MammoClean: Toward Reproducible and Bias-Aware AI in Mammography through Dataset Harmonization

2025-11-04

AI Summary

Breast X-ray AI models suffer from poor generalizability and bias due to substantial heterogeneity across public datasets in image quality, metadata standards, and demographic representation—hindering clinical deployment. To address this, we propose the first open-source framework for standardization and bias quantification specifically designed for mammographic data, enabling systematic identification and mitigation of biases related to breast density, lesion distribution, and other domain-specific factors. Our framework harmonizes heterogeneous datasets—including CBIS-DDSM, CMMD, and VinDr-Mammo—via standardized case selection, left-right view correction, intensity normalization, and unified metadata annotation. Experiments demonstrate that the standardized data significantly reduces distributional shift: models trained on cleaned data achieve an average 3.2% AUC improvement on cross-domain test sets and exhibit a 41% reduction in bias metrics, concurrently enhancing both generalizability and fairness.

Technology Category

• Computer Vision: Bias, Fairness & Privacy
• Machine Learning: Calibration & Uncertainty Quantification
• Humans and AI: Explainable AI (XAI) for Human Understanding

Application Category

• Social Networks and Social Media: Fairness and bias in social network and social media analysis
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

The development of clinically reliable artificial intelligence (AI) systems for mammography is hindered by profound heterogeneity in data quality, metadata standards, and population distributions across public datasets. This heterogeneity introduces dataset-specific biases that severely compromise the generalizability of the model, a fundamental barrier to clinical deployment. We present MammoClean, a public framework for standardization and bias quantification in mammography datasets. MammoClean standardizes case selection, image processing (including laterality and intensity correction), and unifies metadata into a consistent multi-view structure. We provide a comprehensive review of breast anatomy, imaging characteristics, and public mammography datasets to systematically identify key sources of bias. Applying MammoClean to three heterogeneous datasets (CBIS-DDSM, TOMPEI-CMMD, VinDr-Mammo), we quantify substantial distributional shifts in breast density and abnormality prevalence. Critically, we demonstrate the direct impact of data corruption: AI models trained on corrupted datasets exhibit significant performance degradation compared to their curated counterparts. By using MammoClean to identify and mitigate bias sources, researchers can construct unified multi-dataset training corpora that enable development of robust models with superior cross-domain generalization. MammoClean provides an essential, reproducible pipeline for bias-aware AI development in mammography, facilitating fairer comparisons and advancing the creation of safe, effective systems that perform equitably across diverse patient populations and clinical settings. The open-source code is publicly available from: https://github.com/Minds-R-Lab/MammoClean.

Problem

• Addressing data heterogeneity and bias in mammography AI datasets
• Standardizing mammography data processing and metadata across diverse sources
• Improving AI model generalization through bias quantification and mitigation

Innovation

• Standardizes case selection and image processing
• Unifies metadata into consistent multi-view structure
• Quantifies and mitigates dataset-specific bias sources