Abstract
Cross-site generalizability in medical AI can be affected by selection bias, a structural mechanism where patient demographics (e.g., age, severity) non-randomly dictate hospital assignment. Conventional Domain Generalization (DG) paradigms, which predominantly target image-level distribution shifts, fail to address the resulting spurious correlations between site-specific variations and diagnostic labels. To address this challenge, we propose CIV-DG, a causal framework that leverages Conditional Instrumental Variables to reduce reliance on site-specific artifacts. By conditioning IV validity on observed demographics, CIV-DG targets complex clinical scenarios where hospital selection is endogenously driven by patient demographics. We instantiate this theory via a Deep Generalized Method of Moments (DeepGMM) architecture, employing a conditional critic to minimize moment violations and enforce instrument-error orthogonality within demographic strata. Extensive experiments on the Camelyon17 benchmark and large-scale Chest X-Ray datasets show improved predictive robustness, fairness, and calibration, supporting the utility of conditional moment constraints for medical AI.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/MLMI_023.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to Open Review
Open Review Page: https://openreview.net/forum?id=y5vliAlDZ6
BibTex
@InProceedings{BaiSha_CIVDG_MICCAISAT2026,
author = { Bai, Shaojin AND Su, Yuting AND Nie, Weizhi},
title = { { CIV-DG: Conditional Instrumental Variables for Domain Generalization in Medical Imaging } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17272},
month = {pending},
page = {pending}
}