Unsupervised Anomaly Detection for Image Dataset Quality Assurance in Multi-Center Breast MRI
2026-08-17 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionArtificial Intelligence
AI summaryⓘ
The authors studied how to automatically find bad or unusual data in breast MRI scans that could mess up medical AI. They created a test set with different types of mistakes and anomalies from several datasets to see how well various detection methods work. They found some methods are good at spotting very different data but struggle with data that looks almost normal, especially from new hospitals or unusual cases like implants. Their work helps guide future efforts to automatically check medical imaging data quality without needing labeled examples.
anomaly detectionout-of-distribution detectionbreast MRIdataset quality assurance3D reconstructionunsupervised learningmedical AIdynamic contrast-enhanced MRIAUROCdomain adaptation
Authors
Chiara Tappermann, Steffen Renisch, Lars Ole Schwen, Hans Meine, Horst K. Hahn, Eike Petersen
Abstract
Corrupted, inconsistent, or anomalous data silently threatens the safety and reliability of medical AI. Despite growing regulatory recognition of dataset quality assurance (QA) for high-risk medical AI, scalable automated detection remains underdeveloped. We employ unsupervised anomaly detection (AD) and out-of-distribution (OOD) detection as an automated dataset QA mechanism for multi-center dynamic contrast-enhanced breast MRI. We build a controlled AD benchmark of 17 realistic QA-relevant anomaly types from six public datasets (protocol violations, processing errors, incorrect anatomical regions) and propose a taxonomy of radiological image anomalies based on human visual perception, enabling fine-grained analysis of AD failure modes. The benchmark includes near-, medium-far-, far-OOD samples, as well as in-distribution and external normal data. Four methods are evaluated: a projection-based method extended with a domain-specific feature extractor and a novel positional encoding, a reconstruction-based approach extended to full 3D volumes with an augmented training objective, and two unmodified hybrid OOD detection methods. Medium-far- and far-OOD samples are detected reliably, whereas near-OOD samples and external normal data from unseen institutions expose method-specific differences. The 3D reconstruction-based approach best balances detection performance (AUROC: 0.936) and generalization to unseen institutions. The projection-based method with positional encoding achieves the highest overall detection performance (AUROC: 0.954). Both hybrid methods exhibit critical failure modes, confirming that methods validated for one modality or anatomy may not generalize without domain-specific adaptation. Implants and mastectomies remain an open challenge for all methods. Our results establish a foundation and practical guidance on scalable unsupervised QA in medical AI pipelines.