Search Results (90)

The accurate segmentation of pigmented skin lesions is a critical prerequisite for reliable melanoma detection, yet approximately 30% of lesions exhibit fuzzy or poorly defined borders. This ambiguity makes the definition of a single contour unreliable and limits the effectiveness of computer-assisted diagnosis (CAD) systems. While clinical assessment based on the ABCDE criteria (asymmetry, border, color, diameter, and evolution), dermoscopic imaging, and scoring systems remains the standard, these methods are inherently subjective and vary with clinician experience. We address this challenge by reframing segmentation into three distinct regions: background, border, and lesion core. These regions are delineated using superpixels generated via the Simple Linear Iterative Clustering (SLIC) algorithm, which provides meaningful structural units for analysis. Our contributions are fourfold: (1) redefining lesion borders as regions, rather than sharp lines; (2) generating superpixel-level embeddings with a transformer-based autoencoder; (3) incorporating these embeddings as features for superpixel classification; and (4) integrating neighborhood information to construct enhanced feature vectors. Unlike pixel-level algorithms that often overlook boundary context, our pipeline fuses global class information with local spatial relationships, significantly improving precision and recall in challenging border regions. An evaluation on the HAM10000 melanoma dataset demonstrates that our superpixel–RAG–transformer (region adjacency graph) pipeline achieves exceptional performance (100% F1 score, accuracy, and precision) in classifying background, border, and lesion core superpixels. By transforming raw dermoscopic images into region-based structured representations, the proposed method generates more informative inputs for downstream deep learning models. This strategy not only advances melanoma analysis but also provides a generalizable framework for other medical image segmentation and classification tasks. Full article

Modern external beam radiotherapy (EBRT) on lung cancer improved dose distribution thanks to advanced dose calculation algorithms, but side effects and relapses can occur in any case onset. Differential diagnosis of relapses and side effects is difficult, and when computed tomography (CT) is uncertain 18-fluorodeoxyglucose-positron emission tomography/computed tomography (18F-FDG-PET/CT) can support the diagnosis, even if it can also be difficult to construe. The aim of this retrospective analysis was to evaluate 18F-FDG PET/CT qualitative patterns and semiquantitative parameters, both automatic and preceded by physicians, in interpreting lung lesions in the radiotherapy (RT) lung irradiation field. In total, 94 patients (pts) submitted to EBRT (3 months before) for stage II lung cancer were included (74 men, 20 women, mean age of 68 years old, range of 49–84 years old). CT scans were performed on pts, which showed lung lesions in the RT field. 18F-FDG-PET/CT scans were analyzed qualitatively as negative or positive, and the presence of the lung area with a high 18F-FDG uptake pattern was distinguished as the following: focal/wide, deep/shade, or homogeneous/inhomogeneous. Furthermore, the following semiquantitative parameters were collected: gSUVmax (global standardized uptake value max), MTV (tumor metabolic volume), metabolic spatial distribution (MSD) = proximal SUVmax/distal SUVmax, and intratumoral difference in spatial distribution (IDSD%) = [distal SUVmax/proximal SUVmax] × 100. 18F-FDG PET/CT was related to the pts’ outcome (biopsy and/or clinical–instrumental follow-up): positive for lung relapse, negative if the lesions were phlogistic. The following diagnostic performance parameters of 18F-FDG PET/CT were calculated: sensitivity (Sens), specificity (Spec), diagnostic accuracy (DA), positive predictive value (PPV), and negative predictive value (NPV). Qualitative variables were compared by Chi-squared test, while for semiquantitative parameters Student’s t-test was applied; p < 0.05 was considered statistically significant. Statistics tests were performed with MedCalc V.22.018 ©2024. In 76/94 (80.8%) pts, 18F-FDG uptake was higher compared to the background; in 18/94 (19.2%) no high 18F-FDG uptake areas were detected. Outcome was positive for lung relapse in 49/94 pts, while negative in 45/94, with disease prevalence of 52.13% (95%CI = 41.57–62.54%). In the 18/94 pts without high 18F-FDG uptake, the outcome was negative for lung relapse. In 49/76 pts with higher 18F-FDG uptake, the outcome confirmed the presence of relapse, while in 27/76 the lesion was phlogistic. Results about the Sens, Spec, DA, PPV, and NPV (95%CI) were, respectively: 100% (92.75–100%), 40% (25.7–55.67%), 71.28% (61.02–80.14%), 64.47% (58.84–69.73%), and 100% (81.47–100%). Chi-square test showed significant statistical difference between the positive and negative outcome for patterns focal/wide (p = 0.02) and deep/shade (p < 0.00001). A total of 35/49 (71.4%) pts with lung relapse had a focal lesion and 15/27 (55.6%) with phlogosis had a wide pattern. A total of 34/49 (69.4%) pts with lung relapse had a deep pattern and 25/27 (92.6%) with lung phlogosis had the shade one. Significant difference was observed in evaluating the three patterns (p = 0.00007), with prevalence of “focal/deep/homogeneous” patterns in lung relapse and “wide/shade/inhomogeneous” in phlogosis. gSUVmax, MTV, MSD, and IDSD% were in the following order: in the 76 pts, 5.63 (1.4–24.7), 42.49 (4.94–193), 3.61 (1–5.54), and 70.7% (18–100%); in the 49/76 true positive pts, 6.93 (1.5–24.7), 35.28 (4.94–85.99), 3.30 (1.05–5.54), and (18–95%); in the 27/76 false positive pts, 3.27 (1.4–19.2), 38.37 (4.94–193), 1.57 (1–2.13), and 78.6% (4.7–100%). The difference was statistically significant only for MSD (t = 2.779; p = 0.0069) and IDSD% (t = 2.769; p = 0.0071). 18F-FDG-PET/CT confirms its high sensitivity and NPV in evaluating lung lesions after RT. To improve physician confidence in interpreting lung 18F-FDG uptake without further support, MSD and IDSD% could be considered. Heterogeneity of lung lesions, especially in radiotreated tissue, can be turned from a drawback to a resource and analyzed for differentiating relapses from EBRT side effects. Considering the calculation of semiquantitative parameters that require “human intelligence”, even if slightly more time-consuming, can improve the nuclear physician’s confidence in interpreting 18F-FDG PET/CT images. Full article

Background: Lattice Radiation Therapy (LRT), a spatially fractionated stereotactic radiotherapy (SBRT) technique, has shown promising results in the palliative treatment of large tumors. The focus of our first analysis of 56 lesions ≥7 cm was on the extent of shrinkage following palliative LRT (mean 50%) and assessment of its effect duration (: mean 6 months). Herewith we present an updated analysis of our single-center LRT cohort, with a focus on LRT outcome across diagnoses and applied LRT regimens. Methods: We assessed the clinical outcome following LRT in 66 patients treated for 81 lesions between 01.2022 and 05.2025. LRT protocols included simultaneous integrated boost (sib-) LRT in 49 lesions (5 × 4–5 Gy to the entire mass with sib of 9–13 Gy to lattice vertices). Alternatively mainly in pre-irradiated and/or very large lesions—a single-fraction stereotactic LRT (SBRT-LRT) of 1 × 20 Gy to vertices only was delivered to 26 lesions. In six cases with modest response to single fraction SBRT-LRT, the sib-LRT schedule was added 4–8 weeks later. Results: The median age was 68 years (18–93). Main tumor locations were abdomino-pelvic (n = 34) and thoracic (n = 17). Histopathological diagnoses included carcinoma (n = 34), sarcoma (n = 31), and melanoma (n = 16). 31% of all lesions have been previously irradiated. 73% of cases underwent concurrent or peri-LRT systemic therapy. The mean/median overall survival (OS) time of the cohort was 7.6/4.6 months (0.4–40.2), 11.9/5.8 months in 16/66 alive, and 6.4/4.3 months in deceased patients, respectively. 82% of symptomatic patients reported immediate subjective improvement (PROM), with a lifelong response duration in most cases. Progressive disease (PD: >10% increase in initial volume) was found in 9%, stable disease (SD +/−10% of initial volume) in 19% of scanned lesions, and shrinkage (>10% reduction in initial volume) in 75%, with a mean/median tumor volume reduction of 51/60%. The extent of shrinkage was found to be 11–30%/31–60%/61–100% in 38/24/38% of lesions. Response rates (PD, SD, shrinkage) following the two applied LRT regimens, as well as those related to sarcoma and carcinoma diagnoses, were found to be comparable. Treatment tolerance was excellent (G0-1). Conclusions: Palliative LRT provides rapid subjective relief in ~80% of symptomatic patients. Radiologic shrinkage was stated in 75% of FU-scanned lesions, with a lifelong effect duration in most patients. LRT was found effective across histologies, with a similar extent of shrinkage in carcinoma and sarcoma following 1F SBRT- and 5F sib-LRT regimens, respectively. Full article

Background: Recently, the adoption of AI-based technologies has been accelerating in the field of medical image analysis. For the early diagnosis and treatment planning of breast cancer, Automated Breast Ultrasound (ABUS) has emerged as a safe and non-invasive imaging method, especially for women with dense breasts. However, the increasing computational cost due to the minute size and complexity of 3D ABUS data remains a major challenge. Methods: In this study, we propose a novel model based on the Mamba state–space model architecture for 3D tumor segmentation in ABUS images. The model uses Mamba blocks to effectively capture the volumetric spatial features of tumors, and integrates a deep spatial pyramid pooling (DASPP) module to extract multiscale contextual information from lesions of different sizes. Results: On the TDSC-2023 ABUS dataset, the proposed model achieved a Dice Similarity Coefficient (DSC) of 0.8062, and Intersection over Union (IoU) of 0.6831, using only 3.08 million parameters. Conclusions: These results show that the proposed model improves the performance of tumor segmentation in ABUS, offering both diagnostic precision and computational efficiency. The reduced computational space suggests a strong potential for real-world medical applications, where accurate early diagnosis can reduce costs and improve patient survival. Full article

Deep learning has achieved remarkable performance in skin lesion classification, but its lack of interpretability often remains a critical barrier to clinical adoption. In this study, we investigate the spatial properties of saliency-based model explanations, focusing on symmetry and other morphometric features. We benchmark five deep learning architectures (ResNet-50, EfficientNetV2-S, ConvNeXt-Tiny, Swin-Tiny, and MaxViT-Tiny) on a nine-class skin lesion dataset from the International Skin Imaging Collaboration (ISIC) archive, generating saliency maps with Grad-CAM++ and LayerCAM. The best-performing model, Swin-Tiny, achieved an accuracy of 78.2% and a macro-F1 score of 71.2%. Our morphometric analysis reveals statistically significant differences in the explanation maps between correct and incorrect predictions. Notably, the transformer-based models exhibit highly significant differences ($p<0.001$) in metrics related to attentional focus (Entropy and Gini), indicating that their correct predictions are associated with more concentrated saliency maps. In contrast, convolutional models show less consistent differences, and only at a standard significance level ($p<0.05$). These findings suggest that the quantitative morphometric properties of saliency maps could serve as valuable indicators of predictive reliability in medical AI. Full article

Background: Vascular risk factors (VRFs) are known to influence cerebral small vessel disease (cSVD) burden and progression. However, their specific impact on the presence and distribution of each cSVD imaging marker (white matter hyperintensity [WMH], perivascular spaces [PVSs], lacunes, and cerebral microbleeds [CMBs]) and their spatial distribution remains unclear. Methods: We conducted a retrospective analysis of 93 patients with lacunar stroke with a standardized investigational magnetic resonance imaging protocol using a 3T scanner. WMH and PVSs were segmented semi-automatically, and lacunes and CMBs were manually segmented. We assessed the univariable associations of four common VRFs (hypertension, hyperlipidemia, diabetes, and smoking) with the load of each cSVD marker. Then, we assessed the independent associations of these VRFs in multivariable regression models adjusted for age and sex. Spatial lesion patterns were explored with regional volumetric comparisons using Pearson’s coefficient analysis, which was adjusted for multiple comparisons, and by visually examining heatmap lesion distributions. Results: Hypertension was the VRF that exhibited stronger associations with the cSVD markers in the univariable analysis. In the multivariable analysis, only lacunes (p = 0.009) and PVSs in the basal ganglia (p = 0.014) and white matter (p = 0.016) were still associated with hypertension. In the regional analysis, hypertension showed a higher WMH load in deep structures and white matter, particularly in the posterior periventricular regions. In patients with hyperlipidemia, WMH was preferentially found in hippocampal regions. Conclusions: Hypertension was confirmed to be the VRF with the most impact on cSVD load, especially for lacunes and PVSs, while the lesion topography was variable for each VRF. These findings shed light on the complexity of cSVD expression in relation to factors detrimental to vascular health. Full article

Background/Objectives: Accurate segmentation of brain stroke lesions from MRI images is a critical task in medical image analysis that is essential for timely diagnosis and treatment planning. Methods: This paper presents a novel approach for segmenting brain stroke lesions using a deep learning model based on the U-Net neural network architecture. We enhanced the traditional U-Net by integrating a transformer-based backbone, specifically the Mix Vision Transformer (MiT), and compared its performance against other commonly used backbones such as ResNet and EfficientNet. Additionally, we implemented a 2.5D method, which leverages 2D networks to process three-dimensional data slices, effectively balancing the rich spatial context of 3D methods and the simplicity of 2D methods. The 2.5D approach captures inter-slice dependencies, leading to improved lesion delineation without the computational complexity of full 3D models. Utilizing the 2015 ISLES dataset, which includes MRI images and corresponding lesion masks for 20 patients, we conducted our experiments with 4-fold cross-validation to ensure robustness and reliability. To evaluate the effectiveness of our method, we conducted comparative experiments with several state-of-the-art (SOTA) segmentation models, including CNN-based UNet, nnU-Net, TransUNet, and SwinUNet. Results: Our proposed model outperformed all competing methods in terms of Dice Coefficient and Intersection over Union (IoU), demonstrating its robustness and superiority. Our extensive experiments demonstrate that the proposed U-Net with the MiT Backbone, combined with 2.5D data preparation, achieves superior performance metrics, specifically achieving DICE and IoU scores of 0.8153 ± 0.0101 and 0.7835 ± 0.0079, respectively, outperforming other backbone configurations. Conclusions: These results indicate that the integration of transformer-based backbones and 2.5D techniques offers a significant advancement in the accurate segmentation of brain stroke lesions, paving the way for more reliable and efficient diagnostic tools in clinical settings. Full article

Background: Integrating artificial intelligence into clinical photography offers great potential for monitoring skin conditions such as actinic keratosis (AK) and skin field cancerization. Identifying the extent of AK lesions often requires more than analyzing lesion morphology—it also depends on contextual cues, such as surrounding photodamage. This highlights the need for models that can combine fine-grained local features with a comprehensive global view. Methods: To address this challenge, we propose AKTransU-net, a hybrid U-net-based architecture. The model incorporates Transformer blocks to enrich feature representations, which are passed through ConvLSTM modules within the skip connections. This configuration allows the network to maintain semantic coherence and spatial continuity in AK detection. This global awareness is critical when applying the model to whole-image detection via tile-based processing, where continuity across tile boundaries is essential for accurate and reliable lesion segmentation. Results: The effectiveness of AKTransU-net was demonstrated through comparative evaluations with state-of-the-art segmentation models. A proprietary annotated dataset of 569 clinical photographs from 115 patients with actinic keratosis was used to train and evaluate the models. From each photograph, crops of 512 × 512 pixels were extracted using translation lesion boxes that encompassed lesions in different positions and captured different contexts. AKtransU-net exhibited a more robust context awareness and achieved a median Dice score of 65.13%, demonstrating significant progress in whole-image assessments. Conclusions: Transformer-driven context modeling offers a promising approach for robust AK lesion monitoring, supporting its application in real-world clinical settings where accurate, context-aware analysis is crucial for managing skin field cancerization. Full article

Bitter melon, an important medicinal and edible economic crop, is often threatened by diseases such as downy mildew, powdery mildew, viral diseases, anthracnose, and blight during its growth. Efficient and accurate disease detection is of significant importance for achieving sustainable disease management in bitter melon cultivation. To address the issues of weak generalization ability and high computational demands in existing deep learning models in complex field environments, this study proposes an improved lightweight YOLOv8-LSW model. The model incorporates the inverted bottleneck structure of LeYOLO-small to design the backbone network, utilizing depthwise separable convolutions and cross-stage feature reuse modules to achieve lightweight design, reducing the number of parameters while enhancing multi-scale feature extraction capabilities. It also integrates the ShuffleAttention mechanism, strengthening the feature response in lesion areas through channel shuffling and spatial attention dual pathways. Finally, WIoUv3 replaces the original loss function, optimizing lesion boundary regression based on a dynamic focusing mechanism. The results show that YOLOv8-LSW achieves a precision of 95.3%, recall of 94.3%, mAP50 of 98.1%, mAP50-95h of 95.6%, and F1-score of 94.80%, which represent improvements of 2.2%, 2.7%, 1.2%, 2.2%, and 2.46%, respectively, compared to the original YOLOv8n. The effectiveness of the improvements was verified through heatmap analysis and ablation experiments. The number of parameters and GFLOPS were reduced by 20.58% and 20.29%, respectively, with an FPS of 341.58. Comparison tests with various mainstream deep learning models also demonstrated that YOLO-LSW performs well in the bitter melon disease detection task. This research provides a technical solution with both lightweight design and strong generalization ability for real-time detection of bitter melon diseases in complex environments, which holds significant application value in promoting precision disease control in smart agriculture. Full article

Background and Objectives: Single-photon emission computed tomography (SPECT) myocardial perfusion imaging (MPI) is a well-established technique for evaluating myocardial perfusion and function in patients with suspected or known coronary artery disease. While conventional dual-detector SPECT scanners have limitations in spatial resolution and photon detection sensitivity, recent advancements, including full-ring solid-state cadmium zinc telluride (CZT) detectors, offer enhanced image quality and improved diagnostic accuracy. This study aimed to compare the performance of Veriton-CT, a full-ring CZT SPECT system, with GE Discovery 530c, a dedicated cardiac fixed-angle gamma camera, in myocardial perfusion imaging and their correlation with coronary angiography findings. Materials and Methods: This was a prospective study that analyzed 21 patients who underwent MPI at the Department of Nuclear Medicine, Lithuanian University of Health Sciences, Kauno Klinikos. A one-day stress–rest protocol using 99mTc-Sestamibi was employed, with stress testing performed via bicycle ergometry or pharmacological induction. MPI was first conducted using GE Discovery 530c (GE Health Care, Boston, MA, USA), followed by imaging on Veriton-CT, which included low-dose CT for attenuation correction. The summed stress score (SSS), summed rest score (SRS), and summed difference score (SDS) were analyzed and compared between both imaging modalities. Coronary angiography results were retrospectively collected, and lesion-based analysis was performed to assess the correlation between imaging results and the presence of significant coronary artery stenosis (≥35% and ≥70% narrowing). Image quality and the certainty of distinguishing the inferior myocardial wall from extracardiac structures were also evaluated by two independent researchers with differing levels of experience. Results: Among the 14 patients included in the final analysis, Veriton-CT was more likely to classify MPI scans as normal (64.3%) compared to GE Discovery 530c (28.6%). Additionally, Veriton-CT provided a better assessment of the right coronary artery (RCA) basin, showing greater agreement with coronary angiography findings than GE Discovery 530c, although the difference was not statistically significant. No significant differences in lesion overlap were observed for the left anterior descending artery (LAD) or left circumflex artery (LCx) basins. Furthermore, the image quality assessment revealed slightly better delineation of extracardiac structures using Veriton-CT (Spectrum Dynamics Medical, Caesarea, Israel), particularly when evaluated by an experienced researcher. However, no significant difference was observed when assessed by a less experienced observer. Conclusions: Our findings suggest that Veriton-CT, with its full-ring CZT detector system, may offer advantages over fixed-angle gamma cameras in improving image quality and reducing attenuation artifacts in MPI. Although the difference in correlations with coronary angiography findings was not statistically significant, Veriton-CT showed a trend toward better agreement, particularly in the RCA basin. These results indicate that full-ring SPECT imaging could improve the diagnostic accuracy of non-invasive MPI, potentially reducing the need for unnecessary invasive angiography. Further studies with larger patient cohorts are required to confirm these findings and evaluate the clinical impact of full-ring SPECT technology in myocardial perfusion imaging. Full article

IgG4–related disease (IgG4–RD) is an autoimmune condition marked by IgG4–positive plasma cell infiltration, causing inflammation, fibrosis, and tumor–like lesions, especially in the lacrimal gland (LG). Current diagnostic criteria, based primarily on serum IgG4 levels, face limitations in predicting clinical outcomes and treatment responses. To address this, we conducted a multiplex immaunohistochemical analysis of LG tissues to assess immune checkpoint interactions and immune cell distribution in relation to mass size, fibrosis, and treatment response. Our findings revealed that PD–L1 (Programmed Death–Ligand 1), an immune checkpoint molecule, plays a key role in shaping an immunosuppressive environment that varies by clinical group. In non–responsive patients, increased co–expression of PD–L1 and CD11c+ dendritic cells (DCs) suggested a link to treatment resistance. Spatial analysis highlighted more active immune responses in non–fibrotic areas, while fibrotic regions exhibited stabilized immune interactions driven by PD–L1 expression. These results indicate that PD–L1 contributes to immune regulation and disease progression in IgG4–RD and emphasize its potential as a therapeutic target. This study provides new insights into the immunological landscape of IgG4–RD and paves the way for the development of personalized treatment strategies. Full article

Background/Objectives: Plain radiography is the primary diagnostic tool for calcific tendinitis of the shoulder. Several qualitative grading methods have been proposed to represent the pathophysiologic phase and guide treatment decisions. However, these methods have demonstrated low reliability, complicating their effectiveness for such purposes. This study aims to perform the first quantitative analysis of calcific lesions using radiographic imaging and explore their correlation with ultrasonographic parameters to enhance their diagnostic utility. Methods: A total of 57 shoulders presenting with painful calcific tendinitis in either the supraspinatus or subscapularis tendon were reviewed. The calcific deposits and tendon regions of interest were meticulously identified and annotated. Image brightness was reduced to 256 grayscale levels, and descriptive and heterogeneity parameters, including skewness, kurtosis, complexity, and entropy, were quantified and analyzed. Results: In the region of calcification, the average grayscale values were 21.69 units higher than those of tendon tissue. All spatial heterogeneity parameters, except for skewness, demonstrated statistically significant differences when compared with the adjacent tendon. Notably, entropy and complexity were the most distinctive features, with an area under the curve of 0.93 and cut-off values of 4.62 and 4.18, respectively. Significant correlations were observed between the heterogeneity parameters and ultrasonographic findings, such as bursal contact and peri-calcific hypoattenuation. Conclusions: Calcific deposits demonstrated not only increased brightness in grayscale levels but also distinct spatial heterogeneity. The correlation with ultrasonographic findings indicates that these heterogeneity parameters may reflect underlying pathophysiological characteristics. Future prospective research could explore the whole temporal changes of calcifications more thoroughly. Full article

Image quality plays a critical role in medical image analysis, significantly impacting diagnostic outcomes. Sharp and detailed images are essential for accurate diagnoses, but acquiring high-resolution medical images often demands sophisticated and costly equipment. To address this challenge, this study proposes a convolutional neural network (CNN)-based super-resolution architecture, utilizing a melanoma dataset to enhance image resolution through deep learning techniques. The proposed model incorporates a convolutional self-attention block that combines channel and spatial attention to emphasize important image features. Channel attention uses global average pooling and fully connected layers to enhance high-frequency features within channels. Meanwhile, spatial attention applies a single-channel convolution to emphasize high-frequency features in the spatial domain. By integrating various attention blocks, feature extraction is optimized and further expanded through subpixel convolution to produce high-quality super-resolution images. The model uses L1 loss to generate realistic and smooth outputs, outperforming existing deep learning methods in capturing contours and textures. Evaluations with the ISIC 2020 dataset—containing 33126 training and 10982 test images for skin lesion analysis—showed a 1–2% improvement in peak signal-to-noise ratio (PSNR) compared to very deep super-resolution (VDSR) and enhanced deep super-resolution (EDSR) architectures. Full article

Manual labeling of lesions in medical image analysis presents a significant challenge due to its labor-intensive and inefficient nature, which ultimately strains essential medical resources and impedes the advancement of computer-aided diagnosis. This paper introduces a novel medical image-segmentation framework named Efficient Generative-Adversarial U-Net (EGAUNet), designed to facilitate rapid and accurate multi-organ labeling. To enhance the model’s capability to comprehend spatial information, we propose the Global Spatial-Channel Attention Mechanism (GSCA). This mechanism enables the model to concentrate more effectively on regions of interest. Additionally, we have integrated Efficient Mapping Convolutional Blocks (EMCB) into the feature-learning process, allowing for the extraction of multi-scale spatial information and the adjustment of feature map channels through optimized weight values. Moreover, the proposed framework progressively enhances its performance by utilizing a generative-adversarial learning strategy, which contributes to improvements in segmentation accuracy. Consequently, EGAUNet demonstrates exemplary segmentation performance on public multi-organ datasets while maintaining high efficiency. For instance, in evaluations on the CHAOS T2SPIR dataset, EGAUNet achieves approximately $2\%$ higher performance on the Jaccard metric, $1\%$ higher on the Dice metric, and nearly $3\%$ higher on the precision metric in comparison to advanced networks such as Swin-Unet and TransUnet. Full article

This case study highlights the use of cinematic rendering (CR) in preoperative planning for the excision of a cyst in the oral and maxillofacial region of a 60-year-old man. The patient presented with a firm, non-tender mass in the right cheek, clinically suspected to be an epidermoid cyst. Conventional imaging, including dental magnetic resonance imaging (MRI) protocols, confirmed the lesion’s size, location, and benign nature. CR reconstructions, combining advanced algorithms and novel skin presets, allow for the generation of highly realistic, three-dimensional visualizations from conventional imaging datasets. CR provided an enhanced, detailed depiction of the lesion within its anatomical context, significantly improving spatial understanding for surgical planning. The surgical excision was performed without complications, and histological analysis confirmed the diagnosis of a benign epidermoid cyst with no evidence of dysplasia or malignancy. This case demonstrates the potential of CR to refine preoperative planning, especially in complex anatomical regions such as the face and jaw, by offering superior visualization of superficial and deep structures. Thus, the integration of CR into clinical workflows has the potential to lead to improved diagnostic accuracy and better surgical outcomes. Full article