Research Appraisals
Evidence-based critical appraisals of the latest medical research, systematically evaluated using Oxford CEBM methodology.
Showing 2 appraisals
Journal of medical systems
RetCond: A Conditional Diffusion Model for Self-Explanatory Multi-Class Fundus Image Classification
Vision impairment remains a major global health challenge, and early diagnosis of retinal diseases is essential to prevent avoidable vision loss. Color fundus photography is widely used for retinal assessment, but manual interpretation is time-consuming and prone to inter-clinician variability. Although deep learning has improved automated retinal disease detection, most classification models lack transparency. We introduce RetCond, a self-explanatory generative classifier designed to provide both accurate predictions and counterfactual visual explanations. RetCond is a diffusion model repurposed as a classifier whose generative capabilities enable counterfactual synthesis for built-in explanations. These counterfactual images depict the retinal changes associated with alternative diagnoses, offering insight into the model's decisions. We curated a diverse dataset of 19,565 images containing five retinal disease conditions to train and evaluate the model. Performance is assessed using standard classification metrics and by qualitatively and quantitatively evaluating the self-explanatory component. Comparisons are made against strong discriminative baselines. RetCond achieves a classification accuracy of 96.98%, comparable to RETFound (96.99%). In contrast to standard models, RetCond produces counterfactual images that highlight condition-specific retinal features, confirming its learned concepts and demonstrating its self-explanatory properties. These results show that RetCond matches state-of-the-art discriminative performance while addressing limitations of post-hoc interpretability techniques in standard deep learning models. RetCond, therefore, constitutes a contribution toward more trustworthy AI solutions in ophthalmology.
23 May 2026
Read appraisal →American journal of ophthalmology
Optical Coherence Tomography Radiomics and Machine Learning Enable Accurate Detection of Forme Fruste Keratoconus
PURPOSE: To evaluate the diagnostic performance of a radiomics-based machine learning approach applied to corneal optical coherence tomography (OCT) images for detecting forme fruste keratoconus (FFKC). DESIGN: Evaluation of machine learning diagnostic algorithms. METHODS: OCT images from 307 eyes (234 normal, 73 FFKC) were acquired along eight meridians (M1-M8). All images underwent preprocessing before texture-based radiomics feature extraction. Three machine learning classifiers-Random Forest, C5.0, and XGBoost-were trained using a feature subset selected by recursive feature elimination (RFE). Model performance was evaluated using the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and accuracy. RESULTS: A total of 3752 features were extracted per eye, of which 41 were selected for model training. All three models demonstrated strong diagnostic performance in the test set (AUCs > 0.92), with no significant differences between models (P > .05). The XGBoost model achieved the highest performance (AUC = 0.93, 95% CI: 0.829-1.0, sensitivity 0.857, specificity = 0.978, accuracy = 0.950). Among the top 10 XGBoost features ranked by importance, a preferred meridional distribution was observed, with most features concentrated along M1 to M3, corresponding to the inferotemporal corneal region. CONCLUSION: Radiomics analysis of corneal OCT images combined with machine learning enables accurate FFKC detection using a single imaging device, providing diagnostic information beyond conventional morphological assessment and suggesting a potential imaging biomarker for early keratoconus screening.
5 May 2026
Read appraisal →