Medical imaging models are often deployed without demographic, acquisition, and quality metadata necessary for subgroup auditing. When such metadata is missing, clinically critical failure modes can be masked by strong aggregate performance, and many robust learning methods lose the group structure they depend on. We present CAPRA, a calibrated proxy-axis framework for hidden subgroup analysis under missing metadata. CAPRA predicts image-derived semantic axes, calibrates axis posteriors on a small metadata-labeled split via patient-level cross-fitting, and organizes those posteriors into a calibrated subgroup interface that supports both deployment-time failure analysis and downstream robust learning without requiring subgroup labels at deployment.
Across fundus, dermoscopy, and chest radiography, CAPRA reveals disparity patterns missed by metadata-only slicing, remains informative under dataset shift, and produces subgroup partitions that align more closely with explicit failure axes than image-only or latent-slice baselines. The same interface can also be reused by downstream robust learners, although those gains are domain-dependent. Overall, CAPRA turns hidden subgroup analysis under missing metadata into a calibrated, interpretable, and reusable subgroup interface for deployment-time analysis and robust transfer.
Blogger's Review: The introduction of CAPRA offers a novel solution in the field of medical imaging analysis, especially in scenarios with missing critical metadata. This calibrated framework enhances model interpretability and improves the ability to discover hidden subgroups, laying the groundwork for future robust learning. Its potential applications could significantly enhance diagnostic accuracy in medical imaging.