Search

Introduction The abdomen is the most prevalent site of primarily unidentified fatal injury after blunt trauma, and represents the third major affected part. However, follow-up strategies of intra-abdominal injuries remain controversial. The aim of this study is to describe the characteristics of children with blunt abdominal trauma (BAT) and who presented radiographic amendments in re-imaging studies and predict factors that could identify patients group requiring control in this population. Methods A retrospective study was conducted in the department of pediatric surgery and the intensive care unit of Monastir involving all patients under the age of 14 years old who were admitted for blunt abdominal trauma between January 2010 and December 2021.An analysis of epidemiological, clinical, radiological and therapeutic management characteristics were performed. Results A total of 151 patients were included in this study with a mean age of six years. The sex ratio (m/f) was 2.7. Most trauma circumstances were related to road accidents (43.7%) and falls (32.4%). The liver, the spleen, and the kidneys were the most frequently affected organs. A combination of intra-abdominal lesions was recorded in 32 cases. 84% of the cases (110/131) of the reimaged patients had no complications, compared to 16% (21/131) who developed intra-abdominal complications. The complicated reimaging findings were significantly associated with high grade hepatic and renal injuries (p=0.019 and 0.002, respectively), and patient symptom persistence or development (p=0.001). Conclusion It is safe to avoid performing reimaging studies if clinical progression remains uneventful in children with low-grade hepatic, splenic, and renal lesions. However, the clinical assessment is the most crucial consideration during BAT management in children.

Abstract Background Metanephric adenoma is a rare benign renal tumor of the kidney, uncommonly observed in children. It is often misdiagnosed preoperatively as a malignant neoplasm, leading to an unnecessary nephrectomy. The challenge is to make the right diagnosis preoperatively and therefore manage it with conservative surgery. We report a case of a child with metanephric adenoma who underwent nephron-sparing surgery. Case presentation A renal tumor was discovered fortuitously in an 18-month-old Caucasian girl with several congenital malformations. Investigations showed a 28 × 27 × 27 mm left renal mass centrally located, well defined, nonvascularized, with no calcifications and which compressed the adjacent renal tissue. Furthermore, there were no signs of metastasis. The decision of a multidisciplinary meeting was to perform a computed tomography (CT)-scan-guided biopsy. Histologic examination concluded it was a metanephric adenoma. We performed a left open partial nephrectomy via a flank retroperitoneal incision. The final histopathological examination confirmed the diagnosis. The postoperative course was uneventful. Conclusion Preoperative diagnosis of metanephric adenoma is challenging. Because of the high probability of unnecessary radical nephrectomy, preoperative biopsy can be safe and determining to guide a more conservative approach so nephron-sparing surgery can be performed.

PurposeTo compare change over time in eye-specific optical coherence tomography (OCT) retinal nerve fiber layer (RNFL)-based region-of-interest (ROI) maps developed using unsupervised deep-learning auto-encoders (DL-AE) to circumpapillary RNFL (cpRNFL) thickness for the detection of glaucomatous progression.MethodsForty-four progressing glaucoma eyes (by stereophotograph assessment), 189 nonprogressing glaucoma eyes (by stereophotograph assessment), and 109 healthy eyes were followed for ≥3 years with ≥4 visits using OCT. The San Diego Automated Layer Segmentation Algorithm was used to automatically segment the RNFL layer from raw three-dimensional OCT images. For each longitudinal series, DL-AEs were used to generate individualized eye-based ROI maps by identifying RNFL regions of likely progression and no change. Sensitivities and specificities for detecting change over time and rates of change over time were compared for the DL-AE ROI and global cpRNFL thickness measurements derived from a 2.22-mm to 3.45-mm annulus centered on the optic disc.ResultsThe sensitivity for detecting change in progressing eyes was greater for DL-AE ROIs than for global cpRNFL annulus thicknesses (0.90 and 0.63, respectively). The specificity for detecting not likely progression in nonprogressing eyes was similar (0.92 and 0.93, respectively). The mean rates of change in DL-AE ROI were significantly faster than for cpRNFL annulus thickness in progressing eyes (-1.28 µm/y vs. -0.83 µm/y) and nonprogressing eyes (-1.03 µm/y vs. -0.78 µm/y).ConclusionsEye-specific ROIs identified using DL-AE analysis of OCT images show promise for improving assessment of glaucomatous progression.Translational relevanceThe detection and monitoring of structural glaucomatous progression can be improved by considering eye-specific regions of likely progression identified using deep learning.

This study characterizes differences in glaucomatous eyes with and without high axial myopia using custom automated analysis of OCT images. 452 eyes of 277 glaucoma patients were stratified into non (n = 145 eyes), mild (n = 214 eyes), and high axial myopia (axial length (AL) > 26 mm, n = 93 eyes). Optic disc ovality index, tilt and rotation angle of Bruch´s membrane opening (BMO) and peripapillary choroidal thickness (PCT) were calculated using automated and deep learning strategies. High myopic optic discs were more oval and had larger BMO tilt than mild and non-myopic discs (both p

PurposeThe purpose of this study was to determine if the rate of change in the depth of the surface of the lamina cribrosa due to glaucomatous remodeling differs between glaucoma patients of African descent (AD) and European descent (ED).MethodsThere were 1122 images taken longitudinally over an average of 3 years (range = 0.9-4.1 years) from 122 patients with glaucoma followed in the African Descent and Glaucoma Evaluation Study (ADAGES) and Diagnostic Intervention and Glaucoma Study (DIGS) were automatically segmented to compute anterior lamina cribrosa surface depth (ALCSD). The rate of ALCSD change was compared across racial groups after adjusting for baseline characteristics known to be associated with ALCSD or disease progression (visual field, ALCSD, corneal thickness, optic disk size, and age).ResultsAfter adjusting for all other covariates, the ED group had significantly greater ALCSD posterior migration (deepening) than the AD group (difference = 2.57 µm/year, P = 0.035). There was a wider range of ALCSD change in the ED compared with the AD group, and more individuals had greater magnitude of both deepening and shallowing. No other covariates measured at baseline had independent effects on the longitudinal changes in ALCSD (baseline visual field severity, baseline ALCSD, corneal thickness, Bruch's membrane opening [BMO] area, or age).ConclusionsGlaucomatous remodeling of the lamina cribrosa differs between AD and ED patients with glaucoma. Unlike the cross-sectional associations seen with aging, in which a deeper ALCSD was seen with age in the ED group, glaucomatous remodeling in this longitudinal study resulted in more posterior migration of ALCSD in ED compared to AD patients.

The Cancel Culture Between Quest for Justice and Overflow Cancel culture has developed in the various arts in multiple facets in order to publicly denounce individuals, groups or institutions responsible for acts, behaviours or remarks considered inadmissible or despicable by human society. Even if the concept of "the cancel culture " appeared at the end of the 2010s. Thus, more than an awareness strategy, the objective of the cancel culture in the arts is part of a real confrontation and continual struggle to minimize the impact of groups and behaviours that threaten modern societies. Since September 11, 2001 and the attacks on the twin towers of the World Trade Center and the Pentagon, several directors have chosen to fight against religious fanaticism and terrorism by directing films that denounce Islamic fundamentalism and fanaticism in all its forms. The appearance of Daesh in Syria and Iraq has led to the proliferation of cinematographic works that denounce the brutality, religious fanaticism and terrorism of this terrorist group. However, we note that a large part of the cinematographic works and successful series have attached the label of terrorist to Muslims and to the inhabitants of the Arab-Muslim world. In other words, by fighting against certain behaviours and stereotypes, some of the filmmakers have contributed to creating new stereotypes and nourishing them through artistic creations that stigmatize a particular religious category. We will first seek to present the different social representations of the phenomenon of cancel culture in popular culture and on social networks. In this first part, we will demonstrate the role of the cancel culture in delivering justice to victims in new, less legal and more instantaneous forms. Secondly, we will illustrate through different cinematographic examples how the concept of the cancel culture has led to the appearance of films and series that denounce intolerance, religious fundamentalism and terrorism. Finally, we will focus on the limits of this concept which has overflowed in several artistic representations to turn into the stigmatization and lynching of a particular social and religious category.

In the description of the automated delineation algorithm, it was incorrectly stated that measurements were made 1.8 µm from the BMO centroid, whereas the correct value was 18 µm. The affected sentence is in the Image Acquisition and Processing section of the Methods and originally read: “The mean of ALCSD measurements from all 48 B-scans within 1.8 µm of the center of the BMO from each individual eye was used for the analysis after qualitative review of the accuracy of the BMO and ALCSD segmentation.” The error has been corrected in the article online and does not affect the conclusions of the article.

PurposeTo compare gradient-boosting classifier (GBC) analysis of optical coherence tomography angiography (OCTA)-measured vessel density (VD) and OCT-measured tissue thickness to standard OCTA VD and OCT thickness parameters for classifying healthy eyes and eyes with early to moderate glaucoma.DesignComparison of diagnostic tools.MethodsA total of 180 healthy eyes and 193 glaucomatous eyes with OCTA and OCT imaging of the macula and optic nerve head (ONH) were studied. Four GBCs were evaluated that combined 1) all macula VD and thickness measurements (Macula GBC), 2) all ONH VD and thickness measurements (ONH GBC), 3) all VD measurements from the macula and ONH (vessel density GBC), and 4) all thickness measurements from the macula and ONH (thickness GBC). ROC curve (AUROC) analyses compared the diagnostic accuracy of GBCs to that of standard instrument-provided parameters. A fifth GBC that combined all parameters (full GBC) also was investigated.ResultsGBCs had better diagnostic accuracy than standard OCTA and OCT parameters with AUROCs ranging from 0.90 to 0.93 and 0.64 to 0.91, respectively. The full GBC (AUROC = 0.93) performed significantly better than the ONH GBC (AUROC = 0.91; P = .036) and the vessel density GBC (AUROC = 0.90; P = .010). All other GBCs performed similarly. The mean relative influence of each parameter included in the full GBC identified a combination of macular thickness and ONH VD measurements as the greatest contributors.ConclusionsGBCs that combine OCTA and OCT macula and ONH measurements can improve diagnostic accuracy for glaucoma detection compared to most but not all instrument provided parameters.

PurposeTo compare performance of independently developed deep learning algorithms for detecting glaucoma from fundus photographs and to evaluate strategies for incorporating new data into models.MethodsTwo fundus photograph datasets from the Diagnostic Innovations in Glaucoma Study/African Descent and Glaucoma Evaluation Study and Matsue Red Cross Hospital were used to independently develop deep learning algorithms for detection of glaucoma at the University of California, San Diego, and the University of Tokyo. We compared three versions of the University of California, San Diego, and University of Tokyo models: original (no retraining), sequential (retraining only on new data), and combined (training on combined data). Independent datasets were used to test the algorithms.ResultsThe original University of California, San Diego and University of Tokyo models performed similarly (area under the receiver operating characteristic curve = 0.96 and 0.97, respectively) for detection of glaucoma in the Matsue Red Cross Hospital dataset, but not the Diagnostic Innovations in Glaucoma Study/African Descent and Glaucoma Evaluation Study data (0.79 and 0.92; P < .001), respectively. Model performance was higher when classifying moderate-to-severe compared with mild disease (area under the receiver operating characteristic curve = 0.98 and 0.91; P < .001), respectively. Models trained with the combined strategy generally had better performance across all datasets than the original strategy.ConclusionsDeep learning glaucoma detection can achieve high accuracy across diverse datasets with appropriate training strategies. Because model performance was influenced by the severity of disease, labeling, training strategies, and population characteristics, reporting accuracy stratified by relevant covariates is important for cross study comparisons.Translational relevanceHigh sensitivity and specificity of deep learning algorithms for moderate-to-severe glaucoma across diverse populations suggest a role for artificial intelligence in the detection of glaucoma in primary care.

PurposeTo investigate the relationship between corneal hysteresis (CH) and anterior lamina cribrosa surface (ALCS) displacement over time in a cohort of patients with glaucoma.DesignProspective observational case series.MethodsIn this study, 147 eyes from 96 glaucoma or glaucoma suspect patients were followed for a mean of 3.5 years and 7.9 visits. Baseline CH measurements were obtained using the Ocular Response Analyzer (ORA; Reichert Ophthalmic Instruments Inc, Depew, New York, USA). The mean anterior lamina cribrosa surface depth (ALCSD) and choroidal thickness were by automated segmentation of spectral-domain optical coherence tomography (SD-OCT) scans. The rate of change of ALCSD was calculated using linear mixed effects models. Relationship between baseline CH and follow-up ALCSD rate of change was adjusted for confounding factors, including age, intraocular pressure (IOP), and choroidal thickness.ResultsThe mean baseline CH was 9.4 mm Hg (95% confidence interval [CI] 9.1-9.7). Overall, the ALCS was displaced posteriorly at a rate of 0.78 μm/y (95% CI -1.82, 0.26). Seventeen eyes (11.5%) showed a significant posterior displacement of ALCS, whereas 22 eyes (15.0%) showed a significant anterior displacement of ALCS. The choroidal thickness thinned at a rate of -1.09 μm/y during the follow-up (P = .001). Multivariable mixed modeling showed that choroidal thinning, lower IOP change, and lower corneal hysteresis were significantly associated with posterior ALCS displacement over time (P = .034, P = .037, and P = .048). Each 1 mm lower CH was associated with 0.66 μm/y posterior displacement of the ALCS.ConclusionsLower corneal hysteresis was significantly associated with posterior displacement of the anterior lamina cribrosa over time. These data provide additional support for lower corneal hysteresis being a risk factor for glaucoma progression.

PurposeTo develop and evaluate a deep learning system for differentiating between eyes with and without glaucomatous visual field damage (GVFD) and predicting the severity of GFVD from spectral domain OCT (SD OCT) optic nerve head images.DesignEvaluation of a diagnostic technology.ParticipantsA total of 9765 visual field (VF) SD OCT pairs collected from 1194 participants with and without GVFD (1909 eyes).MethodsDeep learning models were trained to use SD OCT retinal nerve fiber layer (RNFL) thickness maps, RNFL en face images, and confocal scanning laser ophthalmoscopy (CSLO) images to identify eyes with GVFD and predict quantitative VF mean deviation (MD), pattern standard deviation (PSD), and mean VF sectoral pattern deviation (PD) from SD OCT data.Main outcome measuresDeep learning models were compared with mean RNFL thickness for identifying GVFD using area under the curve (AUC), sensitivity, and specificity. For predicting MD, PSD, and mean sectoral PD, models were evaluated using R2 and mean absolute error (MAE).ResultsIn the independent test dataset, the deep learning models based on RNFL en face images achieved an AUC of 0.88 for identifying eyes with GVFD and 0.82 for detecting mild GVFD significantly (P < 0.001) better than using mean RNFL thickness measurements (AUC = 0.82 and 0.73, respectively). Deep learning models outperformed standard RNFL thickness measurements in predicting all quantitative VF metrics. In predicting MD, deep learning models based on RNFL en face images achieved an R2 of 0.70 and MAE of 2.5 decibels (dB) compared with 0.45 and 3.7 dB for RNFL thickness measurements. In predicting mean VF sectoral PD, deep learning models achieved high accuracy in the inferior nasal (R2 = 0.60) and superior nasal (R2 = 0.67) sectors, moderate accuracy in inferior (R2 = 0.26) and superior (R2 = 0.35) sectors, and lower accuracy in the central (R2 = 0.15) and temporal (R2 = 0.12) sectors.ConclusionsDeep learning models had high accuracy in identifying eyes with GFVD and predicting the severity of functional loss from SD OCT images. Accurately predicting the severity of GFVD from SD OCT imaging can help clinicians more effectively individualize the frequency of VF testing to the individual patient.

The advancement of technology has paved the way for a new type of bullying, which often leads to negative stigma in the social setting. Cyberbullying is a cybercrime wherein one individual becomes the target of harassment and hatred. It has recently become more prevalent due to a rise in the usage of social media platforms, and, in some severe situations, it has even led to victims’ suicides. In the literature, several cyberbullying detection methods are proposed, but they are mainly focused on word-based data and user account attributes. Furthermore, most of them are related to the English language. Meanwhile, only a few papers have studied cyberbullying detection in Arabic social media platforms. This paper, therefore, aims to use machine learning in the Arabic language for automatic cyberbullying detection. The proposed mechanism identifies cyberbullying using the Support Vector Machine (SVM) classifier algorithm by using a real dataset obtained from YouTube and Twitter to train and test the classifier. Moreover, we include the Farasa tool to overcome text limitations and improve the detection of bullying attacks.

It is widely known that an appropriately built unpowered bipedal robot can walk down an inclined surface with a passive steady gait. The features of such gait are determined by the robot's geometry and inertial properties, as well as the slope angle. The energy needed to keep the biped moving steadily comes from the gravitational potential energy as it descends the inclined surface. The study of such passive natural motions could lead to ideas for managing active walking devices and a better understanding of the human locomotion. The major goal of this study is to further investigate order, chaos and bifurcations and then to demonstrate the complexity of the passive bipedal walk of the compass-gait biped robot by examining different bifurcation diagrams and also by studying the variation of the eigenvalues of the Poincaré map's Jacobian matrix and the variation of the Lyapunov exponents. We reveal also the exhibition of some additional results by changing the inertial and geometrical parameters of the bipedal robot model.

Introduction: The abdomen is the most prevalent site of primarily unidentified fatal injury after blunt trauma, and represents the third major affected part. However, follow-up strategies of intra-abdominal injuries remain controversial. The aim of this study is to describe the characteristics of children with blunt abdominal trauma (BAT) and who presented radiographic amendments in re-imaging studies and predict factors that could identify patients group requiring control in this population. Methods: A retrospective study was conducted in the department of pediatric surgery and the intensive care unit of Monastir involving all patients under the age of 14 years old who were admitted for blunt abdominal trauma between January 2010 and December 2021.An analysis of epidemiological, clinical, radiological and therapeutic management characteristics were performed. Results: A total of 151 patients were included in this study with a mean age of six years. The sex ratio (m/f) was 2.7. Most trauma circumstances were related to road accidents (43.7%) and falls (32.4%). The liver, the spleen, and the kidneys were the most frequently affected organs. A combination of intra-abdominal lesions was recorded in 32 cases. 84% of the cases (110/131) of the reimaged patients had no complications, compared to 16% (21/131) who developed intra-abdominal complications. The complicated reimaging findings were significantly associated with high grade hepatic and renal injuries (p=0.019 and 0.002, respectively), and patient symptom persistence or development (p=0.001). Conclusion: It is safe to avoid performing reimaging studies if clinical progression remains uneventful in children with low-grade hepatic, splenic, and renal lesions. However, the clinical assessment is the most crucial consideration during BAT management in children.

A longitudinal ophthalmic dataset was used to investigate multi-modal machine learning (ML) models incorporating patient demographics and history, clinical measurements, optical coherence tomography (OCT), and visual field (VF) testing in predicting glaucoma surgical interventions. The cohort included 369 patients who underwent glaucoma surgery and 592 patients who did not undergo surgery. The data types used for prediction included patient demographics, history of systemic conditions, medication history, ophthalmic measurements, 24-2 VF results, and thickness measurements from OCT imaging. The ML models were trained to predict surgical interventions and evaluated on independent data collected at a separate study site. The models were evaluated based on their ability to predict surgeries at varying lengths of time prior to surgical intervention. The highest performing predictions achieved an AUC of 0.93, 0.92, and 0.93 in predicting surgical intervention at 1 year, 2 years, and 3 years, respectively. The models were also able to achieve high sensitivity (0.89, 0.77, 0.86 at 1, 2, and 3 years, respectively) and specificity (0.85, 0.90, and 0.91 at 1, 2, and 3 years, respectively) at an 0.80 level of precision. The multi-modal models trained on a combination of data types predicted surgical interventions with high accuracy up to three years prior to surgery and could provide an important tool to predict the need for glaucoma intervention.

PurposeTo determine if variation in anterior lamina cribrosa surface depth (ALCSD) differs between glaucoma patients of African (AD) and European descent (ED).MethodsA total of 178 eyes from 123 glaucoma patients in the African Descent and Glaucoma Evaluation Study (ADAGES) and Diagnostic Innovations in Glaucoma Study (DIGS) were included. ALCSD and choroidal thickness were measured using the San Diego Automated Layer Segmentation Algorithm (SALSA). ALCSD was defined by both Bruch's membrane opening (BMO)-based (ALCSD-BMO) and scleral-based (ALCSD-Scl) reference planes. Racial differences in ALCSD were evaluated using cross-sectional univariate and multivariable models.ResultsA deeper ALCSD-Scl was found in males (52.4 μm, P = 0.0401), AD individuals (78.6 μm, P = 0.0004), younger individuals (-3.1 μm/year, P < 0.0213), and eyes with larger discs (81.0 μm/mm2, P = 0.024), increased visual field loss (mean defect, MD: -6.4 μm/dB [decibel], P = 0.0106), and higher intraocular pressure (IOP: 14.1 μm/mm Hg, P = 0.0256). Significant deepening of ALSCD was observed with increasing IOP and visual field severity only in the AD group. Race modified the relationship between ALCSD-Scl and age (P = 0.0145) with ALCSD-Scl in AD individuals becoming more shallow with increasing age (-3.1 μm/year, P = 0.0213), while there was no significant association in the ED group (2.1 μm/mm Hg, P < 0.2026).ConclusionsThis study demonstrates that a deeper ALCSD, regardless of the ALCSD reference plane used, is associated with more severe glaucoma and higher IOP in the ADAGES cohort, particularly in individuals of AD. These results suggest that characterizing ALCSD morphology and its relationships to IOP, aging, and glaucoma progression may help explain racial differences in disease susceptibility.

PurposeTo compare race-related differences in estimated rate of change of Bruch's membrane opening-minimum rim width (BMO-MRW) and circumpapillary retinal nerve fiber layer thickness (RNFLT) in healthy, glaucoma suspect, and glaucoma eyes of individuals of European (ED) and African descent (AD).DesignProspective cohort study.MethodsThis study investigated rate of change of BMO-MRW and RNFLT in 124 healthy, 227 glaucoma suspect, and 177 glaucoma eyes followed for approximately 3 years and tested with optical coherence tomography every 6 months. Suspect eyes had a history of untreated intraocular pressure (IOP) ≥ 22 mm Hg or suspicion of glaucoma by optic disc photograph assessment without repeatable abnormal standard automated perimetry (SAP) results. Glaucoma eyes had repeatable abnormal SAP results (GHT ONL or PSD ≤ 5%). Mixed-effects models were used to estimate the rate of change after controlling for age, mean follow-up IOP, central corneal thickness, axial length, and BMO area.ResultsA race-related difference in rate of change of global BMO-MRW but not average RNFLT in suspect eyes was observed. Rate of change of BMO-MRW was -1.82 μm/year and -2.20 μm/year in ED and AD suspect eyes, respectively (P = .03). Rate of change of RNFLT was -0.64 μm/year and -0.75 μm/year in ED and AD suspect eyes, respectively (P = .75). No race-related differences in change rate were found in healthy or glaucoma eyes.ConclusionRace is an important consideration when assessing structural change, particularly minimum rim width, in glaucoma suspect eyes. Differences in rate of structural change may help explain racial disparities in glaucoma susceptibility.

Purpose: To compare the diagnostic accuracy and explainability of a Vision Transformer deep learning technique, Data-efficient image Transformer (DeiT), and ResNet-50, trained on fundus photographs from the Ocular Hypertension Treatment Study (OHTS) to detect primary open-angle glaucoma (POAG) and identify the salient areas of the photographs most important for each model’s decision-making process. Design: Evaluation of a diagnostic technology. Subjects, Participants, and Controls: Overall 66 715 photographs from 1636 OHTS participants and an additional 5 external datasets of 16 137 photographs of healthy and glaucoma eyes. Methods: Data-efficient image Transformer models were trained to detect 5 ground-truth OHTS POAG classifications: OHTS end point committee POAG determinations because of disc changes (model 1), visual field (VF) changes (model 2), or either disc or VF changes (model 3) and Reading Center determinations based on disc (model 4) and VFs (model 5). The best-performing DeiT models were compared with ResNet-50 models on OHTS and 5 external datasets. Main Outcome Measures: Diagnostic performance was compared using areas under the receiver operating characteristic curve (AUROC) and sensitivities at fixed specificities. The explainability of the DeiT and ResNet-50 models was compared by evaluating the attention maps derived directly from DeiT to 3 gradient-weighted class activation map strategies. Results: Compared with our best-performing ResNet-50 models, the DeiT models demonstrated similar performance on the OHTS test sets for all 5 ground-truth POAG labels; AUROC ranged from 0.82 (model 5) to 0.91 (model 1). Data-efficient image Transformer AUROC was consistently higher than ResNet-50 on the 5 external datasets. For example, AUROC for the main OHTS end point (model 3) was between 0.08 and 0.20 higher in the DeiT than ResNet-50 models. The saliency maps from the DeiT highlight localized areas of the neuroretinal rim, suggesting important rim features for classification. The same maps in the ResNet-50 models show a more diffuse, generalized distribution around the optic disc. Conclusions: Vision Transformers have the potential to improve generalizability and explainability in deep learning models, detecting eye disease and possibly other medical conditions that rely on imaging for clinical diagnosis and management.