The study proposes a diagnostic-guided approach to choose between stochastic and deterministic longitudinal imaging models based on whether inter-visit changes are driven by disease progression or acquisition variability. Applied to a large, heterogeneous Optos fundus autofluorescence archive, the diagnostic revealed weak time-dependent changes and limited benefit from stochastic models, leading to the development of the deterministic Temporal Retinal U‑Net (TRU). TRU outperformed other classical and deep‑learning comparators on image‑level and eye‑specific progression metrics across held‑out and independent zero‑shot transfer cohorts, though with slightly lower precision in a smaller cross‑vendor cohort.
By Liyin Chen, Souvick Mukherjee, Ines Maria De Carvalho Lains, Nazlee Zebardast, Mengyu Wang, Tobias Elze, Jason I. Comander
Deep learning framework GLAM predicts glaucoma progression rates from longitudinal Humphrey 24‑2 visual field data and five clinical features, achieving a mean absolute error of 0.139 dB yr⁻¹ and an AUC of 0.990 for fast‑progressor detection. Using attention‑based fusion and aleatoric uncertainty, GLAM outperforms a ridge regression baseline by 73.5% in MD‑rate prediction. The model demonstrates that visual field data alone can match multimodal pipelines for progression prognostication.
By Taiabur Rahman, Siddiqur Rahman, Muhammad Moniruzzaman, Ummay Kawsar, Sayedatunnessa Ratna, Shadman Siddique, Rafsan Siddique, Tausif Ahmad, Tahsin Ahmad, Golam Rabbani
arXiv:2609.12834v1 Announce Type: new
Abstract: Modelling how a disease progresses over time requires longitudinal imaging cohorts, which are scarce and small, whereas cross-sectional data -- one ima...
By Ifeoma Veronica Nwabufo, Julius Gervelmeyer, Sarah M\"uller, Philipp Berens
arXiv:2509.25549v3 Announce Type: replace-cross
Abstract: Choroidal nevi are common benign pigmented lesions in the eye, with a small risk of transforming into melanoma. Early detection is critical t...
By Mohammadmahdi Eshragh, Emad A. Mohammed, Behrouz Far, Ezekiel Weis, Carol L Shields, Sandor R Ferenczy, Trafford Crump
This study presents four deep‑learning pipelines—two‑dimensional and three‑dimensional—for segmenting age‑related macular degeneration (AMD) and diabetic macular edema (DME) lesions in optical coherence tomography (OCT) images. The models achieve Dice scores between 0.76 and 0.82 and demonstrate strong volumetric and surface calibration (r_vol, r_surf ≥ 0.97) on an in‑domain validation set. Generalization was assessed on the OLIVES clinical cohort using proxy metrics such as biomarker AUROC, central subfield thickness correlation, and longitudinal concordance, showing that the predictions still track clinical biomarkers outside the training distribution, albeit with reduced strength.
By Lucia Sundberg, Zhihao Zhao, M. Ali Nasseri
The paper introduces an ordinal latent diffusion model for generating color fundus images that incorporates the ordered structure of diabetic retinopathy (DR) severity, using a scalar disease representation instead of categorical conditioning. Evaluations on the EyePACS dataset show improved visual realism, with reduced Fréchet inception distance for most stages and a higher quadratic weighted κ from 0.79 to 0.87. Interpolation experiments demonstrate the model captures a continuous spectrum of disease progression derived from coarse, ordered labels.
By Gustav Schmidt, Philipp Berens, Sarah M\"uller
arXiv:2508. 03750v2 Announce Type: replace Abstract: Early and accurate glaucoma detection is critical to prevent irreversible vision loss, yet existing AI methods often rely on unimodal inputs and lack interpretability.
By Cheng Huang, Zeyu Han, Weizheng Xie, Karanjit Kooner, Tsengdar Lee, Jui-Kai Wang, Jia Zhang
The study evaluates fundus-specific foundation models (FM) for detecting diabetic macular edema (DME) in retinal images. It compares two popular FM—RETFound and FLAIR—against a lightweight EfficientNet-B0 backbone across multiple datasets (IDRiD, MESSIDOR-2, and OCT-and-Eye-FundusImages). Results indicate that FM do not consistently outperform fine‑tuned CNNs; EfficientNet-B0 often matches or exceeds FM performance, with FLAIR being the most competitive FM.
By Franco Javier Arellano, Jos\'e Ignacio Orlando
We introduce SMART, a framework for learning a flexible, interpretable, and scalable spatio-temporal brain atlas from longitudinal high-resolution 3D medical images. Existing approaches to spatio-temporal atlas construction rely on black-box generative models that lack flexibility, limit interpretability, and struggle to scale to high-dimensional data.
arXiv:2609.31573v1 Announce Type: new
Abstract: Biomedical image segmentation is central to medical image analysis, but practical deployment often faces limited annotations, memory constraints, and c...
By Ziyao Shang, Pouya Sadeghi, Letian Jiang, Alexander Wong, Sirisha Rambhatla
arXiv:2604. 22700v2 Announce Type: replace Abstract: Modeling and predicting neurodegenerative disease progression from medical images remains a major challenge in medical AI, with significant implications for early diagnosis, disease monitoring, and treatment planning.
By Nivetha Jayakumar, Swakshar Deb, Bahram Jafrasteh, Qingyu Zhao, Miaomiao Zhang
OptiModNet is a lightweight UNet‑Transformer hybrid designed for optic disc and cup segmentation. It incorporates grouped‑query and channel attention across multiple stages, along with an Aggregated Pyramid Loss to improve gradient flow and structural consistency. Evaluated on the REFUGE2 dataset, it surpasses existing methods by over 2.5 % while using only 3.73 GFLOPs and 1.93 M parameters.
By Soumili Ghosh, Debapriya Roy, Aryan Das, Bikash Santra