SAC-Diff: A Scan-Aware Consistency-Enhanced Diffusion Framework for Unsupervised Chest CT Anomaly Detection

Xinyuan Zheng, Yoshihisa Shinagawa, Sepehr Farhand, Chi Liu, Gerardo Hermosillo Valadez, Xueqi Guo
Proceedings of The 9th International Conference on Medical Imaging with Deep Learning, PMLR 315:3723-3749, 2026.

Abstract

Anomaly detection in medical imaging is important but challenging due to diverse and imbalanced pathologies. Supervised methods rely on large annotated datasets and generalize poorly to unseen conditions. Unsupervised generative methods, especially diffusion models, can learn normal anatomy and detect outliers, but often hallucinate because of the Gaussian noise design and insufficient anatomical guidance. To address these challenges, we propose SAC-Diff, a Scan-Aware Consistency-Enhanced Diffusion framework for unsupervised anomaly detection in automated lung disease screening using chest CT. SAC-Diff adopts simplex noise for detail-preserving diffusion perturbation, integrates scan awareness via (A) subject-aware anatomical priors into conditional diffusion and (B) background-aware masking for scan-specific variations and heterogeneous lung anomalies, and enhances robustness by enforcing consistency and quantifying uncertainty through multi-sample ensembling. We evaluate SAC-Diff on two diseased datasets with various anomalies, COVID-19 and interstitial lung disease (ILD), and observe substantial improvements over prior methods. On COVID-19, SAC-Diff achieves an IoU of 0.39 (+3.75% improvement compared to existing methods) and Dice of 0.53 (+2.99%); on ILD, it improves IoU to 0.31 (+74.45%) and Dice to 0.44 (+60.40%). Our results demonstrate promise toward robust and annotation-free CT anomaly detection in hospital deployment.

Cite this Paper

BibTeX
@InProceedings{pmlr-v315-zheng26b, title = {SAC-Diff: A Scan-Aware Consistency-Enhanced Diffusion Framework for Unsupervised Chest CT Anomaly Detection}, author = {Zheng, Xinyuan and Shinagawa, Yoshihisa and Farhand, Sepehr and Liu, Chi and Valadez, Gerardo Hermosillo and Guo, Xueqi}, booktitle = {Proceedings of The 9th International Conference on Medical Imaging with Deep Learning}, pages = {3723--3749}, year = {2026}, editor = {Huo, Yuankai and Gao, Mingchen and Kuo, Chang-Fu and Jin, Yueming and Deng, Ruining}, volume = {315}, series = {Proceedings of Machine Learning Research}, month = {08--10 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v315/main/assets/zheng26b/zheng26b.pdf}, url = {https://proceedings.mlr.press/v315/zheng26b.html}, abstract = {Anomaly detection in medical imaging is important but challenging due to diverse and imbalanced pathologies. Supervised methods rely on large annotated datasets and generalize poorly to unseen conditions. Unsupervised generative methods, especially diffusion models, can learn normal anatomy and detect outliers, but often hallucinate because of the Gaussian noise design and insufficient anatomical guidance. To address these challenges, we propose SAC-Diff, a Scan-Aware Consistency-Enhanced Diffusion framework for unsupervised anomaly detection in automated lung disease screening using chest CT. SAC-Diff adopts simplex noise for detail-preserving diffusion perturbation, integrates scan awareness via (A) subject-aware anatomical priors into conditional diffusion and (B) background-aware masking for scan-specific variations and heterogeneous lung anomalies, and enhances robustness by enforcing consistency and quantifying uncertainty through multi-sample ensembling. We evaluate SAC-Diff on two diseased datasets with various anomalies, COVID-19 and interstitial lung disease (ILD), and observe substantial improvements over prior methods. On COVID-19, SAC-Diff achieves an IoU of 0.39 (+3.75% improvement compared to existing methods) and Dice of 0.53 (+2.99%); on ILD, it improves IoU to 0.31 (+74.45%) and Dice to 0.44 (+60.40%). Our results demonstrate promise toward robust and annotation-free CT anomaly detection in hospital deployment.} }
Endnote
%0 Conference Paper %T SAC-Diff: A Scan-Aware Consistency-Enhanced Diffusion Framework for Unsupervised Chest CT Anomaly Detection %A Xinyuan Zheng %A Yoshihisa Shinagawa %A Sepehr Farhand %A Chi Liu %A Gerardo Hermosillo Valadez %A Xueqi Guo %B Proceedings of The 9th International Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2026 %E Yuankai Huo %E Mingchen Gao %E Chang-Fu Kuo %E Yueming Jin %E Ruining Deng %F pmlr-v315-zheng26b %I PMLR %P 3723--3749 %U https://proceedings.mlr.press/v315/zheng26b.html %V 315 %X Anomaly detection in medical imaging is important but challenging due to diverse and imbalanced pathologies. Supervised methods rely on large annotated datasets and generalize poorly to unseen conditions. Unsupervised generative methods, especially diffusion models, can learn normal anatomy and detect outliers, but often hallucinate because of the Gaussian noise design and insufficient anatomical guidance. To address these challenges, we propose SAC-Diff, a Scan-Aware Consistency-Enhanced Diffusion framework for unsupervised anomaly detection in automated lung disease screening using chest CT. SAC-Diff adopts simplex noise for detail-preserving diffusion perturbation, integrates scan awareness via (A) subject-aware anatomical priors into conditional diffusion and (B) background-aware masking for scan-specific variations and heterogeneous lung anomalies, and enhances robustness by enforcing consistency and quantifying uncertainty through multi-sample ensembling. We evaluate SAC-Diff on two diseased datasets with various anomalies, COVID-19 and interstitial lung disease (ILD), and observe substantial improvements over prior methods. On COVID-19, SAC-Diff achieves an IoU of 0.39 (+3.75% improvement compared to existing methods) and Dice of 0.53 (+2.99%); on ILD, it improves IoU to 0.31 (+74.45%) and Dice to 0.44 (+60.40%). Our results demonstrate promise toward robust and annotation-free CT anomaly detection in hospital deployment.
APA
Zheng, X., Shinagawa, Y., Farhand, S., Liu, C., Valadez, G.H. & Guo, X.. (2026). SAC-Diff: A Scan-Aware Consistency-Enhanced Diffusion Framework for Unsupervised Chest CT Anomaly Detection. Proceedings of The 9th International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 315:3723-3749 Available from https://proceedings.mlr.press/v315/zheng26b.html.

Related Material