Clinician-Centered Evaluation Framework for Explainable AI Heatmaps in OCT-Based Retinal Disease Classification
Eirini Maliagkani, Ilias Georgalas, Ioannis Datseris, Elpiniki Papageorgiou, Ioannis D. Apostolopoulos · 2026
Read Clinician-Centered Evaluation Framework for Explainable AI Heatmaps in OCT-Based Retinal Disease Classification by Eirini Maliagkani, Ilias Georgalas, Ioannis Datseris, Elpiniki Papageorgiou, Ioannis D. Apostolopoulos (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13207473.txt.
Description
Eirini Maliagkani and co-authors build a clinician-centered way to judge explanation heatmaps from retinal AI. Their 2026 paper concerns models that classify eye scans by disease and highlight the regions behind each call. A heatmap that looks right to engineers may still mislead the doctors who act on it.
Their framework runs in two phases. First, one eye specialist and one AI specialist screened a hundred heatmaps per method for quality and anatomical sense, cutting the field to three methods. Then a panel of specialists rated the finalists across cases for agreement with the diagnosis. The team's own contrast-based map scored highest, a popular gradient method followed, and a fusion map trailed. Ratings reflect clinical plausibility, guiding which explanations deserve trust.
Catalog record
| CREATOR | Eirini Maliagkani, Ilias Georgalas, Ioannis Datseris, Elpiniki Papageorgiou, Ioannis D. Apostolopoulos |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 54 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |