Next Article in Journal
Isocitrate Dehydrogenase 1/2 Wildtype Adult Astrocytoma with WHO Grade 2/3 Histological Features: Molecular Re-Classification, Prognostic Factors, Clinical Outcomes
Previous Article in Journal
Optimal Volume Assessment for Serous Fluid Cytology
Previous Article in Special Issue
Natural Fatty Acid Guards against Brain Endothelial Cell Death and Microvascular Pathology following Ischemic Insult in the Presence of Acute Hyperglycemia
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Biomedicines
Volume 12
Issue 4
10.3390/biomedicines12040896
biomedicines-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Integrating Demographics and Imaging Features for Various Stages of Dementia Classification: Feed Forward Neural Network Multi-Class Approach

by
Eva Y. W. Cheung
1,*,
Ricky W. K. Wu
2,
Ellie S. M. Chu
1 and
Henry K. F. Mak
3,*
1
School of Medical and Health Sciences, Tung Wah College, 31 Wylie Road, HoManTin, Hong Kong
2
Department of Biological and Biomedical Sciences, School of Health and Life Sciences, Glasgow Caledonian University, Glasgow G4 0BA, UK
3
Department of Diagnostic Radiology, School of Clinical Medicine, LKS Faculty of Medicine, University of Hong Kong, Hong Kong
*
Authors to whom correspondence should be addressed.
Biomedicines 2024, 12(4), 896; https://doi.org/10.3390/biomedicines12040896
Submission received: 29 November 2023 / Revised: 5 March 2024 / Accepted: 12 March 2024 / Published: 18 April 2024
(This article belongs to the Special Issue State of the Art: Prerequisites, Prevention and Treatment of Neurodegenerative Diseases)
Browse Figures
Versions Notes

Abstract

:
Background: MRI magnetization-prepared rapid acquisition (MPRAGE) is an easily available imaging modality for dementia diagnosis. Previous studies suggested that volumetric analysis plays a crucial role in various stages of dementia classification. In this study, volumetry, radiomics and demographics were integrated as inputs to develop an artificial intelligence model for various stages, including Alzheimer’s disease (AD), mild cognitive decline (MCI) and cognitive normal (CN) dementia classifications. Method: The Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset was separated into training and testing groups, and the Open Access Series of Imaging Studies (OASIS) dataset was used as the second testing group. The MRI MPRAGE image was reoriented via statistical parametric mapping (SPM12). Freesurfer was employed for brain segmentation, and 45 regional brain volumes were retrieved. The 3D Slicer software was employed for 107 radiomics feature extractions from within the whole brain. Data on patient demographics were collected from the datasets. The feed-forward neural network (FFNN) and the other most common artificial intelligence algorithms, including support vector machine (SVM), ensemble classifier (EC) and decision tree (DT), were used to build the models using various features. Results: The integration of brain regional volumes, radiomics and patient demographics attained the highest overall accuracy at 76.57% and 73.14% in ADNI and OASIS testing, respectively. The subclass accuracies in MCI, AD and CN were 78.29%, 89.71% and 85.14%, respectively, in ADNI testing, as well as 74.86%, 88% and 83.43% in OASIS testing. Balanced sensitivity and specificity were obtained for all subclass classifications in MCI, AD and CN. Conclusion: The FFNN yielded good overall accuracy for MCI, AD and CN categorization, with balanced subclass accuracy, sensitivity and specificity. The proposed FFNN model is simple, and it may support the triage of patients for further confirmation of the diagnosis.

1. Introduction

With an increasingly aging global population, the incidence of dementia is rapidly increasing. In 2016, there were 47 million people living with dementia worldwide. This figure is projected to increase to more than 131 million by 2050 [1]. The most common cause of dementia is Alzheimer’s disease (AD), which accounts for approximately 40% of all dementia cases. With recent pharmacological advancements, drug therapies for ameliorating the progression of AD [2] and improved preventive measures and therapies for AD have been developed. The early detection and accurate diagnosis of the prodromal stage of dementia, i.e., mild cognitive impairment (MCI), are crucial to reduce mortality, improve the quality of life and extend the lifespan of patients with dementia.
MRI magnetization-prepared rapid gradient-echo (MPRAGE) imaging is a down-stream imaging modality, which captures high tissue contrast with superior spatial resolution in a short scan time [3]. The three-dimensional application of whole-brain scans has been extensively used for AD diagnosis and disease progression monitoring. It provides detailed structural images of the brain, allowing physicians to visualize and assess the brain abnormalities associated with dementia. It is easily available and plays a crucial role in dementia diagnosis.
One of the MRI MPRAGE image applications is brain volumetric analysis. A significant volume reduction in the medial temporal lobe, including the hippocampus, precuneus, posterior cingulate, amygdala, parahippocampal gyrus and entorhinal cortex, is a signature for AD patients [4,5,6,7,8]. Through detailed hippocampal volume assessment [9,10], sub-regional corpus callosum atrophy [11,12] and connectivity-based segmentation of amygdala nuclei [13], AD can be effectively diagnosed from a cognitive normal (CN) state. Recent developments in automatic brain regional volume segmentation have improved the segmentation accuracy and can handle large amounts of data effectively. This allows for the comprehensive analysis of yearly MRI MPRAGE images for disease monitoring. Previous studies suggested that AD progression can be predicted based on the rate of volume reduction by monitoring the hippocampal volume change [14,15]. However, for the prodromal stage of AD, which is MCI, the brain regional volume change is subtle and cannot be easily detected by the naked eye. The diagnosis of MCI from AD requires either supplements with an up-stream imaging modality or extensive experience and knowledge from clinical experts. Neither of them is commonly available in memory clinics.
In recent years, the radiomics analysis of MRI MPRAGE has been widely applied in medical imaging. It is a novel technique that incorporates gray-level invariant features (GLIFs) into a data classification algorithm. It has the potential to reveal disease heterogeneity characteristics, which are related to the gray-level matrixes. This method has been adopted for cancer prognosis and recurrence prediction [16,17,18], the prediction of distant metastasis [19] and treatment response [20]. In view of dementia classification, an exploratory study was conducted by Li et al. 2020 using pure radiomics, and 55.9–56% accuracy was achieved in diagnosing preclinical AD. However, the accuracy improved to 76.1% when combined with other high-frequency features [21].
Biological differences and aging are other perspectives on dementia development. Previous studies suggested that women in many cohorts have a higher risk of developing AD [22,23]. Also, a higher incidence of dementia in elderly individuals is observed around the world, and the prevalence ranges from 5 to 7%, even after age standardization [24]. Including age and sex as parameters in the prediction model may have a positive impact on discriminating AD, MCI and CN in different perspectives.
During the past two decades, many studies have applied artificial intelligence to dementia classification using traditional classifiers, including logistic regression [25,26], decision tree (DT) [27], random forest [28,29,30], naïve Bayes [31], K-nearest neighbor [32], support vector machines (SVMs) [33,34,35,36,37,38] and ensemble classifier (EC) [39]. With the improvements in computer processing power, more studies have focused on using discriminative approaches, such as neural networks, in recent years. It is a branch that simulates the human brain, both in terms of structural and learning patterns. Compared to the traditional classifier, it allows for the processing of complicated high-level information by connecting a large number of inputs [40]. In addition, a multiple-layer neural network can capture complex non-linear relationships in data, as well as learning the relevant features automatically. The feed-forward neural network (FFNN) is one of the most popular neural networks being employed. It processes information from the input layer, through hidden layers to the output layer in one direction, without any feedback connections. It has only a few hidden layers, which requires less computation power to process, and is able to provide a good classification with a smaller dataset when compared to deep learning models. Previous studies showed good accuracies in identifying AD from CN (>85%) and MCI from CN (>80%) [41,42,43,44]. However, most studies relied on a single dataset to train and test the model. The models were not tested against unseen data, which may affect the generalizability of the built model and limit its application in clinical settings. Also, a binary classifier, i.e., to classify AD from CN or MCI from CN, was employed in most studies. In real-world scenarios, patients’ images were retrieved from multiple stages. The classification may be required to fit patients’ images into several binary classifiers to confirm the diagnosis. Instead of training and managing the multiple binary classifiers of each class, a multi-class classifier is designed to handle multiple classes simultaneously. Although it is more challenging to train and yields lower accuracy [45], the deployment of the multi-class model provides only one output for various stages of diseases. It is simple and efficient to precisely identify these diseases.
In this study, we aimed to develop a reliable artificial intelligence multi-class model to classify AD, MCI and CN using patient demographics and MRI imaging features. The first objective was to use various combinations of demographics and image features to build the models using FFNN and various traditional artificial intelligence algorithms, including DT, EC and SVM, as a comparison. The second objective was to compare the classification performances of the FFNN with those of the above-developed models to identify the algorithm that could provide a more accurate classification of AD, MCI and CN.

2. Materials and Methods

In this study, two cohorts of patients were used to build and validate the artificial intelligence models. For each patient, the demographics were recorded. In addition, the brain regional volumes, as well as radiomics from the whole brain, were retrieved from the MRI images as image features. The patients’ demographics, brain regional volumes and radiomics were integrated as inputs for model building using FFNN, DT, EC and SVM algorithms. The model classification performances were analyzed.

2.1. Patient Dataset

The datasets used in this study are the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu) [46] and the Open Access Series of Imaging Studies (OASIS) database (oasis-brains.org) [47]. The use of the above datasets was approved by the institutional review board at each site, and all participants provided written consent. All eligible participants underwent brain MRI MPRAGE scanning and clinical diagnosis with demographics collected.

2.1.1. ADNI Dataset

There were 25 memory centers from the USA which joined the ADNI project. A total of 582 images were collected from 25 centers. Further, 406 images (70% of all images) from 21 memory centers were partitioned as the training dataset, and 176 images (30% of all images) from the remaining 4 centers were used as validation datasets. The distribution of images is listed in Table 1.

2.1.2. OASIS Dataset

An independent cohort dataset (OASIS dataset) was collected from the Washington University Knight Alzheimer Disease Research Center. The entire OASIS dataset consists of 1552 patients. Thus, 176 patients, 28 AD, 91 MCI and 57 CN, were picked randomly. The total number of patients and the distribution of subclasses were the same as the testing dataset from ADNI 4 centers. This was to ensure the result of testing using dataset from ADNI 4 centers and that using the OASIS dataset would not be influenced by the number of patients and its subclass distribution.

2.2. Brain Segmentation and Regional Volume Analysis

FreeSurfer v7.1.0 image analysis suite was employed to perform brain segmentation and volumetric analysis. The procedures and algorithms employed were documented in previous publications [48,49,50,51,52,53,54] and are freely available from the website (http://surfer.nmr.mgh.harvard.edu/ (accessed on 22 January 2023)). Forty-five brain regional volumes were obtained. Details of the brain regions are listed in Table 2 and illustrated in Figure 1.

2.3. Radiomics Features

Reorientation of images was performed for each of the MPRAGE MRI images by SPM12 software [55]. The individualized whole-brain mask template was fused onto the MPRAGE image for brain regional configuration, which is shown in Figure 2. Further, 3D slicer software (The Slicer Community; V.4.11.20210226) with the PyRadiomics extension (Computational Imaging and Bioinformatics Lab, Harvard Medical School) was employed for the radiomics feature extraction [56]. One hundred and seven radiomics features were extracted within the whole brain from the MRI MPRAGE image for every patient. The definition of radiomics features was subdivided into eight classes [57], which are listed in Table 3.

2.4. Demographics

The MRI MPRAGE dataset and patients’ demographics were retrieved from the ADNI and OASIS website. The demographics of age and sex were recorded.

2.5. Integration of Patients’ Demographics and Image Features

The patients’ demographics, brain regional volumes and radiomics were integrated in the following 5 groups, which were used as inputs for building the artificial intelligence models: radiomics only with 107 features (R only), radiomics and patents’ demographics with 109 features (RD), volumes only with 45 features (V only), volumes and patients’ demographics with 47 features (VD) and volumes, radiomics and patients’ demographics with 154 features (VRD). Details are listed in Table 4.

2.6. Model Building

Patients from the ADNI dataset of 21 centers were used to build the models. The 5 groups of features obtained in Section 2.5 were used as input to build the models.
The proposed FFNN was built using Matlab® (R2021a) Neural Network toolbox. The neural network training employed Levenberg–Marquardt as a training algorithm with the random data division method. It had 5 layers, including 1 input layer, 3 hidden layers and 1 output layer. The input layers were the 5 groups of features in Section 2.5. The 3 hidden layers included 50 nodes in the first layer, 30 nodes in the second layer and 20 nodes in the last hidden layer for processing. In each hidden layer, the weight (w) and bias (b) are valued as a single vector, as shown in Figure 3. The FFNN is trained to fit input data; then, its weight and bias values are formed (+) into a vector (curve in the diagram) and fitted to the next layer. The output layer gave the result of classification. In this model, three subclasses, either AD, MCI or CN, were classified.
During model building, the hyper-parameter optimization algorithm was employed to control the learning process so as to optimally solve the problem. The maximum epoch was set to 50, and no training time limit was applied. The three-layered FFNN was trained using mean squared error performance function and a regularization value of 0.01. This was the early stopping-based optimization, which was used to stop training when the performance function and a regularization value of 0.01 were achieved. Details of the FFNN model building are listed in Figure 3.
In addition, the Matlab Classification Learner toolbox was employed to build the models using traditional artificial intelligence algorithms, including DT, EC and SVM, as a comparison. Hyper-parameter tuning was employed, with Bayesian optimization as optimizer, expected improvement per second plus as the acquisition function, the maximum iterations set as 30 and no training time limit applied, in DT, EC and SVM model building, so as to reduce the instability and provide simple models [58].
To improve the generalizability of the built models and avoid overfitting, 10-fold cross-validation was employed during each of the model-building processes. The dataset was divided into ten groups with an equal number of samples. The first neural network training process used the initial nine groups as training data and the remaining group as testing data. The second training process continued with another nine groups as training data and the rest as testing data. This process was undertaken 10 times. The performance of each model was the average result computed in these 10 rounds of training [59].
The performance of each model was assessed in terms of the overall accuracy, the classification ability of each subclass, i.e., MCI, AD and CN by class accuracy, sensitivity and specificity.

2.7. Model Testing and Data Analysis

Each model was tested using two independent cohorts of patients, including patients from the 4 centers of the ADNI dataset and those from the OASIS dataset. The performance of each model was assessed considering the overall accuracy, the classification ability of each subclass, i.e., MCI, AD and CN by class accuracy, sensitivity and specificity.

3. Results

We used five groups of features (R only, V only, RD, VD and VRD) to build models using four algorithms (FFNN, DT, EC and SVM); as a result, 20 models were built. Firstly, the value of the integration of multiple features was assessed through a performance evaluation using the same algorithm, with various features included in building the models. Secondly, the performance of the proposed FFNN was evaluated for various stages of dementia classification.

3.1. Dataset Demographics

The ADNI dataset comprises patients from 25 centers. Further, 406 patients from 21 centers (ADNI 21 centers) were selected to build the model, and 176 patients from the remaining 5 centers (ADNI 5 centers) were used to test the model. Another independent dataset from the OASIS database was used, with 176 patients used for secondary validation on the models built in Section 2.6. The demographics of the study cohort are shown in Table 5.

3.2. Performance Comparison in View of the Various Features Employed for Model Building

When comparing models built using the same model-building algorithm, those models built using volumes performed better, with higher overall accuracy, accuracy in characterizing MCI, AD and CN, sensitivity and specificity when compared to those models built using radiomics. Including demographics as features for either volumes or radiomics improved the overall accuracy when compared to the use of volume or radiomics alone. However, in SVM algorithms, the specificity of AD classification was zero when using VD or VRD features. Overall, in all models, the integration of volumes, radiomics and demographics attained the highest overall accuracy, balanced sensitivity and specificity, as well as the best accuracy in classification in MCI, AD and CN. The results are listed in Table 6.

3.3. Performance Evaluation of FFNN when Compared to Traditional Classifiers

The results from Section 3.2 suggest that the models using features from volumes, radiomics and demographics achieved the highest overall accuracy when compared to those built from either volumes or radiomics alone. Thus, we focused on analyzing models using all three features. In Table 7 e, it can be seen that the performance of FFNN was the best when compared to traditional classifiers. FFNN showed 76.57% and 73.14% overall accuracy in tests for patients from ADNI 4 centers and the OASIS database, respectively. In particular, the FFNN model attained good sensitivity and specificity.

4. Discussion

4.1. The Value of Integrating Image Features and Patient Demographics in AD, MCI and CN Classification

Our previous study suggested that structural MRI images aided in differentiating AD and MCI from CN using artificial intelligence [30]. However, that study was limited to binary classification, i.e., differentiating AD from CN, AD from MCI or MCI from CN. In clinical situations, CN may progress to MCI and then to AD in a matter of years. Multi-class classification is more useful considering three stages of disease. For two decades, brain regional volumes have been employed to diagnose AD from CN. Hippocampal atrophy is a widely used biomarker for the diagnosis of AD, but the low sensitivity and specificity limit its application as a confirmation of diagnosis [60]. Sørensen and his team suggested using other imaging features, including cortical thickness, hippocampal shape and its texture for the differential diagnosis of MCI from AD, and they achieved a classification accuracy of 62.7% for CN from AD and MCI [61]. Similar results were obtained by Koikkanlainen and his team, where 74% of AD could be accurately classified from other types of dementia using structural MRI [62]. Both authors suggested that other features might be required to attain higher accuracy in classification. Our results demonstrated that, using the volumes of only 45 brain regions, the overall classification accuracy achieved was 73.14% and 68% (EC) in validation for patients from ADNI 4 centers and OASIS, respectively. The results are similar to those obtained in previous studies. In subclass classification, however, the sensitivity was under 70% in AD and CN for all four algorithms. This suggested that the models built using brain regional volumes alone were unsatisfactory in identifying AD from CN.
In recent years, radiomics has been employed in the classification of AD, MCI and CN. Du and his team used radiomics features of the hippocampus for diagnosing early-onset and late-onset AD, which achieved 77% and 78%, respectively. However, their sample size was small, with only 144 patients included in training (36 patients in each group) and another 60 patients (15 patients in each group) for testing [63]. The limited sample size may restrict the generalizability of the classification model. Our results demonstrated that, using radiomics as the only feature for model building, the overall accuracy achieved 40.57% (SVM) to 51.43% (FFNN) in tests using patients from ADNI 4 centers and 35.43% (SVM) to 58.29 (EC) for patients from OASIS, respectively. The models built using radiomics alone were well below satisfactory.
To improve the classification accuracy, previous studies suggested building the model using multiple image features. Li and his team included 30,128 image features, including 24,910 features from structural MRI, 4988 features from functional MRI and 200 features from MRI Diffusion Tensor Imaging (DTI). They achieved overall accuracy of 90.2% and sensitivity and specificity of 79.8% and 86%, respectively [21]. The current study included only structural MRI image features, and, by adding patients’ demographics to the brain regional volumes and radiomics, the overall accuracy improved to 76.57% and 73.14% (FFNN) in tests for patients from ADNI 4 centers and OASIS, respectively. Also, the accuracy of MCI, AD and CN was 78.29%, 89.71% and 85.14% in tests for patients from ADNI 4 centers, which remained consistently high in tests for OASIS patients (74.86%, 88% and 83.43%).
In addition to the overall accuracy, the sensitivity and specificity, which refer to a model’s ability to classify patients with AD as AD, and to classify patients who were not AD as MCI and CN, respectively, were balanced. This illustrated that the model demonstrated high capability in classifying the corresponding groups accurately. To address the issue of an imbalanced subclass dataset, we used the precision and F1 score to evaluate the models. Precision was used to measure how many predictions for one group made by the model were correct. Recall was used to measure the number of one-class samples present in the dataset that were correctly identified by the model. The F1 score combines precision and recall using their harmonic mean. The high F1 score illustrates maximized precision and recall simultaneously. In the current study, FFNN demonstrated the highest precision and F1 score when compared to other algorithms, which demonstrated that the FFNN model can concurrently attain high precision and high recall, indicating well-balanced performance.

4.2. The Value of the Feed-Forward Neural Network in Classification of AD, MCI and CN

In this study, the FFNN showed the best performance in terms of accuracy, specificity and sensitivity for dementia classification when compared to other traditional algorithms. FFNN is a multi-layer artificial neural network, with a connection between the input layer, hidden layers and output layers. The training process allows information to move in one direction, from the input layer and hidden layers to the output layer, without looping (backpropagation) [64]. This simulates the thinking process of a physician in clinical decision making and diagnosis confirmation, based on information from patients’ demographics and imaging features.
The FFNN networks built in this study were relatively small in view of network training. There were only five layers (one input layer, three hidden layers and one output layer). The processing time is within 2 min when running on most computers in the clinical setting. The Levenberg–Marquardt algorithm used in FFNN offered significant accuracy, with fewer errors during the training, validation and testing phases [65].

4.3. The Value of Multi-Classes in Classification of AD, MCI and CN

Previous studies achieved good classification accuracies; for example, Zhang et al. 2019 achieved 96% accuracy in discriminating AD from CN, with sensitivity and specificity of 89% and 98%, respectively [66]. In addition, Mendoza-Leon and his team developed an auto-encoder model, which achieved accuracy of 90%, with sensitivity and specificity of 85% and 95%, respectively, in discriminating AD from CN [67]. Ning and his team demonstrated over 95% accuracy in classifying AD from CN [43]. A previous study from our team also achieved excellent classification accuracy, with 99.92% in differentiating MCI from CN, 99.86% to differentiate MCI from AD and 99.94% to differentiate AD from CN. However, these models were binary classifiers [30]. In real-world scenarios, however, patients can be taken from either stage. The multi-class model is a one-stop model, which can differentiate AD, MCI and CN distinctively. Technically, Borchert et al. 2023 highlighted that building a multi-class classifier model is more challenging than a binary classifier in view of the machine learning algorithm, and it usually yielded lower accuracy, sensitivity and specificity [68]. Compared to similar studies—one conducted by Moore and his team, where their model achieved accuracy of 99%, 59% and 29% in CN, MCI and AD, respectively [69], and another study conducted by Cárdenas-Pẽna and his team, where their model achieved 71.4%, 53.4% and 75.1% in CN, MCI and AD, respectively [42]—our proposed FFNN with VRD features yielded 83.43% 74.86% and 88% accuracies in CN, MCI and AD, respectively, in the OASIS test dataset. The balanced accuracies in various stages demonstrated that the model has the capability to classify all three stages of disease with satisfactory results, leading to a precise stage classification in real-world scenarios.

4.4. The Value of Testing against Independent Cohort of Patients

Another important asset of the current study is the use of an independent dataset for validation. Compared to those studies using cross-validation or other similar methods, the use of an independent dataset demonstrated much lower accuracy [70]. For instance, Cohen et al. 2019 achieved accuracies of 93.1%, 82.3% and 88.6% in CN, MCI and AD, respectively [44]; however, their algorithm was not tested against unseen data. A review study concluded that, when compared to studies using cross-validation alone, studies using an unseen dataset for validation usually reported lower accuracy, especially when using a local population [68]. Recent studies have addressed the risk of overfitting for models built using a single dataset [71,72] and suggested conducting model validation using an independent dataset to report the model accuracy. In the current study, we reported accuracies for both validations: validation by an independent part of the same dataset (i.e., test 1) and by an unseen independent dataset (i.e., OASIS dataset). Our proposed FFNN with VRD features yielded 85.14%, 85.71% and 78.29% in CN, MCI and AD, respectively, in the Test 1 dataset, and 83.43%, 74.86% and 88% accuracies in CN, MCI and AD, respectively, in the OASIS test dataset. Both results were satisfactory and indicate the model’s capability to generalize to new data.

4.5. Potential Clinical Application and Development of the Proposed Model

In the current study, the brain regional volumes and radiomics were retrieved from the MRI images manually using the chosen software. With the improved computer power and database management, script encoding is available. The retrieval of brain regional volumes and radiomics can be carried out after image acquisition in the image storage database. Together with the demographics obtained from the patient management system, the obtained features can be passed to the proposed neural network as the input for dementia classification. The predicted diagnosis from the neural network may help to triage AD and MCI patients from the CN and lead to a higher priority for clinicians to determine the diagnosis.

4.6. Main Findings of Study

In this study, we utilized the brain regional volumes, radiomics retrieved from MPRAGE MRI images and patients’ demographics to build a classification for dementia patients; further, we evaluated the performance of the networks built in terms of overall accuracy, subclass accuracy, sensitivity and specificity. The proposed FFNN model using all three types of features demonstrated the best distinguishing ability and achieved very good performance in dementia classification.

4.7. Study Limitations and Future Directions

In this study, two cohorts of neurodegenerative patients from a public database were used for model development and testing. The sample size was relatively small, even though it consisted of balanced samples in various groups. Small sample sizes provide less reliable estimates of the underlying data distribution, meaning that the developed model may miss subtle data patterns present in the data [70]. Further study is recommended using another independent local cohort of patients with a larger sample size to verify the proposed model.
Radiomics of MPRAGE MRI images and demographic data were used as input to develop the classification model. In future studies, the model can be improved by incorporating image features from various imaging modalities, e.g., PET/CT with 18F-Flumetemetamol as a radionuclide for an amyloid study [73], T2-weighted MRI imaging for white matter hyper-intensity [74], arterial spin labeling MRI imaging for cerebral blood flow study [75] and resting state functional MRI imaging for interhemispheric functional connectivity [76], so as to develop a more comprehensive model.
Furthermore, other clinical parameters, such as the Montreal Cognitive Assessment (MoCA) result, plasma amyloid-β level [77,78], can be included as input to develop or modify the networks, so as to improve the classification capabilities with more relevant parameters.

5. Conclusions

This study established a feed-forward neural network model by integrating image features and demographics for various stages of dementia classification. The FFNN yielded good overall accuracies for MCI, AD and CN classification, with balanced subclass accuracy, sensitivity and specificity. The proposed FFNN model is simple and can be operated using a general-purpose computer in radiology departments. The application can be used as a reliable classification tool to prioritize patients with AD or MCI from CN. It may support the triage of patient for further testing, which shortens the diagnosis confirmation pathway.

Author Contributions

Conceptualization, E.Y.W.C.; methodology, E.Y.W.C. and H.K.F.M.; software, E.Y.W.C. and R.W.K.W.; formal analysis, E.Y.W.C., R.W.K.W. and E.S.M.C.; investigation, E.Y.W.C., R.W.K.W. and E.S.M.C.; writing—original draft preparation, E.Y.W.C., R.W.K.W. and E.S.M.C.; writing—review and editing, E.Y.W.C., R.W.K.W., E.S.M.C. and H.K.F.M.; visualization, E.Y.W.C.; project administration, E.Y.W.C. and E.S.M.C.; funding acquisition, E.Y.W.C. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by a College Research Grant of Tung Wah College, grant number 2021/01. The APC was funded by a College Research Grant and the Staff Development Fund of Tung Wah College.

Institutional Review Board Statement

Approval was obtained from the ADNI Data and Publications Committee (DPC) prior to publication.

Informed Consent Statement

All patients formally consented through the participating institution during data collection.

Data Availability Statement

Publicly available datasets were analyzed in this study. The ADNI dataset can be found at: http://adni.loni.usc.edu/ (accessed on 22 January 2023). Data used in the preparation of this article were obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). As such, the investigators within the ADNI contributed to the design and implementation of ADNI and/or provided data but did not participate in the analysis or writing of this report. A complete listing of ADNI investigators can be found at: https://adni.loni.usc.edu/data-samples/access-data/#access_data (accessed on 22 January 2022). OASIS dataset: https://www.oasis-brains.org/#data. Approval was obtained on 28 February 2022 from the ADNI Data and Publications Committee (DPC) for the use of ADNI data for publication (accessed on 22 January 2023).

Acknowledgments

ADNI Dataset: Data collection and sharing for this project were funded by the Alzheimer’s Disease Neuroimaging Initiative (ADNI) (National Institutes of Health Grant U01 AG024904) and DOD ADNI (Department of Defense award number W81XWH-12-2-0012). ADNI is funded by the National Institute on Aging, the National Institute of Biomedical Imaging and Bioengineering, and through generous contributions from the following: AbbVie, Alzheimer’s Association; Alzheimer’s Drug Discovery Foundation; Araclon Biotech; BioClinica, Inc.; Biogen; Bristol-Myers Squibb Company; CereSpir, Inc.; Cogstate; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company; Euro-Immun; F. Hoffmann-La Roche Ltd. and its affiliated company Genentech, Inc.; Fujirebio; GE Healthcare; IXICO Ltd.; Janssen Alzheimer Immunotherapy Research & Development, LLC.; Johnson & Johnson Pharmaceutical Research & Development LLC.; Lumosity; Lundbeck; Merck & Co., Inc.; Meso Scale Diagnostics, LLC.; NeuroRx Research; Neurotrack Technologies; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Piramal Imaging; Servier; Takeda Pharmaceutical Company; and Transition Therapeutics. The Canadian Institutes of Health Research is providing funds to support ADNI clinical sites in Canada. Private sector contributions are facilitated by the Foundation for the National Institutes of Health (www.fnih.org). The grantee organization is the Northern California Institute for Research and Education, and the study is coordinated by the Alzheimer’s Therapeutic Research Institute at the University of Southern California. ADNI data are disseminated by the Laboratory for Neuro Imaging at the University of Southern California. OASIS Dataset: Data were provided by OASIS OASIS-3: Longitudinal Multimodal Neuroimaging: Principal Investigators: T. Benzinger, D. Marcus, J. Morris; NIH P30 AG066444, P50 AG00561, P30 NS09857781, P01 AG026276, P01 AG003991, R01 AG043434, UL1 TR000448, R01 EB009352. AV-45 doses were provided by Avid Radiopharmaceuticals, a wholly owned subsidiary of Eli Lilly. We thank the Washington University ADRC and the Conte Center for clinical assistance and participant recruitment; Elizabeth Grant for assistance with data preparation; Susan Larson, Amy Sanders, Laura Williams, Jamie Parker, and Glenn Foster for assistance with MRI data collection; Avi Snyder for development of analytic techniques; Tim Olsen, Mohana Ramaratnam, Kevin Archie, and Mikhail Milchenko for development of database and web tools. Anders Dale assisted with the original selection of imaging parameters. The acquisition of this data and support for data analysis and management were provided by NIH grants P50 AG05681, P01 AG03991, R01 AG021910, P20 MH071616, RR14075, RR 16594, BIRN002, the Alzheimer‘s Association, the James S. McDonnell Foundation, the Mental Illness and Neuroscience Discovery Institute, and the Howard Hughes Medical Institute.

Conflicts of Interest

All authors declare no conflicts of interest.

References

  1. Prince, M.; Comas-Herrera, A.; Knapp, M.; Guerchet, M.; Karagiannidou, M. World Alzheimer Report 2016; Alzheimer’s Disease International: London, UK, 2016. [Google Scholar]
  2. Esiri, M.M. Is an Effective Immune Intervention for Alzheimer’s Disease in Prospect? Trends Pharmacol. Sci. 2001, 22, 2–3. [Google Scholar] [CrossRef] [PubMed]
  3. Wang, J.; He, L.; Zheng, H.; Lu, Z.-L. Optimizing the Magnetization-Prepared Rapid Gradient-Echo (MP-RAGE) Sequence. PLoS ONE 2014, 9, e96899. [Google Scholar] [CrossRef] [PubMed]
  4. van der Flier, W.M.; van Straaten, E.C.W.; Barkhof, F.; Ferro, J.M.; Pantoni, L.; Basile, A.M.; Inzitari, D.; Erkinjuntti, T.; Wahlund, L.O.; Rostrup, E.; et al. Medial Temporal Lobe Atrophy and White Matter Hyperintensities Are Associated with Mild Cognitive Deficits in Non-Disabled Elderly People: The LADIS Study. J. Neurol. Neurosurg. Psychiatry 2005, 76, 1497–1500. [Google Scholar] [CrossRef] [PubMed]
  5. Visser, P.J.; Verhey, F.R.J.; Hofman, P.A.M.; Scheltens, P.; Jolles, J. Medial Temporal Lobe Atrophy Predicts Alzheimer’s Disease in Patients with Minor Cognitive Impairment. J. Neurol. Neurosurg. Psychiatry 2002, 72, 491–497. [Google Scholar] [CrossRef] [PubMed]
  6. Scheltens, P.; Fox, N.; Barkhof, F.; De Carli, C. Structural Magnetic Resonance Imaging in the Practical Assessment of Dementia: Beyond Exclusion. Lancet Neurol. 2002, 1, 13–21. [Google Scholar] [CrossRef] [PubMed]
  7. Glodzik-Sobanska, L.; Rusinek, H.; Mosconi, L.; Li, Y.; Zhan, J.; de Santi, S.; Convit, A.; Rich, K.; Brys, M.; de Leon, M.J. The Role of Quantitative Structural Imaging in the Early Diagnosis of Alzheimer’s DiseaseQ4. Neuroimaging Clin. N. Am. 2005, 15, 803–826. [Google Scholar] [CrossRef] [PubMed]
  8. Cheung, E.Y.; Chiu, P.K.; Kwan, J.S.; Shea, Y.F.; Mak, H. Brain Regional Volume Analysis to Differentiate Alzheimer’s Disease (AD) and Vascular Dementia (VD) from Healthy Control (HC): Machine Learning Approach. Alzheimers Dement. 2021, 17, e058343. [Google Scholar] [CrossRef]
  9. Cherbuin, N.; Anstey, K.J.; Réglade-Meslin, C.; Sachdev, P.S. In Vivo Hippocampal Measurement and Memory: A Comparison of Manual Tracing and Automated Segmentation in a Large Community-Based Sample. PLoS ONE 2009, 4, e5265. [Google Scholar] [CrossRef] [PubMed]
  10. Pardoe, H.R.; Pell, G.S.; Abbott, D.F.; Jackson, G.D. Hippocampal Volume Assessment in Temporal Lobe Epilepsy: How Good Is Automated Segmentation? Epilepsia 2009, 50, 2586–2592. [Google Scholar] [CrossRef]
  11. Qiu, Y.; Liu, S.; Hilal, S.; Loke, Y.M.; Ikram, M.K.; Xu, X.; Yeow Tan, B.; Venketasubramanian, N.; Chen, C.L.-H.; Zhou, J. Inter-Hemispheric Functional Dysconnectivity Mediates the Association of Corpus Callosum Degeneration with Memory Impairment in AD and Amnestic MCI. Sci. Rep. 2016, 6, 32573. [Google Scholar] [CrossRef]
  12. Cheung, E.Y.; Shea, Y.F.; Chiu, P.K.; Kwan, J.S.; Mak, H. The Role of Corpus Callosum Sub-regional Volume and Its Association to Interhemispheric Functional Connectivity in Alzheimer’s Disease (AD) and Amyloid Mild Cognitive Impairment (amyMCI). Alzheimers Dement. 2021, 17, e050099. [Google Scholar] [CrossRef]
  13. Saygin, Z.M.; Osher, D.E.; Augustinack, J.; Fischl, B.; Gabrieli, J.D.E. Connectivity-Based Segmentation of Human Amygdala Nuclei Using Probabilistic Tractography. NeuroImage 2011, 56, 1353–1361. [Google Scholar] [CrossRef] [PubMed]
  14. Ramani, A.; Jensen, J.H.; Helpern, J.A. Quantitative MR Imaging in Alzheimer Disease. Radiology 2006, 241, 26–44. [Google Scholar] [CrossRef] [PubMed]
  15. Moodley, K.K.; Chan, D. The Hippocampus in Neurodegenerative Disease. Front. Neurol. Neurosci. 2014, 34, 95–108. [Google Scholar] [CrossRef] [PubMed]
  16. Fh, T.; Cyw, C.; Eyw, C. Radiomics AI Prediction for Head and Neck Squamous Cell Carcinoma (HNSCC) Prognosis and Recurrence with Target Volume Approach. BJR|Open 2021, 3, 20200073. [Google Scholar] [CrossRef] [PubMed]
  17. Tang, F.-H.; Cheung, E.-Y.-W.; Wong, H.-L.; Yuen, C.-M.; Yu, M.-H.; Ho, P.-C. Radiomics from Various Tumour Volume Sizes for Prognosis Prediction of Head and Neck Squamous Cell Carcinoma: A Voted Ensemble Machine Learning Approach. Life 2022, 12, 1380. [Google Scholar] [CrossRef] [PubMed]
  18. Zhang, Y.; Oikonomou, A.; Wong, A.; Haider, M.A.; Khalvati, F. Radiomics-Based Prognosis Analysis for Non-Small Cell Lung Cancer. Sci. Rep. 2017, 7, 46349. [Google Scholar] [CrossRef] [PubMed]
  19. Zhang, L.; Dong, D.; Li, H.; Tian, J.; Ouyang, F.; Mo, X.; Zhang, B.; Luo, X.; Lian, Z.; Pei, S.; et al. Development and Validation of a Magnetic Resonance Imaging-Based Model for the Prediction of Distant Metastasis before Initial Treatment of Nasopharyngeal Carcinoma: A Retrospective Cohort Study. EBioMedicine 2019, 40, 327–335. [Google Scholar] [CrossRef] [PubMed]
  20. Das, A.K.; Bell, M.H.; Nirodi, C.S.; Story, M.D.; Minna, J.D. Radiogenomics Predicting Tumor Responses to Radiotherapy in Lung Cancer. Semin. Radiat. Oncol. 2010, 20, 149–155. [Google Scholar] [CrossRef]
  21. Li, T.-R.; Wu, Y.; Jiang, J.-J.; Lin, H.; Han, C.-L.; Jiang, J.-H.; Han, Y. Radiomics Analysis of Magnetic Resonance Imaging Facilitates the Identification of Preclinical Alzheimer’s Disease: An Exploratory Study. Front. Cell Dev. Biol. 2020, 8, 605734. [Google Scholar] [CrossRef]
  22. Hogervorst, E.; Temple, S.; O’Donnell, E. Sex Differences in Dementia. In Sex Differences in Brain Function and Dysfunction; Gibson, C., Galea, L.A.M., Eds.; Current Topics in Behavioral Neurosciences; Springer International Publishing: Cham, Switzerland, 2022; Volume 62, pp. 309–331. ISBN 978-3-031-26722-2. [Google Scholar]
  23. Dong, C.; Zhou, C.; Fu, C.; Hao, W.; Ozaki, A.; Shrestha, N.; Virani, S.S.; Mishra, S.R.; Zhu, D. Sex Differences in the Association between Cardiovascular Diseases and Dementia Subtypes: A Prospective Analysis of 464,616 UK Biobank Participants. Biol. Sex Differ. 2022, 13, 21. [Google Scholar] [CrossRef]
  24. Lopez, O.L.; Kuller, L.H. Epidemiology of Aging and Associated Cognitive Disorders: Prevalence and Incidence of Alzheimer’s Disease and Other Dementias. In Handbook of Clinical Neurology; Elsevier: Amsterdam, The Netherlands, 2019; Volume 167, pp. 139–148. ISBN 978-0-12-804766-8. [Google Scholar]
  25. Chary, E.; Amieva, H.; Pérès, K.; Orgogozo, J.; Dartigues, J.; Jacqmin-Gadda, H. Short- versus Long-term Prediction of Dementia among Subjects with Low and High Educational Levels. Alzheimers Dement. 2013, 9, 562–571. [Google Scholar] [CrossRef]
  26. Barnes, D.E.; Covinsky, K.E.; Whitmer, R.A.; Kuller, L.H.; Lopez, O.L.; Yaffe, K. Commentary on “Developing a National Strategy to Prevent Dementia: Leon Thal Symposium 2009.” Dementia Risk Indices: A Framework for Identifying Individuals with a High Dementia Risk. Alzheimers Dement. 2010, 6, 138–141. [Google Scholar] [CrossRef]
  27. Maroco, J.; Silva, D.; Rodrigues, A.; Guerreiro, M.; Santana, I.; De Mendonça, A. Data Mining Methods in the Prediction of Dementia: A Real-Data Comparison of the Accuracy, Sensitivity and Specificity of Linear Discriminant Analysis, Logistic Regression, Neural Networks, Support Vector Machines, Classification Trees and Random Forests. BMC Res. Notes 2011, 4, 299. [Google Scholar] [CrossRef]
  28. Lebedev, A.V.; Westman, E.; Van Westen, G.J.P.; Kramberger, M.G.; Lundervold, A.; Aarsland, D.; Soininen, H.; Kłoszewska, I.; Mecocci, P.; Tsolaki, M.; et al. Random Forest Ensembles for Detection and Prediction of Alzheimer’s Disease with a Good between-Cohort Robustness. NeuroImage Clin. 2014, 6, 115–125. [Google Scholar] [CrossRef]
  29. Cheung, E.Y. Radiomics Deep Learning Application to Differentiate Mild Cognitive Impairment Patients (MCI) from Healthy Control (HC). Alzheimers Dement. 2023, 19, e066747. [Google Scholar] [CrossRef]
  30. Cheung, E.Y.W.; Chau, A.C.M.; Tang, F.H.; on behalf of the Alzheimer’s Disease Neuroimaging Initiative. Radiomics-Based Artificial Intelligence Differentiation of Neurodegenerative Diseases with Reference to the Volumetry. Life 2022, 12, 514. [Google Scholar] [CrossRef]
  31. Wei, W.; Visweswaran, S.; Cooper, G.F. The Application of Naive Bayes Model Averaging to Predict Alzheimer’s Disease from Genome-Wide Data. J. Am. Med. Inform. Assoc. 2011, 18, 370–375. [Google Scholar] [CrossRef]
  32. Kruthika, K.R.; Rajeswari; Maheshappa, H.D. Multistage Classifier-Based Approach for Alzheimer’s Disease Prediction and Retrieval. Inform. Med. Unlocked 2019, 14, 34–42. [Google Scholar] [CrossRef]
  33. Seo, K.; Pan, R.; Lee, D.; Thiyyagura, P.; Chen, K. Visualizing Alzheimer’s Disease Progression in Low Dimensional Manifolds. Heliyon 2019, 5, e02216. [Google Scholar] [CrossRef]
  34. Sheng, J.; Wang, B.; Zhang, Q.; Liu, Q.; Ma, Y.; Liu, W.; Shao, M.; Chen, B. A Novel Joint HCPMMP Method for Automatically Classifying Alzheimer’s and Different Stage MCI Patients. Behav. Brain Res. 2019, 365, 210–221. [Google Scholar] [CrossRef]
  35. Alzheimer’s Disease Neuroimaging Initiative; Yamashita, A.Y.; Falcão, A.X.; Leite, N.J. The Residual Center of Mass: An Image Descriptor for the Diagnosis of Alzheimer Disease. Neuroinformatics 2019, 17, 307–321. [Google Scholar] [CrossRef]
  36. Yang, W.; Chen, X.; Cohen, D.S.; Rosin, E.R.; Toga, A.W.; Thompson, P.M.; Huang, X. Classification of MRI and Psychological Testing Data Based on Support Vector Machine. Int. J. Clin. Exp. Med. 2017, 10, 16004–16026. [Google Scholar]
  37. Zhao, W.; Luo, Y.; Zhao, L.; Mok, V.; Su, L.; Yin, C.; Sun, Y.; Lu, J.; Shi, L.; Han, Y. Automated Brain MRI Volumetry Differentiates Early Stages of Alzheimer’s Disease from Normal Aging. J. Geriatr. Psychiatry Neurol. 2019, 32, 354–364. [Google Scholar] [CrossRef]
  38. Zheng, Y.; Guo, H.; Zhang, L.; Wu, J.; Li, Q.; Lv, F. Machine Learning-Based Framework for Differential Diagnosis Between Vascular Dementia and Alzheimer’s Disease Using Structural MRI Features. Front. Neurol. 2019, 10, 1097. [Google Scholar] [CrossRef]
  39. Danso, S.O.; Zeng, Z.; Muniz-Terrera, G.; Ritchie, C.W. Developing an Explainable Machine Learning-Based Personalised Dementia Risk Prediction Model: A Transfer Learning Approach with Ensemble Learning Algorithms. Front. Big Data 2021, 4, 613047. [Google Scholar] [CrossRef]
  40. Suk, H.-I. An Introduction to Neural Networks and Deep Learning. In Deep Learning for Medical Image Analysis; Elsevier: Amsterdam, The Netherlands, 2017; pp. 3–24. ISBN 978-0-12-810408-8. [Google Scholar]
  41. Cheung, E.Y.; Mak, H.; Chiu, P.K.; Shea, Y.F.; Chau, A.C.; Kwan, J.S. Artificial Neural Network Application to Characterize Patients with Alzheimer’s Disease and Mild Cognitive Impairment from Cognitive Normal. Alzheimers Dement. 2023, 19, e079624. [Google Scholar] [CrossRef]
  42. Cárdenas-Peña, D.; Collazos-Huertas, D.; Castellanos-Dominguez, G. Centered Kernel Alignment Enhancing Neural Network Pretraining for MRI-Based Dementia Diagnosis. Comput. Math. Methods Med. 2016, 2016, 9523849. [Google Scholar] [CrossRef]
  43. Ning, K.; Chen, B.; Sun, F.; Hobel, Z.; Zhao, L.; Matloff, W.; Toga, A.W. Classifying Alzheimer’s Disease with Brain Imaging and Genetic Data Using a Neural Network Framework. Neurobiol. Aging 2018, 68, 151–158. [Google Scholar] [CrossRef]
  44. Cohen, D.S.; Carpenter, K.A.; Jarrell, J.T.; Huang, X.; Alzheimer’s Disease Neuroimaging Initiative. Deep Learning-Based Classification of Multi-Categorical Alzheimer’s Disease Data. Curr. Neurobiol. 2019, 10, 141–147. [Google Scholar]
  45. Frizzell, T.O.; Glashutter, M.; Liu, C.C.; Zeng, A.; Pan, D.; Hajra, S.G.; D’Arcy, R.C.N.; Song, X. Artificial Intelligence in Brain MRI Analysis of Alzheimer’s Disease over the Past 12 Years: A Systematic Review. Ageing Res. Rev. 2022, 77, 101614. [Google Scholar] [CrossRef]
  46. Jack, C.R.; Bernstein, M.A.; Fox, N.C.; Thompson, P.; Alexander, G.; Harvey, D.; Borowski, B.; Britson, P.J.; Whitwell, J.L.; Ward, C.; et al. The Alzheimer’s Disease Neuroimaging Initiative (ADNI): MRI Methods. J. Magn. Reson. Imaging 2008, 27, 685–691. [Google Scholar] [CrossRef]
  47. LaMontagne, P.J.; Benzinger, T.L.S.; Morris, J.C.; Keefe, S.; Hornbeck, R.; Xiong, C.; Grant, E.; Hassenstab, J.; Moulder, K.; Vlassenko, A.G.; et al. OASIS-3: Longitudinal Neuroimaging, Clinical, and Cognitive Dataset for Normal Aging and Alzheimer Disease. medRxiv, 2019; preprint. [Google Scholar]
  48. Dale, A.M.; Fischl, B.; Sereno, M.I. Cortical Surface-Based Analysis. I. Segmentation and Surface Reconstruction. Neuroimage 1999, 9, 179–194. [Google Scholar] [CrossRef]
  49. Fischl, B.; Liu, A.; Dale, A.M. Automated Manifold Surgery: Constructing Geometrically Accurate and Topologically Correct Models of the Human Cerebral Cortex. IEEE Trans. Med. Imaging 2001, 20, 70–80. [Google Scholar] [CrossRef]
  50. Fischl, B.; Salat, D.H.; van der Kouwe, A.J.W.; Makris, N.; Ségonne, F.; Quinn, B.T.; Dale, A.M. Sequence-Independent Segmentation of Magnetic Resonance Images. NeuroImage 2004, 23, S69–S84. [Google Scholar] [CrossRef]
  51. Fischl, B.; Sereno, M.I.; Dale, A.M. Cortical Surface-Based Analysis. II: Inflation, Flattening, and a Surface-Based Coordinate System. Neuroimage 1999, 9, 195–207. [Google Scholar] [CrossRef]
  52. Jovicich, J.; Czanner, S.; Greve, D.; Haley, E.; van der Kouwe, A.; Gollub, R.; Kennedy, D.; Schmitt, F.; Brown, G.; Macfall, J.; et al. Reliability in Multi-Site Structural MRI Studies: Effects of Gradient Non-Linearity Correction on Phantom and Human Data. Neuroimage 2006, 30, 436–443. [Google Scholar] [CrossRef]
  53. Reuter, M.; Rosas, H.D.; Fischl, B. Highly Accurate Inverse Consistent Registration: A Robust Approach. Neuroimage 2010, 53, 1181–1196. [Google Scholar] [CrossRef]
  54. Reuter, M.; Schmansky, N.J.; Rosas, H.D.; Fischl, B. Within-Subject Template Estimation for Unbiased Longitudinal Image Analysis. Neuroimage 2012, 61, 1402–1418. [Google Scholar] [CrossRef]
  55. Friston, K.J. (Ed.) Statistical Parametric Mapping: The Analysis of Funtional Brain Images, 1st ed.; Elsevier: Amsterdam, The Netherlands; Academic Press: Boston, MA, USA, 2007; ISBN 978-0-12-372560-8. [Google Scholar]
  56. Aerts, H.J.W.L.; Velazquez, E.R.; Leijenaar, R.T.H.; Parmar, C.; Grossmann, P.; Carvalho, S.; Cavalho, S.; Bussink, J.; Monshouwer, R.; Haibe-Kains, B.; et al. Decoding Tumour Phenotype by Noninvasive Imaging Using a Quantitative Radiomics Approach. Nat. Commun. 2014, 5, 4006. [Google Scholar] [CrossRef]
  57. Zwanenburg, A.; Vallières, M.; Abdalah, M.A.; Aerts, H.J.W.L.; Andrearczyk, V.; Apte, A.; Ashrafinia, S.; Bakas, S.; Beukinga, R.J.; Boellaard, R.; et al. The Image Biomarker Standardization Initiative: Standardized Quantitative Radiomics for High-Throughput Image-Based Phenotyping. Radiology 2020, 295, 328–338. [Google Scholar] [CrossRef]
  58. Nuti, G.; Jiménez Rugama, L.A.; Cross, A.-I. An Explainable Bayesian Decision Tree Algorithm. Front. Appl. Math. Stat. 2021, 7, 598833. [Google Scholar] [CrossRef]
  59. Cross-Validation: Evaluating Estimator Performance. Available online: https://scikit-learn/stable/modules/cross_validation.html (accessed on 8 October 2022).
  60. de Flores, R.; La Joie, R.; Chételat, G. Structural Imaging of Hippocampal Subfields in Healthy Aging and Alzheimer’s Disease. Neuroscience 2015, 309, 29–50. [Google Scholar] [CrossRef]
  61. Sørensen, L.; Igel, C.; Pai, A.; Balas, I.; Anker, C.; Lillholm, M.; Nielsen, M. Differential Diagnosis of Mild Cognitive Impairment and Alzheimer’s Disease Using Structural MRI Cortical Thickness, Hippocampal Shape, Hippocampal Texture, and Volumetry. NeuroImage Clin. 2017, 13, 470–482. [Google Scholar] [CrossRef]
  62. Koikkalainen, J.; Rhodius-Meester, H.; Tolonen, A.; Barkhof, F.; Tijms, B.; Lemstra, A.W.; Tong, T.; Guerrero, R.; Schuh, A.; Ledig, C.; et al. Differential Diagnosis of Neurodegenerative Diseases Using Structural MRI Data. Neuroimage Clin. 2016, 11, 435–449. [Google Scholar] [CrossRef]
  63. Du, Y.; Zhang, S.; Fang, Y.; Qiu, Q.; Zhao, L.; Wei, W.; Tang, Y.; Li, X. Radiomic Features of the Hippocampus for Diagnosing Early-Onset and Late-Onset Alzheimer’s Disease. Front. Aging Neurosci. 2022, 13, 789099. [Google Scholar] [CrossRef]
  64. Zell, A. Simulation Neuronaler Netze, 1st ed.; Addison-Wesley: Bonn, Germany, 1996; ISBN 978-3-89319-554-1. [Google Scholar]
  65. Gavin, H.P. The Levenberg-Marquardt Algorithm for Nonlinear Least Squares Curve-Fitting Problems; Department of Civil and Environmental Engineering, Duke University: Durham, NC, USA, 2019; 19p. [Google Scholar]
  66. Zhang, F.; Tian, S.; Chen, S.; Ma, Y.; Li, X.; Guo, X. Voxel-Based Morphometry: Improving the Diagnosis of Alzheimer’s Disease Based on an Extreme Learning Machine Method from the ADNI Cohort. Neuroscience 2019, 414, 273–279. [Google Scholar] [CrossRef]
  67. Mendoza-Léon, R.; Puentes, J.; Uriza, L.F.; Hernández Hoyos, M. Single-Slice Alzheimer’s Disease Classification and Disease Regional Analysis with Supervised Switching Autoencoders. Comput. Biol. Med. 2020, 116, 103527. [Google Scholar] [CrossRef]
  68. Borchert, R.J.; Azevedo, T.; Badhwar, A.; Bernal, J.; Betts, M.; Bruffaerts, R.; Burkhart, M.C.; Dewachter, I.; Gellersen, H.M.; Low, A.; et al. Artificial Intelligence for Diagnostic and Prognostic Neuroimaging in Dementia: A Systematic Review. Alzheimers Dement. 2023, 19, 5885–5904. [Google Scholar] [CrossRef] [PubMed]
  69. Moore, P.J.; Lyons, T.J.; Gallacher, J.; for the Alzheimer’s Disease Neuroimaging Initiative. Random Forest Prediction of Alzheimer’s Disease Using Pairwise Selection from Time Series Data. PLoS ONE 2019, 14, e0211558. [Google Scholar] [CrossRef]
  70. Bucholc, M.; James, C.; Khleifat, A.A.; Badhwar, A.; Clarke, N.; Dehsarvi, A.; Madan, C.R.; Marzi, S.J.; Shand, C.; Schilder, B.M.; et al. Artificial Intelligence for Dementia Research Methods Optimization. Alzheimers Dement. 2023, 19, 5934–5951. [Google Scholar] [CrossRef]
  71. Sørensen, L.; Nielsen, M. Ensemble Support Vector Machine Classification of Dementia Using Structural MRI and Mini-Mental State Examination. J. Neurosci. Methods 2018, 302, 66–74. [Google Scholar] [CrossRef]
  72. Qiu, S.; Joshi, P.S.; Miller, M.I.; Xue, C.; Zhou, X.; Karjadi, C.; Chang, G.H.; Joshi, A.S.; Dwyer, B.; Zhu, S.; et al. Development and Validation of an Interpretable Deep Learning Framework for Alzheimer’s Disease Classification. Brain 2020, 143, 1920–1933. [Google Scholar] [CrossRef]
  73. Cheung, E.Y.W.; Chau, A.C.M.; Shea, Y.-F.; Chiu, P.K.C.; Kwan, J.S.K.; Mak, H.K.F. Level of Amyloid-β (Aβ) Binding Leading to Differential Effects on Resting State Functional Connectivity in Major Brain Networks. Biomedicines 2022, 10, 2321. [Google Scholar] [CrossRef] [PubMed]
  74. Shu, Z.-Y.; Shao, Y.; Xu, Y.-Y.; Ye, Q.; Cui, S.-J.; Mao, D.-W.; Pang, P.-P.; Gong, X.-Y. Radiomics Nomogram Based on MRI for Predicting White Matter Hyperintensity Progression in Elderly Adults. J. Magn. Reson. Imaging 2020, 51, 535–546. [Google Scholar] [CrossRef]
  75. Cheung, E.Y.; Shea, Y.F.; Chiu, P.K.; Kwan, J.S.; Mak, H. Aberrant Interhemispheric Functional Connectivity and Cerebral Blood Flow in Vascular Dementia (VD) and Non-amyloid Mild Cognitive Impairment (NamyMCI). Alzheimers Dement. 2021, 17, e049381. [Google Scholar] [CrossRef]
  76. Cheung, E.Y.W.; Shea, Y.F.; Chiu, P.K.C.; Kwan, J.S.K.; Mak, H.K.F. Diagnostic Efficacy of Voxel-Mirrored Homotopic Connectivity in Vascular Dementia as Compared to Alzheimer’s Related Neurodegenerative Diseases—A Resting State fMRI Study. Life 2021, 11, 1108. [Google Scholar] [CrossRef] [PubMed]
  77. Hampel, H.; Hardy, J.; Blennow, K.; Chen, C.; Perry, G.; Kim, S.H.; Villemagne, V.L.; Aisen, P.; Vendruscolo, M.; Iwatsubo, T.; et al. The Amyloid-β Pathway in Alzheimer’s Disease. Mol. Psychiatry 2021, 26, 5481–5503. [Google Scholar] [CrossRef]
  78. Gabelle, A.; Richard, F.; Gutierrez, L.-A.; Schraen, S.; Delva, F.; Rouaud, O.; Buée, L.; Dartigues, J.-F.; Touchon, J.; Lambert, J.-C.; et al. Plasma Amyloid-β Levels and Prognosis in Incident Dementia Cases of the 3-City Study. JAD 2012, 33, 381–391. [Google Scholar] [CrossRef]
Biomedicines 12 00896 g001
Figure 1. Extraction of 45 brain regional volumes in the software. Inf: inferior; Lat: lateral; Mid: middle; DC: diencephalon; WM: white matter; CC: corpus callosum.
Figure 1. Extraction of 45 brain regional volumes in the software. Inf: inferior; Lat: lateral; Mid: middle; DC: diencephalon; WM: white matter; CC: corpus callosum.
Biomedicines 12 00896 g001
Biomedicines 12 00896 g002
Figure 2. Individualized whole-brain mask (the green region) was used to quantify the whole brain for the retrieval of 107 radiomics features.
Figure 2. Individualized whole-brain mask (the green region) was used to quantify the whole brain for the retrieval of 107 radiomics features.
Biomedicines 12 00896 g002
Biomedicines 12 00896 g003
Figure 3. Details of the feed-forward neural network.
Figure 3. Details of the feed-forward neural network.
Biomedicines 12 00896 g003
Table 1. Images collected from ADNI and OASIS databases.
Table 1. Images collected from ADNI and OASIS databases.
ADNI Dataset
21 Centers
for Training		ADNI Dataset
4 Centers
for Testing		OASIS Dataset
for Testing
Alzheimer’s Disease692828
Mild Cognitive Decline2029191
Health Control1355757
Total406176176
Table 2. Details for the 45 brain regional volumes.
Table 2. Details for the 45 brain regional volumes.
45 Brain Regional Volumes Segmented by FreeSurfer
Left-Lateral-VentricleRight-Lateral-VentricleCSF
Left-Inf-Lat-VentRight-Inf-Lat-VentThird-Ventricle
Left-Cerebellum-White-MatterRight-Cerebellum-White-MatterForth-Ventricle
Left-Cerebellum-CortexRight-Cerebellum-CortexFifth-Ventricle
Left-ThalamusRight-ThalamusBrain-Stem
Left-CaudateRight-CaudateWM-hypointensities
Left-PutamenRight-Putamennon-WM-hypointensities
Left-PallidumRight-PallidumOptic-Chiasm
Left-HippocampusRight-HippocampusCC_Posterior
Left-AmygdalaRight-AmygdalaCC_Mid_Posterior
Left-Accumbens-areaRight-Accumbens-areaCC_Central
Left-Ventral DCRight-Ventral DCCC_Mid_Anterior
Left-vesselRight-vesselCC_Anterior
Left-choroid-plexusRight-choroid-plexus
Left-WM-hypointensitiesRight-WM-hypointensities
Left-non-WM-hypointensitiesRight-non-WM-hypointensities
Inf: inferior; Lat: lateral; Mid: middle; DC: diencephalon; WM: white matter; CC: corpus callosum.
Table 3. Eight classes of radiomics features.
Table 3. Eight classes of radiomics features.
Radiomics FeaturesNo. of Features
First-order statistics14
2D-shaped based features9
3D-shaped based features13
Gray-level co-occurrence matrix (GLCM)22
Gray-level run length matrix (GLRLM)16
Gray-level size zone matrix (GLSZM)16
Gray-level dependence matrix (GLDM)12
Neighboring gray tone difference matrix (NGTDM)5
Total107
Table 4. Integration of patients’ demographics and image features.
Table 4. Integration of patients’ demographics and image features.
Radiomics
107 Features		Volumes
45 Features		Patients’ Demographics
2 Features		Total Number of Features
R only✔ 107
RD✔ ✔109
V only ✔ 45
VD ✔✔47
VRD✔✔✔154
Table 5. Demographics of the ADNI and OASIS datasets.
Table 5. Demographics of the ADNI and OASIS datasets.
ADNI
21 Centers
for Training		ADNI
4 Centers
for Testing		Oasis Dataset
for Testing
Age range55–9065–9074–89
Sex Ratio (M:F)205:20199:7792:84
Alzheimer’s Disease692828
Mild Cognitive Decline2029191
Health Control1355757
Total406176176
Table 6. Various features employed for model building using 4 algorithms. Training was the result when building the model using patients from ADNI 21 centers; Test 1 was the result when testing the model by patients from ADNI 4 centers; Oasis was the result when testing the model by patients from OASIS database. The red fonts highlight results over 70%. (a) Various features employed for model building using SVM. (b) Various features employed for model building using ensemble classifier (EC). (c) Various features employed for model building using decision tree (DT). (d) Various features employed for model building using feed-forward neural network (FFNN).
Table 6. Various features employed for model building using 4 algorithms. Training was the result when building the model using patients from ADNI 21 centers; Test 1 was the result when testing the model by patients from ADNI 4 centers; Oasis was the result when testing the model by patients from OASIS database. The red fonts highlight results over 70%. (a) Various features employed for model building using SVM. (b) Various features employed for model building using ensemble classifier (EC). (c) Various features employed for model building using decision tree (DT). (d) Various features employed for model building using feed-forward neural network (FFNN).
(a) Various features employed for model building using SVM.
SVM MCIADCN
Overall
Accuracy		AccuracySensitivitySpecificityPrecisionF1 ScoreAccuracySensitivitySpecificityPrecisionF1 ScoreAccuracySensitivitySpecificityPrecisionF1 Score
R onlyTrain67.74%73.45%70.85%76.67%79.00%74.70%83.62%52.63%88.73%43.48%47.62%78.41%69.11%82.50%63.43%66.15%
Test140.57%56.00%56.99%54.88%58.89%57.92%74.29%9.52%83.12%7.14%8.16%50.86%26.23%64.04%28.07%27.12%
Oasis35.43%52.57%58.54%50.75%26.67%36.64%66.29%10.26%82.35%14.29%11.94%52.00%35.79%71.25%59.65%44.74%
RDTrain62.03%68.49%63.88%77.14%84.00%72.57%81.14%33.33%83.77%10.14%15.56%74.44%63.03%79.23%55.97%59.29%
Test160.00%66.86%63.11%75.47%85.56%72.64%81.71%0.00%83.63%0.00%#DIV/0!71.43%57.14%76.98%49.12%52.83%
Oasis54.29%61.14%57.05%94.74%98.89%72.36%83.43%0.00%83.91%0.00%#DIV/0!64.00%33.33%67.52%10.53%16.00%
V onlyTrain98.26%98.51%98.02%99.00%99.00%98.51%98.76%97.06%99.10%95.65%96.35%99.26%99.25%99.26%98.51%98.88%
Test170.29%75.43%75.82%75.00%76.67%76.24%84.57%60.00%85.29%10.71%18.18%80.57%64.56%93.75%89.47%75.00%
Oasis61.14%66.29%68.24%64.44%64.44%66.29%76.57%19.05%84.42%14.29%16.33%79.43%65.22%88.68%78.95%71.43%
VDTrain94.79%96.28%93.02%100.00%100.00%96.39%96.03%100.00%95.43%76.81%86.89%97.27%95.56%98.13%96.27%95.91%
Test171.43%74.86%70.18%83.61%88.89%78.43%84.00%#DIV/0!84.00%0.00%#DIV/0!84.00%73.77%89.47%78.95%76.27%
Oasis71.43%71.43%67.24%79.66%86.67%75.73%84.00%#DIV/0!84.00%0.00%#DIV/0!87.43%79.66%91.38%82.46%81.03%
VRDTrain81.14%83.62%78.15%91.52%93.00%84.93%89.58%100.00%88.83%39.13%56.25%89.08%82.61%92.45%85.07%83.82%
Test171.43%73.14%66.93%89.58%94.44%78.34%82.86%0.00%83.82%0.00%#DIV/0!86.86%86.96%86.82%70.18%77.67%
Oasis68.00%70.29%67.27%75.38%82.22%74.00%80.00%23.08%84.57%10.71%14.63%85.71%80.77%87.80%73.68%77.06%
(b) Various features employed for models building using ensemble classifier (EC).
Ensemble MCIADCN
Overall
Accuracy		AccuracySensitivitySpecificityPrecisionF1 ScoreAccuracySensitivitySpecificityPrecisionF1 ScoreAccuracySensitivitySpecificityPrecisionF1 Score
R onlyTrain64.52%68.98%64.71%76.35%82.50%72.53%83.62%54.84%86.02%24.64%34.00%76.43%66.67%80.42%58.21%62.15%
Test149.14%58.86%57.14%63.27%80.00%66.67%78.29%8.33%83.44%3.57%5.00%61.14%35.14%68.12%22.81%27.66%
Oasis58.29%65.14%62.39%70.69%81.11%70.53%81.14%14.29%83.93%3.57%5.71%70.29%54.90%76.61%49.12%51.85%
RDTrain67.49%70.97%66.27%79.05%84.50%74.29%83.62%54.05%86.61%28.99%37.74%80.40%74.77%82.53%61.94%67.76%
Test153.14%61.71%58.91%69.57%84.44%69.41%81.14%22.22%84.34%7.14%10.81%63.43%40.54%69.57%26.32%31.91%
Oasis52.00%60.57%63.64%58.16%54.44%58.68%74.29%20.69%84.93%21.43%21.05%69.14%52.17%80.19%63.16%57.14%
V onlyTrain96.28%96.28%94.26%98.45%98.50%96.33%97.52%96.83%97.65%88.41%92.42%98.76%99.24%98.53%97.01%98.11%
Test173.14%73.14%69.03%80.65%86.67%76.85%82.86%0.00%83.82%0.00%#DIV/0!90.29%83.33%93.91%87.72%85.47%
Oasis68.00%69.71%68.32%71.62%76.67%72.25%82.29%36.36%85.37%14.29%20.51%84.00%73.02%90.18%80.70%76.67%
VDTrain95.29%96.28%93.84%98.96%99.00%96.35%96.28%100.00%95.70%78.26%87.80%98.01%95.65%99.25%98.51%97.06%
Test177.91%77.91%73.21%86.67%91.11%81.19%83.72%#DIV/0!83.72%0.00%#DIV/0!94.19%86.67%98.21%96.30%91.23%
Oasis70.86%70.86%67.57%76.56%83.33%74.63%84.57%57.14%85.71%14.29%22.86%86.29%78.95%89.83%78.95%78.95%
VRDTrain93.55%94.29%91.94%96.88%97.00%94.40%96.53%100.00%95.98%79.71%88.71%96.28%93.43%97.74%95.52%94.46%
Test174.86%76.00%70.00%89.09%93.33%80.00%84.00%50.00%84.39%3.57%6.67%89.71%86.79%90.98%80.70%83.64%
Oasis69.71%72.00%74.12%70.00%70.00%72.00%80.57%40.00%88.97%42.86%41.38%86.86%78.33%91.30%82.46%80.34%
(c) Various features employed for model building using decision tree (DT).
Decision Tree MCIADCN
Overall
Accuracy		AccuracySensitivitySpecificityPrecisionF1 ScoreAccuracySensitivitySpecificityPrecisionF1 ScoreAccuracySensitivitySpecificityPrecisionF1 Score
R onlyTrain58.31%65.51%62.87%69.28%74.50%68.19%80.15%36.59%85.08%21.74%27.27%70.97%56.80%77.34%52.99%54.83%
Test147.43%54.86%54.55%55.56%73.33%62.56%76.00%20.83%84.77%17.86%19.23%64.00%40.00%68.97%21.05%27.59%
Oasis54.86%56.00%53.99%83.33%97.78%69.57%85.14%100.00%84.97%7.14%13.33%68.57%60.00%69.09%10.53%17.91%
RDTrain62.78%70.22%65.50%78.62%84.50%73.80%82.63%46.15%83.85%8.70%14.63%72.70%59.09%79.34%58.21%58.65%
Test157.71%64.00%62.86%65.71%73.33%67.69%81.14%22.22%84.34%7.14%10.81%70.29%54.10%78.95%57.89%55.93%
Oasis57.14%61.14%58.09%71.79%87.78%69.91%82.29%36.36%85.37%14.29%20.51%70.86%60.71%72.79%29.82%40.00%
V onlyTrain82.63%87.34%87.44%87.25%87.00%87.22%88.83%65.00%94.74%75.36%69.80%89.08%86.29%90.32%79.85%82.95%
Test162.86%70.86%68.93%73.61%78.89%73.58%78.29%25.00%85.16%17.86%20.83%76.57%65.38%81.30%59.65%62.39%
Oasis64.57%70.86%70.10%71.79%75.56%72.73%78.29%22.22%84.71%14.29%17.39%80.00%68.33%86.09%71.93%70.09%
VDTrain84.12%86.85%86.57%87.13%87.00%86.78%88.34%68.33%91.84%59.42%63.57%93.05%87.32%96.17%92.54%89.86%
Test170.86%74.86%71.70%79.71%84.44%77.55%84.00%50.00%86.96%25.00%33.33%82.86%74.55%86.67%71.93%73.21%
Oasis70.29%74.29%70.64%80.30%85.56%77.39%82.29%41.18%86.71%25.00%31.11%84.00%79.59%85.71%68.42%73.58%
VRDTrain71.22%73.70%68.65%82.12%86.50%76.55%84.86%66.67%86.02%23.19%34.41%83.87%77.17%86.96%73.13%75.10%
Test175.43%75.43%69.11%90.38%94.44%79.81%86.29%83.33%86.39%17.86%29.41%89.14%91.30%88.37%73.68%81.55%
Oasis72.00%75.43%73.27%78.38%82.22%77.49%85.14%55.00%89.03%39.29%45.83%83.43%75.93%86.78%71.93%73.87%
(d) Various features employed for model building using feed-forward neural network (FFNN).
Feed Forward Neural Network MCIADCN
Overall
Accuracy		AccuracySensitivitySpecificityPrecisionF1 ScoreAccuracySensitivitySpecificityPrecisionF1 ScoreAccuracySensitivitySpecificityPrecisionF1 Score
R onlyTrain73.55%75.76%70.00%85.71%89.44%78.54%85.95%78.95%86.34%24.19%37.04%85.40%79.82%87.95%75.21%77.45%
Test151.43%58.29%57.26%60.34%74.44%64.73%82.86%0.00%83.82%0.00%#DIV/0!61.71%41.07%71.43%40.35%40.71%
Oasis45.71%54.86%57.14%53.06%48.89%52.69%82.29%28.57%84.52%7.14%11.43%54.29%37.36%72.62%59.65%45.95%
RDTrain92.01%92.29%93.33%91.26%91.30%92.31%96.14%87.10%98.01%90.00%88.52%95.59%92.56%97.11%94.12%93.33%
Test155.43%64.57%62.96%67.16%75.56%68.69%76.57%21.74%84.87%17.86%19.61%69.71%54.55%74.81%42.11%47.52%
Oasis55.43%68.57%73.97%64.71%60.00%66.26%74.86%16.67%84.11%14.29%15.38%67.43%50.00%81.44%68.42%57.78%
V onlyTrain100.00%100.00%100.00%100.00%100.00%100.00%100.00%100.00%100.00%100.00%100.00%100.00%100.00%100.00%100.00%100.00%
Test161.14%64.57%65.56%63.53%65.56%65.56%79.43%16.67%84.05%7.14%10.00%78.29%63.01%89.22%80.70%70.77%
Oasis48.57%49.71%51.19%48.35%47.78%49.43%79.43%27.78%85.35%17.86%21.74%68.00%50.68%80.39%64.91%56.92%
VDTrain99.72%99.72%99.44%100.00%100.00%99.72%99.72%100.00%99.67%98.41%99.20%100.00%100.00%100.00%100.00%100.00%
Test165.71%67.43%67.74%67.07%70.00%68.85%77.14%22.73%84.97%17.86%20.00%86.86%78.33%91.30%82.46%80.34%
Oasis54.29%54.86%55.24%54.29%64.44%59.49%77.71%13.33%83.75%7.14%9.30%76.00%63.64%81.67%61.40%62.50%
VRDTrain99.72%99.72%99.44%100.00%100.00%99.72%100.00%100.00%100.00%100.00%100.00%99.72%100.00%99.57%99.18%99.59%
Test176.57%78.29%76.53%80.52%83.33%79.79%89.71%72.73%92.16%57.14%64.00%85.14%78.18%88.33%75.44%76.79%
Oasis73.14%74.86%78.75%71.58%70.00%74.12%88.00%62.07%93.15%64.29%63.16%83.43%71.21%90.83%82.46%76.42%
Table 7. Five groups of features employed for model building using 4 algorithms. Training was the result when building the model using patients from ADNI 21 centers; Test 1 was the result when testing the model by patients from ADNI 4 centers; Oasis was the result when testing the model by patients from OASIS database. The red fonts highlighted the result over 70%. (a) Model performance using radiomics only in various model-building algorithms. (b) Model performance using RD in various model-building algorithms. (c) Model performance using volumes only in various model-building algorithms. (d) Model performance using VD in various model-building algorithms. (e) Model performance using VRD in various model-building algorithms.
Table 7. Five groups of features employed for model building using 4 algorithms. Training was the result when building the model using patients from ADNI 21 centers; Test 1 was the result when testing the model by patients from ADNI 4 centers; Oasis was the result when testing the model by patients from OASIS database. The red fonts highlighted the result over 70%. (a) Model performance using radiomics only in various model-building algorithms. (b) Model performance using RD in various model-building algorithms. (c) Model performance using volumes only in various model-building algorithms. (d) Model performance using VD in various model-building algorithms. (e) Model performance using VRD in various model-building algorithms.
(a) Model performance using radiomics only in various model-building algorithms
R Only MCIADCN
Overall
Accuracy		AccuracySensitivitySpecificityPrecisionF1 ScoreAccuracySensitivitySpecificityPrecisionF1 ScoreAccuracySensitivitySpecificityPrecisionF1 Score
SVMTrain67.74%73.45%70.85%76.67%79.00%74.70%83.62%52.63%88.73%43.48%47.62%78.41%69.11%82.50%63.43%66.15%
Test140.57%56.00%56.99%54.88%58.89%57.92%74.29%9.52%83.12%7.14%8.16%50.86%26.23%64.04%28.07%27.12%
Oasis35.43%52.57%58.54%50.75%26.67%36.64%66.29%10.26%82.35%14.29%11.94%52.00%35.79%71.25%59.65%44.74%
ECTrain64.52%68.98%64.71%76.35%82.50%72.53%83.62%54.84%86.02%24.64%34.00%76.43%66.67%80.42%58.21%62.15%
Test149.14%58.86%57.14%63.27%80.00%66.67%78.29%8.33%83.44%3.57%5.00%61.14%35.14%68.12%22.81%27.66%
Oasis58.29%65.14%62.39%70.69%81.11%70.53%81.14%14.29%83.93%3.57%5.71%70.29%54.90%76.61%49.12%51.85%
DTTrain58.31%65.51%62.87%69.28%74.50%68.19%80.15%36.59%85.08%21.74%27.27%70.97%56.80%77.34%52.99%54.83%
Test147.43%54.86%54.55%55.56%73.33%62.56%76.00%20.83%84.77%17.86%19.23%64.00%40.00%68.97%21.05%27.59%
Oasis54.86%56.00%53.99%83.33%97.78%69.57%85.14%100.00%84.97%7.14%13.33%68.57%60.00%69.09%10.53%17.91%
FFNNTrain73.55%75.76%70.00%85.71%89.44%78.54%85.95%78.95%86.34%24.19%37.04%85.40%79.82%87.95%75.21%77.45%
Test151.43%58.29%57.26%60.34%74.44%64.73%82.86%0.00%83.82%0.00%#DIV/0!61.71%41.07%71.43%40.35%40.71%
Oasis45.71%54.86%57.14%53.06%48.89%52.69%82.29%28.57%84.52%7.14%11.43%54.29%37.36%72.62%59.65%45.95%
(b) Model performance using RD in various model-building algorithms
RD MCIADCN
Overall
Accuracy		AccuracySensitivitySpecificityPrecisionF1 ScoreAccuracySensitivitySpecificityPrecisionF1 ScoreAccuracySensitivitySpecificityPrecisionF1 Score
SVMTrain62.03%68.49%63.88%77.14%84.00%72.57%81.14%33.33%83.77%10.14%15.56%74.44%63.03%79.23%55.97%59.29%
Test160.00%66.86%63.11%75.47%85.56%72.64%81.71%0.00%83.63%0.00%#DIV/0!71.43%57.14%76.98%49.12%52.83%
Oasis54.29%61.14%57.05%94.74%98.89%72.36%83.43%0.00%83.91%0.00%#DIV/0!64.00%33.33%67.52%10.53%16.00%
ECTrain67.49%70.97%66.27%79.05%84.50%74.29%83.62%54.05%86.61%28.99%37.74%80.40%74.77%82.53%61.94%67.76%
Test153.14%61.71%58.91%69.57%84.44%69.41%81.14%22.22%84.34%7.14%10.81%63.43%40.54%69.57%26.32%31.91%
Oasis52.00%60.57%63.64%58.16%54.44%58.68%74.29%20.69%84.93%21.43%21.05%69.14%52.17%80.19%63.16%57.14%
DTTrain62.78%70.22%65.50%78.62%84.50%73.80%82.63%46.15%83.85%8.70%14.63%72.70%59.09%79.34%58.21%58.65%
Test157.71%64.00%62.86%65.71%73.33%67.69%81.14%22.22%84.34%7.14%10.81%70.29%54.10%78.95%57.89%55.93%
Oasis57.14%61.14%58.09%71.79%87.78%69.91%82.29%36.36%85.37%14.29%20.51%70.86%60.71%72.79%29.82%40.00%
FFNNTrain92.01%92.29%93.33%91.26%91.30%92.31%96.14%87.10%98.01%90.00%88.52%95.59%92.56%97.11%94.12%93.33%
Test155.43%64.57%62.96%67.16%75.56%68.69%76.57%21.74%84.87%17.86%19.61%69.71%54.55%74.81%42.11%47.52%
Oasis55.43%68.57%73.97%64.71%60.00%66.26%74.86%16.67%84.11%14.29%15.38%67.43%50.00%81.44%68.42%57.78%
(c) Model performance using volumes only in various model-building algorithms
V only MCIADCN
Overall
Accuracy		AccuracySensitivitySpecificityPrecisionF1 ScoreAccuracySensitivitySpecificityPrecisionF1 ScoreAccuracySensitivitySpecificityPrecisionF1 Score
SVMTrain98.26%98.51%98.02%99.00%99.00%98.51%98.76%97.06%99.10%95.65%96.35%99.26%99.25%99.26%98.51%98.88%
Test170.29%75.43%75.82%75.00%76.67%76.24%84.57%60.00%85.29%10.71%18.18%80.57%64.56%93.75%89.47%75.00%
Oasis61.14%66.29%68.24%64.44%64.44%66.29%76.57%19.05%84.42%14.29%16.33%79.43%65.22%88.68%78.95%71.43%
ECTrain96.28%96.28%94.26%98.45%98.50%96.33%97.52%96.83%97.65%88.41%92.42%98.76%99.24%98.53%97.01%98.11%
Test173.14%73.14%69.03%80.65%86.67%76.85%82.86%0.00%83.82%0.00%#DIV/0!90.29%83.33%93.91%87.72%85.47%
Oasis68.00%69.71%68.32%71.62%76.67%72.25%82.29%36.36%85.37%14.29%20.51%84.00%73.02%90.18%80.70%76.67%
DTTrain82.63%87.34%87.44%87.25%87.00%87.22%88.83%65.00%94.74%75.36%69.80%89.08%86.29%90.32%79.85%82.95%
Test162.86%70.86%68.93%73.61%78.89%73.58%78.29%25.00%85.16%17.86%20.83%76.57%65.38%81.30%59.65%62.39%
Oasis64.57%70.86%70.10%71.79%75.56%72.73%78.29%22.22%84.71%14.29%17.39%80.00%68.33%86.09%71.93%70.09%
FFNNTrain100.00%100.00%100.00%100.00%100.00%100.00%100.00%100.00%100.00%100.00%100.00%100.00%100.00%100.00%100.00%100.00%
Test161.14%64.57%65.56%63.53%65.56%65.56%79.43%16.67%84.05%7.14%10.00%78.29%63.01%89.22%80.70%70.77%
Oasis48.57%49.71%51.19%48.35%47.78%49.43%79.43%27.78%85.35%17.86%21.74%68.00%50.68%80.39%64.91%56.92%
(d) Model performance using VD in various model-building algorithms
VD MCIADCN
Overall
Accuracy		AccuracySensitivitySpecificityPrecisionF1 ScoreAccuracySensitivitySpecificityPrecisionF1 ScoreAccuracySensitivitySpecificityPrecisionF1 Score
SVMTrain94.79%96.28%93.02%100.00%100.00%96.39%96.03%100.00%95.43%76.81%86.89%97.27%95.56%98.13%96.27%95.91%
Test171.43%74.86%70.18%83.61%88.89%78.43%84.00%#DIV/0!84.00%0.00%#DIV/0!84.00%73.77%89.47%78.95%76.27%
Oasis71.43%71.43%67.24%79.66%86.67%75.73%84.00%#DIV/0!84.00%0.00%#DIV/0!87.43%79.66%91.38%82.46%81.03%
ECTrain95.29%96.28%93.84%98.96%99.00%96.35%96.28%100.00%95.70%78.26%87.80%98.01%95.65%99.25%98.51%97.06%
Test177.91%77.91%73.21%86.67%91.11%81.19%83.72%#DIV/0!83.72%0.00%#DIV/0!94.19%86.67%98.21%96.30%91.23%
Oasis70.86%70.86%67.57%76.56%83.33%74.63%84.57%57.14%85.71%14.29%22.86%86.29%78.95%89.83%78.95%78.95%
DTTrain84.12%86.85%86.57%87.13%87.00%86.78%88.34%68.33%91.84%59.42%63.57%93.05%87.32%96.17%92.54%89.86%
Test170.86%74.86%71.70%79.71%84.44%77.55%84.00%50.00%86.96%25.00%33.33%82.86%74.55%86.67%71.93%73.21%
Oasis70.29%74.29%70.64%80.30%85.56%77.39%82.29%41.18%86.71%25.00%31.11%84.00%79.59%85.71%68.42%73.58%
FFNNTrain99.72%99.72%99.44%100.00%100.00%99.72%99.72%100.00%99.67%98.41%99.20%100.00%100.00%100.00%100.00%100.00%
Test165.71%67.43%67.74%67.07%70.00%68.85%77.14%22.73%84.97%17.86%20.00%86.86%78.33%91.30%82.46%80.34%
Oasis54.29%54.86%55.24%54.29%64.44%59.49%77.71%13.33%83.75%7.14%9.30%76.00%63.64%81.67%61.40%62.50%
(e) Model performance using VRD in various model-building algorithms
VRD MCIADCN
Overall
Accuracy		AccuracySensitivitySpecificityPrecisionF1 ScoreAccuracySensitivitySpecificityPrecisionF1 ScoreAccuracySensitivitySpecificityPrecisionF1 Score
SVMTrain81.14%83.62%78.15%91.52%93.00%84.93%89.58%100.00%88.83%39.13%56.25%89.08%82.61%92.45%85.07%83.82%
Test171.43%73.14%66.93%89.58%94.44%78.34%82.86%0.00%83.82%0.00%#DIV/0!86.86%86.96%86.82%70.18%77.67%
Oasis68.00%70.29%67.27%75.38%82.22%74.00%80.00%23.08%84.57%10.71%14.63%85.71%80.77%87.80%73.68%77.06%
ECTrain93.55%94.29%91.94%96.88%97.00%94.40%96.53%100.00%95.98%79.71%88.71%96.28%93.43%97.74%95.52%94.46%
Test174.86%76.00%70.00%89.09%93.33%80.00%84.00%50.00%84.39%3.57%6.67%89.71%86.79%90.98%80.70%83.64%
Oasis69.71%72.00%74.12%70.00%70.00%72.00%80.57%40.00%88.97%42.86%41.38%86.86%78.33%91.30%82.46%80.34%
DTTrain71.22%73.70%68.65%82.12%86.50%76.55%84.86%66.67%86.02%23.19%34.41%83.87%77.17%86.96%73.13%75.10%
Test175.43%75.43%69.11%90.38%94.44%79.81%86.29%83.33%86.39%17.86%29.41%89.14%91.30%88.37%73.68%81.55%
Oasis72.00%75.43%73.27%78.38%82.22%77.49%85.14%55.00%89.03%39.29%45.83%83.43%75.93%86.78%71.93%73.87%
FFNNTrain99.72%99.72%99.44%100.00%100.00%99.72%100.00%100.00%100.00%100.00%100.00%99.72%100.00%99.57%99.18%99.59%
Test176.57%78.29%76.53%80.52%83.33%79.79%89.71%72.73%92.16%57.14%64.00%85.14%78.18%88.33%75.44%76.79%
Oasis73.14%74.86%78.75%71.58%70.00%74.12%88.00%62.07%93.15%64.29%63.16%83.43%71.21%90.83%82.46%76.42%
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Cheung, E.Y.W.; Wu, R.W.K.; Chu, E.S.M.; Mak, H.K.F. Integrating Demographics and Imaging Features for Various Stages of Dementia Classification: Feed Forward Neural Network Multi-Class Approach. Biomedicines 2024, 12, 896. https://doi.org/10.3390/biomedicines12040896

AMA Style

Cheung EYW, Wu RWK, Chu ESM, Mak HKF. Integrating Demographics and Imaging Features for Various Stages of Dementia Classification: Feed Forward Neural Network Multi-Class Approach. Biomedicines. 2024; 12(4):896. https://doi.org/10.3390/biomedicines12040896

Chicago/Turabian Style

Cheung, Eva Y. W., Ricky W. K. Wu, Ellie S. M. Chu, and Henry K. F. Mak. 2024. "Integrating Demographics and Imaging Features for Various Stages of Dementia Classification: Feed Forward Neural Network Multi-Class Approach" Biomedicines 12, no. 4: 896. https://doi.org/10.3390/biomedicines12040896

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop