variations in Celiac Trunk and Hepatic Artery by CT Angiography to Aid in Surgeries of Upper Abdominal Region

by Kapil Kumar Malviya, Ashish Verma, Amit Kumar Nayak, Anand Mishra and Raghunath Shahaji More

Diagnostics 2021, 11(12), 2262; https://doi.org/10.3390/diagnostics11122262 - 3 Dec 2021

Cited by 5 | Viewed by 6731

Abstract

Understanding of variations in the course and source of abdominal arteries is crucial for any surgical intervention in the peritoneal space. Intricate surgeries of the upper abdominal region, such as hepato-biliary, pancreatic, gastric and splenic surgeries, require precise knowledge of regular anatomy and [...] Read more.

Understanding of variations in the course and source of abdominal arteries is crucial for any surgical intervention in the peritoneal space. Intricate surgeries of the upper abdominal region, such as hepato-biliary, pancreatic, gastric and splenic surgeries, require precise knowledge of regular anatomy and different variations related to celiac trunk and hepatic artery. In addition, information about the origin of inferior phrenic artery is important in conditions such as hepatocellular carcinoma and gastroesophageal bleeding management. The present study gives an account of anatomical variations in origin and branching pattern of celiac trunk and hepatic artery by the use of CT (computed tomographic) angiography. The study was performed on 110 (66 females and 44 males) patients in a north Indian population. Results unraveled the most common celiac trunk variation as hepatosplenic trunk with left gastric artery, which was observed in 60% of cases, more common in females than in males. Gastrosplenic and hepato-gastric trunk could be seen in 4.55% and 1.82% cases respectively. Gastrosplenic trunk was more commonly found in females, whereas hepato-gastric trunk was more common in males. A gastrosplenic trunk, along with the hepato-mesenteric trunk, was observed in 1.82% cases and was more common in males. A celiacomesenteric trunk, in which the celiac trunk and superior mesenteric artery originated as a common trunk from the aorta, was seen only in 0.91% of cases, and exhibited an origin of right and left inferior phrenic artery from the left gastric artery. The most common variation of hepatic artery, in which the right hepatic artery was replaced and originated from the superior mesenteric artery, was observed in 3.64%, cases with a more common occurrence in males. In 1.82% cases, the left hepatic artery was replaced and originated from the left gastric artery, which was observed only in females. Common hepatic artery originated from the superior mesenteric artery, as observed in 1.82% cases, with slightly higher occurrence in males. These findings not only add to the existing knowledge apart from giving an overview of variations in north Indian population, but also give an account of their correlation with gender. The present study will prove to be important for various surgeries of the upper abdominal region. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

20 pages, 1337 KiB

Open AccessArticle

Measurement of Blood Pressure by Ultrasound—The Applicability of Devices, Algorithms and a View in Local Hemodynamics

by Moritz Meusel, Philipp Wegerich, Berit Bode, Elena Stawschenko, Kristina Kusche-Vihrog, Horst Hellbrück and Hartmut Gehring

Diagnostics 2021, 11(12), 2255; https://doi.org/10.3390/diagnostics11122255 - 2 Dec 2021

Cited by 5 | Viewed by 3319

Abstract

Objective: Due to ongoing technical progress, the ultrasonic measurement of blood pressure (BP) as an alternative to oscillometric measurement (NIBP) or the continuous non-invasive arterial pressure method (CNAP) moves further into focus. The US method offers several advantages over NIBP and CNAP, such [...] Read more.

Objective: Due to ongoing technical progress, the ultrasonic measurement of blood pressure (BP) as an alternative to oscillometric measurement (NIBP) or the continuous non-invasive arterial pressure method (CNAP) moves further into focus. The US method offers several advantages over NIBP and CNAP, such as deep tissue penetration and the utilization of different arterial locations. Approach: Ten healthy subjects (six female, aged 30.9 ± 4.6 years) volunteered in our investigation. In the ultrasonic BP measurement, we differentiated between the directly measured (pulsatile diastolic and systolic vessel diameter) and indirectly calculated variables at three different artery locations on both arms, with two different ultrasound devices in the transversal and longitudinal directions of the transducer. Simultaneously, NIBP monitoring served as reference BP, while CNAP monitored the steady state condition of the arm under investigation. The Moens–Korteweg algorithm (MKE) and the algorithm of the working group of San Diego (SanD) were selected for the indirectly calculated ultrasonic BP data. Main results: With US, we were able to measure the BP at each selected arterial position. Due to the investigation setup, we found small but significant interactions of the main effects. Bland and Altman analysis revealed that US-BP measurement was similar to NIBP, with superior accuracy when compared to the established CNAP method. In addition, US-BP measurement showed that the measurement accuracy of both arms can be regarded as identical. In a detailed comparison of the selected arterial vascular sections, systematic discrepancies between the right and left arm could be observed. Conclusion: In our pilot study, we measured BP effectively and accurately by US using two different devices. Our findings suggest that ultrasonic BP measurement is an adequate alternative for live and continuous hemodynamic monitoring. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

9 pages, 3464 KiB

Open AccessArticle

Impact of Image Resolution on Deep Learning Performance in Endoscopy Image Classification: An Experimental Study Using a Large Dataset of Endoscopic Images

by Vajira Thambawita, Inga Strümke, Steven A. Hicks, Pål Halvorsen, Sravanthi Parasa and Michael A. Riegler

Diagnostics 2021, 11(12), 2183; https://doi.org/10.3390/diagnostics11122183 - 24 Nov 2021

Cited by 63 | Viewed by 5905

Abstract

Recent trials have evaluated the efficacy of deep convolutional neural network (CNN)-based AI systems to improve lesion detection and characterization in endoscopy. Impressive results are achieved, but many medical studies use a very small image resolution to save computing resources at the cost [...] Read more.

Recent trials have evaluated the efficacy of deep convolutional neural network (CNN)-based AI systems to improve lesion detection and characterization in endoscopy. Impressive results are achieved, but many medical studies use a very small image resolution to save computing resources at the cost of losing details. Today, no conventions between resolution and performance exist, and monitoring the performance of various CNN architectures as a function of image resolution provides insights into how subtleties of different lesions on endoscopy affect performance. This can help set standards for image or video characteristics for future CNN-based models in gastrointestinal (GI) endoscopy. This study examines the performance of CNNs on the HyperKvasir dataset, consisting of 10,662 images from 23 different findings. We evaluate two CNN models for endoscopic image classification under quality distortions with image resolutions ranging from 32 × 32 to 512 × 512 pixels. The performance is evaluated using two-fold cross-validation and F1-score, maximum Matthews correlation coefficient (MCC), precision, and sensitivity as metrics. Increased performance was observed with higher image resolution for all findings in the dataset. MCC was achieved at image resolutions between 512 × 512 pixels for classification for the entire dataset after including all subclasses. The highest performance was observed with an MCC value of 0.9002 when the models were trained on the highest resolution and tested on the same resolution. Different resolutions and their effect on CNNs are explored. We show that image resolution has a clear influence on the performance which calls for standards in the field in the future. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

14 pages, 3224 KiB

Open AccessArticle

Diabetic Retinopathy Diagnosis Based on RA-EfficientNet

by San-Li Yi, Xue-Lian Yang, Tian-Wei Wang, Fu-Rong She, Xin Xiong and Jian-Feng He

Appl. Sci. 2021, 11(22), 11035; https://doi.org/10.3390/app112211035 - 22 Nov 2021

Cited by 32 | Viewed by 4090

Abstract

The early detection and grade diagnosis of diabetic retinopathy (DR) are very important for the avoidance of blindness, and using deep learning methods to automatically diagnose DR has attracted great attention. However, the small amount of DR data limits its application. To automatically [...] Read more.

The early detection and grade diagnosis of diabetic retinopathy (DR) are very important for the avoidance of blindness, and using deep learning methods to automatically diagnose DR has attracted great attention. However, the small amount of DR data limits its application. To automatically learn the disease’s features and detect DR more accurately, we constructed a DR grade diagnostic model. To realize the model, the authors performed the following steps: firstly, we preprocess the DR images to solve the existing problems in an APTOS 2019 dataset, such as size difference, information redundancy and the data imbalance. Secondly, to extract more valid image features, a new network named RA-EfficientNet is proposed, in which a residual attention (RA) block is added to EfficientNet to extract more features and to solve the problem of small differences between lesions. EfficientNet has been previously trained on the ImageNet dataset, based on transfer learning technology, to overcome the small sample size problem of DR. Lastly, based on the extracted features, two classifiers are designed, one is a 2-grade classifier and the other a 5-grade classifier. The 2-grade classifier can diagnose DR, and the 5-grade classifier provides 5 grades of diagnosis for DR, as follows: 0 for No DR, 1 for mild DR, 2 for moderate, 3 for severe and 4 for proliferative DR. Experiments show that our proposed RA-EfficientNet can achieve better performance, with an accuracy value of 98.36% and a kappa score of 96.72% in a 2-grade classification and an accuracy value of 93.55% and a kappa score of 91.93% in a 5-grade classification. The results indicate that the proposed model effectively improves DR detection efficiency and resolves the existing limitation of manual feature extraction. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

14 pages, 3433 KiB

Open AccessReview

Imaging Techniques for Cardiac Function

by Vasileios Panis and Erwan Donal

Appl. Sci. 2021, 11(22), 10549; https://doi.org/10.3390/app112210549 - 9 Nov 2021

Cited by 6 | Viewed by 2730

Abstract

Cardiac imaging techniques include a variety of distinct applications with which we can visualize cardiac function non-invasively. Through different applications of physical entities such as sound waves, X-rays, magnetic fields, and nuclear energy, along with highly sophisticated computer hardware and software, it is [...] Read more.

Cardiac imaging techniques include a variety of distinct applications with which we can visualize cardiac function non-invasively. Through different applications of physical entities such as sound waves, X-rays, magnetic fields, and nuclear energy, along with highly sophisticated computer hardware and software, it is now possible to reconstruct the dynamic aspect of cardiac function in many forms, from static images to high-definition videos and real-time three-dimensional projections. In this review, we will describe the fundamental principles of the most widely used techniques and, more specifically, which imaging modality and on what occasion we should use them in order to analyze different aspects of cardiac function. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

13 pages, 2232 KiB

Open AccessArticle

Multiple Instance Classification for Gastric Cancer Pathological Images Based on Implicit Spatial Topological Structure Representation

by Xu Xiang and Xiaofeng Wu

Appl. Sci. 2021, 11(21), 10368; https://doi.org/10.3390/app112110368 - 4 Nov 2021

Cited by 2 | Viewed by 2214

Abstract

Gastric cancer is a malignant tumor with high incidence. Computer-aided screening systems for gastric cancer pathological images can contribute to reducing the workload of specialists and improve the efficiency of disease diagnosis. Due to the high resolution of images, it is common to [...] Read more.

Gastric cancer is a malignant tumor with high incidence. Computer-aided screening systems for gastric cancer pathological images can contribute to reducing the workload of specialists and improve the efficiency of disease diagnosis. Due to the high resolution of images, it is common to divide the whole slide image (WSI) into a set of image patches with overlap before utilizing deep neural networks for further analysis. However, not all patches split from the same cancerous WSI contain information of cancerous issues. This restriction naturally satisfies the assumptions of multiple instance learning (MIL). Moreover, the spatial topological structure relationships between local areas in a WSI are destroyed in the process of patch partitioning. Most existing multiple instance classification (MIC) methods fail to take into account the topological relationships between instances. In this paper, we propose a novel multiple instance classification framework based on graph convolutional networks (GCNs) for gastric microscope image classification. Firstly, patch embeddings were generated by feature extraction. Then, a graph structure was introduced to model the spatial topological structure relationships between instances. Additionally, a graph classification model with hierarchical pooling was constructed to achieve this multiple instance classification task. To certify the effectiveness and generalization of our method, we conducted comparative experiments on two different modes of gastric cancer pathological image datasets. The proposed method achieved average fivefold cross-validation precisions of 91.16% and 98.26% for gastric cancer classification on the two datasets, respectively. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

10 pages, 1717 KiB

Open AccessArticle

A Radioactive-Free Method for the Thorough Analysis of the Kinetics of Cell Cytotoxicity

by Claudia Coronnello, Rosalia Busà, Luca Cicero, Albert Comelli and Ester Badami

J. Imaging 2021, 7(11), 222; https://doi.org/10.3390/jimaging7110222 - 23 Oct 2021

Cited by 2 | Viewed by 3438

Abstract

The cytotoxic activity of T cells and Natural Killer cells is usually measured with the chromium release assay (CRA), which involves the use of 51Chromium (⁵¹Cr), a radioactive substance dangerous to the operator and expensive to handle and dismiss. The accuracy [...] Read more.

The cytotoxic activity of T cells and Natural Killer cells is usually measured with the chromium release assay (CRA), which involves the use of 51Chromium (⁵¹Cr), a radioactive substance dangerous to the operator and expensive to handle and dismiss. The accuracy of the measurements depends on how well the target cells incorporate ⁵¹Cr during labelling which, in turn, depends on cellular division. Due to bystander metabolism, the target cells spontaneously release ⁵¹Cr, producing a high background noise. Alternative radioactive-free methods have been developed. Here, we compare a bioluminescence (BLI)-based and a carboxyfluorescein succinimidyl ester (CFSE)-based cytotoxicity assay to the standard radioactive CRA. In the first assay, the target cells stably express the enzyme luciferase, and vitality is measured by photon emission upon the addition of the substrate d-luciferin. In the second one, the target cells are labelled with CFSE, and the signal is detected by Flow Cytometry. We used these two protocols to measure cytotoxicity induced by treatment with NK cells. The cytotoxicity of NK cells was determined by adding increasing doses of human NK cells. The results obtained with the BLI method were consistent with those obtained with the CRA- or CFSE-based assays 4 hours after adding the NK cells. Most importantly, with the BLI assay, the kinetic of NK cells’ killing was thoroughly traced with multiple time point measurements, in contrast with the single time point measurement the other two methods allow, which unveiled additional information on NK cell killing pathways. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

5 pages, 3384 KiB

Open AccessInteresting Images

Value of ¹⁸F-FDG PET/CT for Assessment of Advanced Lacrimal Sac Non-Keratinizing Squamous Cell Carcinoma Successfully Treated with Concurrent Chemoradiotherapy

by Ching-Yu Liao, Li-An Huang and Yu-Hsuan Lin

Diagnostics 2021, 11(11), 1961; https://doi.org/10.3390/diagnostics11111961 - 22 Oct 2021

Viewed by 1813

Abstract

Non-keratinizing squamous cell carcinoma (NKSCC) of the lacrimal apparatus is extremely rare. It is usually very aggressive in destroying local tissue and has a grave prognosis for relentless recurrence and distant failures. Though the current evidence cannot make confident recommendations regarding the best [...] Read more.

Non-keratinizing squamous cell carcinoma (NKSCC) of the lacrimal apparatus is extremely rare. It is usually very aggressive in destroying local tissue and has a grave prognosis for relentless recurrence and distant failures. Though the current evidence cannot make confident recommendations regarding the best management, curative surgical excision with adjuvant radiotherapy remains the most commonly used strategy. Here, we report a 71-year-old woman presented with progressive right medial canthal swellings for six months. A transnasal endoscopic biopsy revealed NKSCC of the lacrimal sac. She then underwent a combination of magnetic resonance images (MRI) and 2-deoxy-2-(18F)fluoro-D-glucose positron emission tomography/computed tomography (¹⁸F-FDG PET/CT) for staging purposes. Following cisplatin-based concurrent chemo-radiotherapy (CCRT), the post-treatment PET/CT illustrated the absence of an abnormal metabolic accumulation over the suspicious region as observed in post-treatment CT. A further trans-ostia re-biopsy confirmed complete tumor remission. This case demonstrates the remarkable ability of ¹⁸F-FDG PET/CT to differentiate between a persistent malignancy and post-treatment changes. Furthermore, a definite CCRT might provide comparable outcomes to traditional surgery. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

26 pages, 6444 KiB

Open AccessArticle

A Reinforcement Learning Algorithm for Automated Detection of Skin Lesions

by Usman Ahmad Usmani, Junzo Watada, Jafreezal Jaafar, Izzatdin Abdul Aziz and Arunava Roy

Appl. Sci. 2021, 11(20), 9367; https://doi.org/10.3390/app11209367 - 9 Oct 2021

Cited by 27 | Viewed by 4529

Abstract

Skin cancers are increasing at an alarming rate, and detection in the early stages is essential for advanced treatment. The current segmentation methods have limited labeling ability to the ground truth images due to the numerous noisy expert annotations present in the datasets. [...] Read more.

Skin cancers are increasing at an alarming rate, and detection in the early stages is essential for advanced treatment. The current segmentation methods have limited labeling ability to the ground truth images due to the numerous noisy expert annotations present in the datasets. The precise boundary segmentation is essential to correctly locate and diagnose the various skin lesions. In this work, the lesion segmentation method is proposed as a Markov decision process. It is solved by training an agent to segment the region using a deep reinforcement-learning algorithm. Our method is similar to the delineation of a region of interest by the physicians. The agent follows a set of serial actions for the region delineation, and the action space is defined as a set of continuous action parameters. The segmentation model learns in continuous action space using the deep deterministic policy gradient algorithm. The proposed method enables continuous improvement in performance as we proceed from coarse segmentation results to finer results. Finally, our proposed model is evaluated on the International Skin Imaging Collaboration (ISIC) 2017 image dataset, Human against Machine (HAM10000), and PH² dataset. On the ISIC 2017 dataset, the algorithm achieves an accuracy of 96.33% for the naevus cases, 95.39% for the melanoma cases, and 94.27% for the seborrheic keratosis cases. The other metrics are evaluated on these datasets and rank higher when compared with the current state-of-the-art lesion segmentation algorithms. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

17 pages, 1877 KiB

Open AccessArticle

Shared Patterns of Brain Functional Connectivity for the Comorbidity between Migraine and Insomnia

by Kun-Hsien Chou, Chen-Yuan Kuo, Chih-Sung Liang, Pei-Lin Lee, Chia-Kuang Tsai, Chia-Lin Tsai, Ming-Hao Huang, Yi-Chih Hsu, Guan-Yu Lin, Yu-Kai Lin, Ching-Po Lin and Fu-Chi Yang

Biomedicines 2021, 9(10), 1420; https://doi.org/10.3390/biomedicines9101420 - 9 Oct 2021

Cited by 10 | Viewed by 3234

Abstract

Migraine is commonly comorbid with insomnia; both disorders are linked to functional disturbance of the default mode network (DMN). Evidence suggests that DMN could be segregated into multiple subnetworks with specific roles that underline different cognitive processes. However, the relative contributions of DMN [...] Read more.

Migraine is commonly comorbid with insomnia; both disorders are linked to functional disturbance of the default mode network (DMN). Evidence suggests that DMN could be segregated into multiple subnetworks with specific roles that underline different cognitive processes. However, the relative contributions of DMN subnetworks in the comorbidity of migraine and insomnia remain largely unknown. This study sought to identify altered functional connectivity (FC) profiles of DMN subnetworks in the comorbidity of migraine and insomnia. Direct group comparisons with healthy controls, followed by conjunction analyses, were used to identify shared FC alterations of DMN subnetworks. The shared FC changes of the DMN subnetworks in the migraine and insomnia groups were identified in the dorsomedial prefrontal and posteromedial cortex subnetworks. These shared FC changes were primarily associated with motor and somatosensory systems, and consistently found in patients with comorbid migraine and insomnia. Additionally, the magnitude of FC between the posteromedial cortex and postcentral gyrus correlated with insomnia duration in patients with comorbid migraine and insomnia. Our findings point to specific FC alterations of the DMN subnetwork in migraine and insomnia. The shared patterns of FC disturbance may be associated with the underlying mechanisms of the comorbidity of the two disorders. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

16 pages, 32343 KiB

Open AccessArticle

Dermoscopy Images Enhancement via Multi-Scale Morphological Operations

by Julio César Mello-Román, José Luis Vázquez Noguera, Horacio Legal-Ayala, Miguel García-Torres, Jacques Facon, Diego P. Pinto-Roa, Sebastian A. Grillo, Luis Salgueiro Romero, Lizza A. Salgueiro Toledo, Laura Raquel Bareiro Paniagua, Deysi Natalia Leguizamon Correa and Jorge Daniel Mello-Román

Appl. Sci. 2021, 11(19), 9302; https://doi.org/10.3390/app11199302 - 7 Oct 2021

Cited by 4 | Viewed by 2868

Abstract

Skin dermoscopy images frequently lack contrast caused by varying light conditions. Indeed, often low contrast is seen in dermoscopy images of melanoma, causing the lesion to blend in with the surrounding skin. In addition, the low contrast prevents certain details from being seen [...] Read more.

Skin dermoscopy images frequently lack contrast caused by varying light conditions. Indeed, often low contrast is seen in dermoscopy images of melanoma, causing the lesion to blend in with the surrounding skin. In addition, the low contrast prevents certain details from being seen in the image. Therefore, it is necessary to design an approach that can enhance the contrast and details of dermoscopic images. In this work, we propose a multi-scale morphological approach to reduce the impacts of lack of contrast and to enhance the quality of the images. By top-hat reconstruction, the local bright and dark features are extracted from the image. The local bright features are added and the dark features are subtracted from the image. In this way, images with higher contrast and detail are obtained. The proposed approach was applied to a database of 236 color images of benign and malignant melanocytic lesions. The results show that the multi-scale morphological approach by reconstruction is a competitive algorithm since it achieved a very satisfactory level of contrast enhancement and detail enhancement in dermoscopy images. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

15 pages, 2380 KiB

Open AccessArticle

Visual-Saliency-Based Abnormality Detection for MRI Brain Images—Alzheimer’s Disease Analysis

by A. Diana Andrushia, K. Martin Sagayam, Hien Dang, Marc Pomplun and Lien Quach

Appl. Sci. 2021, 11(19), 9199; https://doi.org/10.3390/app11199199 - 2 Oct 2021

Cited by 15 | Viewed by 2728

Abstract

In recent years, medical image analysis has played a vital role in detecting diseases in their early stages. Medical images are rapidly becoming available for various applications to solve human problems. Therefore, complex medical features are needed to develop a diagnostic system for [...] Read more.

In recent years, medical image analysis has played a vital role in detecting diseases in their early stages. Medical images are rapidly becoming available for various applications to solve human problems. Therefore, complex medical features are needed to develop a diagnostic system for physicians to provide better treatment. Traditional methods of abnormality detection suffer from misidentification of abnormal regions in the given data. Visual-saliency detection methods are used to locate abnormalities to improve the accuracy of the proposed work. This study explores the role of a visual saliency map in the classification of Alzheimer’s disease (AD). Bottom-up saliency corresponds to image features, whereas top-down saliency uses domain knowledge in magnetic resonance imaging (MRI) brain images. The novelty of the proposed method lies in the use of an elliptical local binary pattern descriptor for low-level MRI characterization. Ellipse-like topologies help to obtain feature information from different orientations. Extensively directional features at different orientations cover the micro patterns. The brain regions of the Alzheimer’s disease stages were classified from the saliency maps. Multiple-kernel learning (MKL) and simple and efficient MKL (SEMKL) were used to classify Alzheimer’s disease from normal controls. The proposed method used the OASIS dataset and experimental results were compared with eight state-of-the-art methods. The proposed visual saliency-based abnormality detection produces reliable results in terms of accuracy, sensitivity, specificity, and f-measure. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

14 pages, 5819 KiB

Open AccessFeature PaperArticle

Skin Characterizations by Using Contact Capacitive Imaging and High-Resolution Ultrasound Imaging with Machine Learning Algorithms

by Elena Chirikhina, Andrey Chirikhin, Sabina Dewsbury-Ennis, Francesco Bianconi and Perry Xiao

Appl. Sci. 2021, 11(18), 8714; https://doi.org/10.3390/app11188714 - 18 Sep 2021

Cited by 11 | Viewed by 3522

Abstract

We present our latest research on skin characterizations by using Contact Capacitive Imaging and High-Resolution Ultrasound Imaging with Machine Learning algorithms. Contact Capacitive Imaging is a novel imaging technology based on the dielectric constant measurement principle, with which we have studied the skin [...] Read more.

We present our latest research on skin characterizations by using Contact Capacitive Imaging and High-Resolution Ultrasound Imaging with Machine Learning algorithms. Contact Capacitive Imaging is a novel imaging technology based on the dielectric constant measurement principle, with which we have studied the skin water content of different skin sites and performed image classification by using pre-trained Deep Learning Neural Networks through Transfer Learning. The results show lips and nose have the lowest water content, whilst cheek, eye corner and under-eye have the highest water content. The classification yields up to 83.8% accuracy. High-Resolution Ultrasound Imaging is a state-of-the-art ultrasound technology, and can produce high-resolution images of the skin and superficial soft tissue to a vertical resolution of about 40 microns, with which we have studied the thickness of different skin layers, such as stratum corneum, epidermis and dermis, around different locations on the face and around different body parts. The results show the chin has the highest stratum corneum thickness, and the arm has the lowest stratum corneum thickness. We have also developed two feature-based image classification methods which yield promising results. The outcomes of this study could provide valuable guidelines for cosmetic/medical research, and methods developed in this study can also be extended for studying damaged skin or skin diseases. The combination of Contact Capacitive Imaging and High-Resolution Ultrasound Imaging could be a powerful tool for skin studies. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

8 pages, 1587 KiB

Open AccessArticle

Assessment of MicroPET Image Quality Based on Reconstruction Methods and Post-Filtering

by Hyeon-Sik Kim, Byeong-il Lee and Jae-Sung Ahn

Appl. Sci. 2021, 11(18), 8707; https://doi.org/10.3390/app11188707 - 18 Sep 2021

Cited by 1 | Viewed by 1770

Abstract

The accuracy of positron emission tomography (PET) imaging is hampered by the partial volume effect (PVE), which causes image blurring and sampling. The PVE produces spillover phenomena, making PET analysis difficult. Generally, the PVE values vary based on reconstruction methods and filtering. Thus, [...] Read more.

The accuracy of positron emission tomography (PET) imaging is hampered by the partial volume effect (PVE), which causes image blurring and sampling. The PVE produces spillover phenomena, making PET analysis difficult. Generally, the PVE values vary based on reconstruction methods and filtering. Thus, selection of the proper reconstruction and filtering method can ensure accurate and high-quality PET images. This study compared the values of factors (recovery coefficient (RC), uniformity, and spillover ratio (SOR)) associated with different reconstruction and post-filtering methods using a mouse image quality phantom (NEMA NU 4), and we present an effective approach for microPET images. The PET images were obtained using a microPET scanner (Inveon, Siemens Medical Solutions, Malvern, PA, USA). PET data were reconstructed and/or post-filtered. For tumors smaller than 3 mm, iterative reconstruction methods provided better image quality. For tumor sizes bigger than 3 mm, reconstruction methods without post-filtering showed better results. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

17 pages, 1661 KiB

Open AccessArticle

Deep Learning for COVID-19 Diagnosis from CT Images

by Andrea Loddo, Fabio Pili and Cecilia Di Ruberto

Appl. Sci. 2021, 11(17), 8227; https://doi.org/10.3390/app11178227 - 4 Sep 2021

Cited by 23 | Viewed by 3993

Abstract

COVID-19, an infectious coronavirus disease, caused a pandemic with countless deaths. From the outset, clinical institutes have explored computed tomography as an effective and complementary screening tool alongside the reverse transcriptase-polymerase chain reaction. Deep learning techniques have shown promising results in similar medical [...] Read more.

COVID-19, an infectious coronavirus disease, caused a pandemic with countless deaths. From the outset, clinical institutes have explored computed tomography as an effective and complementary screening tool alongside the reverse transcriptase-polymerase chain reaction. Deep learning techniques have shown promising results in similar medical tasks and, hence, may provide solutions to COVID-19 based on medical images of patients. We aim to contribute to the research in this field by: (i) Comparing different architectures on a public and extended reference dataset to find the most suitable; (ii) Proposing a patient-oriented investigation of the best performing networks; and (iii) Evaluating their robustness in a real-world scenario, represented by cross-dataset experiments. We exploited ten well-known convolutional neural networks on two public datasets. The results show that, on the reference dataset, the most suitable architecture is VGG19, which (i) Achieved 98.87% accuracy in the network comparison; (ii) Obtained 95.91% accuracy on the patient status classification, even though it misclassifies some patients that other networks classify correctly; and (iii) The cross-dataset experiments exhibit the limitations of deep learning approaches in a real-world scenario with 70.15% accuracy, which need further investigation to improve the robustness. Thus, VGG19 architecture showed promising performance in the classification of COVID-19 cases. Nonetheless, this architecture enables extensive improvements based on its modification, or even with preprocessing step in addition to it. Finally, the cross-dataset experiments exposed the critical weakness of classifying images from heterogeneous data sources, compatible with a real-world scenario. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

19 pages, 1306 KiB

Open AccessArticle

On the Effectiveness of Leukocytes Classification Methods in a Real Application Scenario

by Andrea Loddo and Lorenzo Putzu

AI 2021, 2(3), 394-412; https://doi.org/10.3390/ai2030025 - 25 Aug 2021

Cited by 12 | Viewed by 9163

Abstract

Automating the analysis of digital microscopic images to identify the cell sub-types or the presence of illness has assumed a great importance since it aids the laborious manual process of review and diagnosis. In this paper, we have focused on the analysis of [...] Read more.

Automating the analysis of digital microscopic images to identify the cell sub-types or the presence of illness has assumed a great importance since it aids the laborious manual process of review and diagnosis. In this paper, we have focused on the analysis of white blood cells. They are the body’s main defence against infections and diseases and, therefore, their reliable classification is very important. Current systems for leukocyte analysis are mainly dedicated to: counting, sub-types classification, disease detection or classification. Although these tasks seem very different, they share many steps in the analysis process, especially those dedicated to the detection of cells in blood smears. A very accurate detection step gives accurate results in the classification of white blood cells. Conversely, when detection is not accurate, it can adversely affect classification performance. However, it is very common in real-world applications that work on inaccurate or non-accurate regions. Many problems can affect detection results. They can be related to the quality of the blood smear images, e.g., colour and lighting conditions, absence of standards, or even density and presence of overlapping cells. To this end, we performed an in-depth investigation of the above scenario, simulating the regions produced by detection-based systems. We exploit various image descriptors combined with different classifiers, including CNNs, in order to evaluate which is the most suitable in such a scenario, when performing two different tasks: Classification of WBC subtypes and Leukaemia detection. Experimental results have shown that Convolutional Neural Networks are very robust in such a scenario, outperforming common machine learning techniques combined with hand-crafted descriptors. However, when exploiting appropriate images for model training, even simpler approaches can lead to accurate results in both tasks. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

11 pages, 832 KiB

Open AccessFeature PaperArticle

Deep Learning for Drug Discovery: A Study of Identifying High Efficacy Drug Compounds Using a Cascade Transfer Learning Approach

by Dylan Zhuang and Ali K. Ibrahim

Appl. Sci. 2021, 11(17), 7772; https://doi.org/10.3390/app11177772 - 24 Aug 2021

Cited by 11 | Viewed by 4076

Abstract

In this research, we applied deep learning to rank the effectiveness of candidate drug compounds in combating viral cells, in particular, SARS-Cov-2 viral cells. For this purpose, two different datasets from Recursion Pharmaceuticals, a siRNA image dataset (RxRx1), which were used to build [...] Read more.

In this research, we applied deep learning to rank the effectiveness of candidate drug compounds in combating viral cells, in particular, SARS-Cov-2 viral cells. For this purpose, two different datasets from Recursion Pharmaceuticals, a siRNA image dataset (RxRx1), which were used to build and calibrate our model for feature extraction, and a SARS-CoV-2 dataset (RxRx19a) was used to train our model for ranking efficacy of candidate drug compounds. The SARS-CoV-2 dataset contained healthy, uninfected control or “mock” cells, as well as “active viral” cells (cells infected with COVID-19), which were the two cell types used to train our deep learning model. In addition, it contains viral cells treated with different drug compounds, which were the cells not used to train but test our model. We devised a new cascade transfer learning strategy to construct our model. We first trained a deep learning model, the DenseNet, with the siRNA set, a dataset with characteristics similar to the SARS-CoV-2 dataset, for feature extraction. We then added additional layers, including a SoftMax layer as an output layer, and retrained the model with active viral cells and mock cells from the SARS-CoV-2 dataset. In the test phase, the SoftMax layer outputs probability (equivalently, efficacy) scores which allows us to rank candidate compounds, and to study the performance of each candidate compound statistically. With this approach, we identified several compounds with high efficacy scores which are promising for the therapeutic treatment of COVID-19. The compounds showing the most promise were GS-441524 and then Remdesivir, which overlapped with these reported in the literature and with these drugs that are approved by FDA, or going through clinical trials and preclinical trials. This study shows the potential of deep learning in its ability to identify promising compounds to aid rapid responses to future pandemic outbreaks. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

19 pages, 6730 KiB

Open AccessFeature PaperArticle

Exploiting Spatial Information to Enhance DTI Segmentations via Spatial Fuzzy c-Means with Covariance Matrix Data and Non-Euclidean Metrics

by Safa Elsheikh, Andrew Fish and Diwei Zhou

Appl. Sci. 2021, 11(15), 7003; https://doi.org/10.3390/app11157003 - 29 Jul 2021

Cited by 3 | Viewed by 1619

Abstract

A diffusion tensor models the covariance of the Brownian motion of water at a voxel and is required to be symmetric and positive semi-definite. Therefore, image processing approaches, designed for linear entities, are not effective for diffusion tensor data manipulation, and the existence [...] Read more.

A diffusion tensor models the covariance of the Brownian motion of water at a voxel and is required to be symmetric and positive semi-definite. Therefore, image processing approaches, designed for linear entities, are not effective for diffusion tensor data manipulation, and the existence of artefacts in diffusion tensor imaging acquisition makes diffusion tensor data segmentation even more challenging. In this study, we develop a spatial fuzzy c-means clustering method for diffusion tensor data that effectively segments diffusion tensor images by accounting for the noise, partial voluming, magnetic field inhomogeneity, and other imaging artefacts. To retain the symmetry and positive semi-definiteness of diffusion tensors, the log and root Euclidean metrics are used to estimate the mean diffusion tensor for each cluster. The method exploits spatial contextual information and provides uncertainty information in segmentation decisions by calculating the membership values for assigning a diffusion tensor at one voxel to different clusters. A regularisation model that allows the user to integrate their prior knowledge into the segmentation scheme or to highlight and segment local structures is also proposed. Experiments on simulated images and real brain datasets from healthy and Spinocerebellar ataxia 2 subjects showed that the new method was more effective than conventional segmentation methods. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

14 pages, 4656 KiB

Open AccessArticle

Detecting the Absence of Lung Sliding in Lung Ultrasounds Using Deep Learning

by Miroslav Jaščur, Marek Bundzel, Marek Malík, Anton Dzian, Norbert Ferenčík and František Babič

Appl. Sci. 2021, 11(15), 6976; https://doi.org/10.3390/app11156976 - 29 Jul 2021

Cited by 15 | Viewed by 3424

Abstract

Certain post-thoracic surgery complications are monitored in a standard manner using methods that employ ionising radiation. A need to automatise the diagnostic procedure has now arisen following the clinical trial of a novel lung ultrasound examination procedure that can replace X-rays. Deep learning [...] Read more.

Certain post-thoracic surgery complications are monitored in a standard manner using methods that employ ionising radiation. A need to automatise the diagnostic procedure has now arisen following the clinical trial of a novel lung ultrasound examination procedure that can replace X-rays. Deep learning was used as a powerful tool for lung ultrasound analysis. We present a novel deep-learning method, automated M-mode classification, to detect the absence of lung sliding motion in lung ultrasound. Automated M-mode classification leverages semantic segmentation to select 2D slices across the temporal dimension of the video recording. These 2D slices are the input for a convolutional neural network, and the output of the neural network indicates the presence or absence of lung sliding in the given time slot. We aggregate the partial predictions over the entire video recording to determine whether the subject has developed post-surgery complications. With a 64-frame version of this architecture, we detected lung sliding on average with a balanced accuracy of 89%, sensitivity of 82%, and specificity of 92%. Automated M-mode classification is suitable for lung sliding detection from clinical lung ultrasound videos. Furthermore, in lung ultrasound videos, we recommend using time windows between 0.53 and 2.13 s for the classification of lung sliding motion followed by aggregation. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

21 pages, 9168 KiB

Open AccessArticle

Efficient and Accurate Hemorrhages Detection in Retinal Fundus Images Using Smart Window Features

by Tamoor Aziz, Ademola E. Ilesanmi and Chalie Charoenlarpnopparut

Appl. Sci. 2021, 11(14), 6391; https://doi.org/10.3390/app11146391 - 10 Jul 2021

Cited by 19 | Viewed by 4653

Abstract

Diabetic retinopathy (DR) is one of the diseases that cause blindness globally. Untreated accumulation of fat and cholesterol may trigger atherosclerosis in the diabetic patient, which may obstruct blood vessels. Retinal fundus images are used as diagnostic tools to screen abnormalities linked to [...] Read more.

Diabetic retinopathy (DR) is one of the diseases that cause blindness globally. Untreated accumulation of fat and cholesterol may trigger atherosclerosis in the diabetic patient, which may obstruct blood vessels. Retinal fundus images are used as diagnostic tools to screen abnormalities linked to diseases that affect the eye. Blurriness and low contrast are major problems when segmenting retinal fundus images. This article proposes an algorithm to segment and detect hemorrhages in retinal fundus images. The proposed method first performs preprocessing on retinal fundus images. Then a novel smart windowing-based adaptive threshold is utilized to segment hemorrhages. Finally, conventional and hand-crafted features are extracted from each candidate and classified by a support vector machine. Two datasets are used to evaluate the algorithms. Precision rate (P), recall rate (R), and F1 score are used for quantitative evaluation of segmentation methods. Mean square error, peak signal to noise ratio, information entropy, and contrast are also used to evaluate preprocessing method. The proposed method achieves a high F1 score with 83.85% for the DIARETDB1 image dataset and 72.25% for the DIARETDB0 image dataset. The proposed algorithm adequately adapts when compared with conventional algorithms, hence will act as a tool for segmentation. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

14 pages, 2193 KiB

Open AccessArticle

Quantitative Analyses of the Left Ventricle Volume and Cardiac Function in Normal and Infarcted Yucatan Minipigs

by Anna V. Naumova, Gregory Kicska, Kiana Pimentel, Lauren E. Neidig, Hiroshi Tsuchida, Kenta Nakamura and Charles E. Murry

J. Imaging 2021, 7(7), 107; https://doi.org/10.3390/jimaging7070107 - 1 Jul 2021

Cited by 1 | Viewed by 3150

Abstract

(1) Background: The accuracy of the left ventricular volume (LVV) and contractility measurements with cardiac magnetic resonance imaging (CMRI) is decreased if the papillary muscles are abnormally enlarged, such as in hypertrophic cardiomyopathy in human patients or in pig models of human diseases. [...] Read more.

(1) Background: The accuracy of the left ventricular volume (LVV) and contractility measurements with cardiac magnetic resonance imaging (CMRI) is decreased if the papillary muscles are abnormally enlarged, such as in hypertrophic cardiomyopathy in human patients or in pig models of human diseases. The purpose of this work was to establish the best method of LVV quantification with CMRI in pigs. (2) Methods: The LVV in 29 Yucatan minipig hearts was measured using two different techniques: the “standard method”, which uses smooth contouring along the endocardial surface and adds the papillary volume to the ventricular cavity volume, and the “detailed method”, which traces the papillary muscles and trabeculations and adds them to the ventricular mass. (3) Results: Papillary muscles add 21% to the LV mass in normal and infarcted hearts of Yucatan minipigs. The inclusion or exclusion of these from the CMRI analysis significantly affected the study results. In the normal pig hearts, the biggest differences were found in measurements of the LVV, ejection fraction (EF), LV mass and indices derived from the LV mass (p < 0.001). The EF measurement in the normal pig heart was 11% higher with the detailed method, and 19% higher in the infarcted pig hearts (p < 0.0001). The detailed method of endocardium tracing with CMRI closely represented the LV mass measured ex vivo. (4) Conclusions: The detailed method, which accounts for the large volume of the papillary muscles in the pig heart, provides better accuracy and interobserver consistency in the assessment of LV mass and ejection fraction, and might therefore be preferable for these analyses. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

16 pages, 692 KiB

Open AccessArticle

3D Volumetric Tensor Velocity Imaging with Low Computational Complexity Using a Row-Column Addressed Array

by Kseniya Chetverikova, Jørgen Arendt Jensen, Marie Sand Traberg and Matthias Bo Stuart

Appl. Sci. 2021, 11(12), 5757; https://doi.org/10.3390/app11125757 - 21 Jun 2021

Viewed by 2389

Abstract

A method for volumetric Tensor Velocity Imaging employing row-column (RC) addressed array with low computational complexity is investigated in simulations. An interleaved and non-interleaved sliding aperture sequence with 11 rows and 11 columns emissions by a 62 + 62 RC addressed array was [...] Read more.

A method for volumetric Tensor Velocity Imaging employing row-column (RC) addressed array with low computational complexity is investigated in simulations. An interleaved and non-interleaved sliding aperture sequence with 11 rows and 11 columns emissions by a 62 + 62 RC addressed array was used. The 3D velocities were estimated by a transverse oscillation (TO) cross-correlation estimator. Parabolic profiles at six different orientations corresponding to combinations of 0, 45 degrees azimuth angles and 90, 75, 60 beam-to-flow angles were investigated with 5 kHz pulse repetition frequencies. The Field II simulations were performed at a depth of 30 mm with peak velocity of 0.3 m/s. Across all vessel orientations, the relative mean bias varied from 2.3% to −14.26%, and the relative standard deviation varied from 0.43% to 5.5%. The best and worst performance was found at beam to flow angles of 90 degrees with 0 degrees rotation angle and 60 degrees beam-to-flow angle with 45 degrees rotation angle respectively. Due to the low channel count of the RC array and the low computational complexity, real-time implementation is feasible on conventional ultrasound systems. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

11 pages, 4778 KiB

Open AccessArticle

Analysis of Compound Stained Cervical Cell Images Using Multi-Spectral Imaging

by Run Fang, Libo Zeng and Fan Yi

Appl. Sci. 2021, 11(12), 5628; https://doi.org/10.3390/app11125628 - 18 Jun 2021

Cited by 2 | Viewed by 1954

Abstract

Multi-spectral imaging technique plays an important role in real-world applications such as medicine and medical detections. This paper proposes a cervical cancer cell screening method to simultaneously adopt TBS classification and DNA quantitative analysis for a single cell smear. Through using compound staining [...] Read more.

Multi-spectral imaging technique plays an important role in real-world applications such as medicine and medical detections. This paper proposes a cervical cancer cell screening method to simultaneously adopt TBS classification and DNA quantitative analysis for a single cell smear. Through using compound staining on a smear, the cytoplasm is stained by Papanicolauo and the nucleus is stained by Feulgen. The main evaluation parameter is the DNA content of the nucleus, not the subjective description of cell morphology, which is more objective than the TBS classification method and reduces the chances of missing a diagnosis due to subjective factors. Each nucleus has its own DI value and color image of the whole cell, which is convenient for doctors as it allows them to review and confirm the morphology of cells with a nucleus DI of over 2.5. Mouse liver smears and cervical cases are utilized as the measuring specimens to evaluate the performance of the microscope multi-spectral imaging system; illustrative results demonstrate that the proposed system qualifies, with high accuracy and reliability, and further presents wide application prospects in the early diagnosis of cervical cancer. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

15 pages, 654 KiB

Open AccessArticle

Automatic Microscopy Analysis with Transfer Learning for Classification of Human Sperm

by Rui Liu, Mingmei Wang, Min Wang, Jianqin Yin, Yixuan Yuan and Jun Liu

Appl. Sci. 2021, 11(12), 5369; https://doi.org/10.3390/app11125369 - 9 Jun 2021

Cited by 6 | Viewed by 7264

Abstract

Infertility is a global problem that affects many couples. Sperm analysis plays an essential role in the clinical diagnosis of human fertility. The examination of sperm morphology is an essential technique because sperm morphology is a proven indicator of biological functions. At present, [...] Read more.

Infertility is a global problem that affects many couples. Sperm analysis plays an essential role in the clinical diagnosis of human fertility. The examination of sperm morphology is an essential technique because sperm morphology is a proven indicator of biological functions. At present, the morphological classification of human sperm is conducted manually by medical experts. However, manual classification is laborious and highly dependent on the experience and capability of clinicians. To address these limitations, we propose a transfer learning method based on AlexNet to automatically classify the sperms into four different categories in terms of the World Health Organization (WHO) standards by analyzing their morphology. We adopt the feature extraction architecture of AlexNet as well as its pre-training parameters. Besides, we redesign the classification network by adding the Batch Normalization layers to improve the performance. The proposed method achieves an average accuracy of 96.0% and an average precision of 96.4% in the freely-available HuSHeM dataset, which exceeds the performance of previous algorithms. Our method shows that automatic sperm classification has great potential to replace manual sperm classification in the future. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

15 pages, 2892 KiB

Open AccessArticle

Evolution of Water-in-Oil Droplets in T-Junction Microchannel by Micro-PIV

by Hawa Ringkai, Khairul Fikri Tamrin, Nadeem Ahmed Sheikh and Shahrol Mohamaddan

Appl. Sci. 2021, 11(11), 5289; https://doi.org/10.3390/app11115289 - 7 Jun 2021

Cited by 7 | Viewed by 3577

Abstract

Water-in-oil droplets have huge importance in chemical and biotechnology applications, despite their difficulty being produced in microfluidics. Moreover, existing studies focus more on the different shape of microchannels instead of their size, which is one of the critical factors that can influence flow [...] Read more.

Water-in-oil droplets have huge importance in chemical and biotechnology applications, despite their difficulty being produced in microfluidics. Moreover, existing studies focus more on the different shape of microchannels instead of their size, which is one of the critical factors that can influence flow characteristics of the droplets. Therefore, the present work aims to study the behaviours of water-in-oil droplets at the interfacial surface in an offset T-junction microchannel, having different radiuses, using micro-PIV software. Food-grade palm olein and distilled water seeded with polystyrene microspheres particles were used as working fluids, and their captured images showing their generated droplets’ behaviours focused on the junction of the respective microfluidic channel, i.e., radiuses of 400 µm, 500 µm, 750 µm and 1000 µm, were analysed via PIVlab. The increasing in the radius of the offset T-junction microchannel leads to the increase in the cross-sectional area and the decrease in the distilled water phase’s velocity. The experimental velocity of the water droplet is in agreement with theoretical values, having a minimal difference as low as 0.004 mm/s for the case of the microchannel with a radius of 750 µm. In summary, a small increase in the channel’s size yields a significant increase in the overall flow of a liquid. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

12 pages, 3024 KiB

Open AccessArticle

Circle-U-Net: An Efficient Architecture for Semantic Segmentation

by Feng Sun, Ajith Kumar V, Guanci Yang, Ansi Zhang and Yiyun Zhang

Algorithms 2021, 14(6), 159; https://doi.org/10.3390/a14060159 - 21 May 2021

Cited by 13 | Viewed by 5164

Abstract

State-of-the-art semantic segmentation methods rely too much on complicated deep networks and thus cannot train efficiently. This paper introduces a novel Circle-U-Net architecture that exceeds the original U-Net on several standards. The proposed model includes circle connect layers, which is the backbone of [...] Read more.

State-of-the-art semantic segmentation methods rely too much on complicated deep networks and thus cannot train efficiently. This paper introduces a novel Circle-U-Net architecture that exceeds the original U-Net on several standards. The proposed model includes circle connect layers, which is the backbone of ResUNet-a architecture. The model possesses a contracting part with residual bottleneck and circle connect layers that capture context and expanding paths, with sampling layers and merging layers for a pixel-wise localization. The results of the experiment show that the proposed Circle-U-Net achieves an improved accuracy of 5.6676%, 2.1587% IoU (Intersection of union, IoU) and can detect 67% classes greater than U-Net, which is better than current results. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

11 pages, 535 KiB

Open AccessArticle

A Multi-Channel and Multi-Spatial Attention Convolutional Neural Network for Prostate Cancer ISUP Grading

by Bochen Yang and Zhifeng Xiao

Appl. Sci. 2021, 11(10), 4321; https://doi.org/10.3390/app11104321 - 11 May 2021

Cited by 10 | Viewed by 2990

Abstract

Prostate cancer (PCa) is one of the most prevalent cancers worldwide. As the demand for prostate biopsies increases, a worldwide shortage and an uneven geographical distribution of proficient pathologists place a strain on the efficacy of pathological diagnosis. Deep learning (DL) is able [...] Read more.

Prostate cancer (PCa) is one of the most prevalent cancers worldwide. As the demand for prostate biopsies increases, a worldwide shortage and an uneven geographical distribution of proficient pathologists place a strain on the efficacy of pathological diagnosis. Deep learning (DL) is able to automatically extract features from whole-slide images of prostate biopsies annotated by skilled pathologists and to classify the severity of PCa. A whole-slide image of biopsies has many irrelevant features that weaken the performance of DL models. To enable DL models to focus more on cancerous tissues, we propose a Multi-Channel and Multi-Spatial (MCMS) Attention module that can be easily plugged into any backbone CNN to enhance feature extraction. Specifically, MCMS learns a channel attention vector to assign weights to channels in the feature map by pooling from multiple attention branches with different reduction ratios; similarly, it also learns a spatial attention matrix to focus on more relevant areas of the image, by pooling from multiple convolutional layers with different kernel sizes. The model is verified on the most extensive multi-center PCa dataset that consists of 11,000 H&E-stained histopathology whole-slide images. Experimental results demonstrate that an MCMS-assisted CNN can effectively boost prediction performance in accuracy (ACC) and quadratic weighted kappa (QWK), compared with prior studies. The proposed model and results can serve as a credible benchmark for future research in automated PCa grading. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

15 pages, 1729 KiB

Open AccessArticle

Numerical Analysis of the Localization of Pulmonary Nodules during Thoracoscopic Surgery by Ultra-Wideband Radio Technology

by Alberto Battistel, Peter Paul Pott and Knut Möller

Appl. Sci. 2021, 11(9), 4282; https://doi.org/10.3390/app11094282 - 9 May 2021

Cited by 2 | Viewed by 1997

Abstract

Worldwide, lung cancer is one of the most common causes of cancer-related death. Detected by computer tomography, it is usually removed through thoracoscopic surgery. During the surgery the lung collapses requiring some strategies to track or localize the new position of the lesion. [...] Read more.

Worldwide, lung cancer is one of the most common causes of cancer-related death. Detected by computer tomography, it is usually removed through thoracoscopic surgery. During the surgery the lung collapses requiring some strategies to track or localize the new position of the lesion. This is particularly challenging in the case of minimally invasive surgeries when mechanical palpation is not possible. Here we undertake a preliminary study with numerical analysis of an ultra-wideband (UWB) radio technology which can be employed directly during thoracoscopic surgery to localize deep solitary pulmonary nodules. This study was conducted through Finite Difference Time Domain (FDTD) simulations, where a spherical target mimicking a nodule located between 1 and 6 cm of depth and an UWB pulse at several frequencies between 0.5 and 5 GHz was used for localization. This investigation quantifies the influence of several parameters, such frequency, lesion depth, and number of acquisitions, on the final confocal image used to locate a cancer in the lung tissue. We also provide extensive discussion on several artifacts that appear in the images. The results show that the cancer localization was possible at operational frequencies below 1 GHz and for deep nodules (>5 cm), while at lower depths and higher frequencies several artifacts hindered its detection. Full article

(This article belongs to the Topic Medical Image Analysis)

► Show Figures

Figure 1

Topic Menu

Topic Editors

Medical Image Analysis

Topic Information

Keywords

Participating Journals

Published Papers (115 papers)

Further Information

Guidelines

MDPI Initiatives

Follow MDPI