Cautious optimism on foundation models in medical imaging balancing privacy and innovation

Foundation models for medical imaging offer transformative healthcare potential but risk patient re-identification through latent demographic and identity signals retained during training. A recent study in npj Digital Medicine reports re-identification rates up to 94 percent in retinal imaging using FMs, yet does not disentangle causality or propose mitigation strategies. Urgent technical safeguards and policy frameworks are needed to balance innovation with individual privacy and protect confidentiality. Foundation models for medical imaging analysis demonstrate remarkable screening and diagnostic potential but also raise novel privacy concerns, especially given the vast information such systems encode during their training steps. A recent study by N... [13101 chars]