Label-free estimation of clinically relevant performance metrics under distribution shifts

2025-07-30

AI Summary

In real-world deployment of medical image classification models, performance monitoring remains challenging due to distribution shift, severe class imbalance, and the absence of ground-truth labels. To address this, we propose a novel confidence-score-based method for estimating the full confusion matrix without requiring any labels—marking the first approach capable of directly predicting all confusion matrix entries in a label-free setting. Our method jointly models covariate and prevalence shift within a unified framework and is validated on chest X-ray data under both synthetic and real-world distribution shifts. Results demonstrate high accuracy in estimating clinically critical metrics (e.g., true positives and false negatives), substantially outperforming existing performance prediction methods. Moreover, our analysis uncovers a systematic failure mechanism of mainstream approaches under prevalence shift. This work establishes an interpretable, robust, and clinically deployable pathway for unsupervised monitoring of AI models in healthcare.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Interpretability, Explainability, and Transparency
• Reasoning under Uncertainty: Graphical Models

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Security and Privacy: Large-scale security measurements

Abstract

Performance monitoring is essential for safe clinical deployment of image classification models. However, because ground-truth labels are typically unavailable in the target dataset, direct assessment of real-world model performance is infeasible. State-of-the-art performance estimation methods address this by leveraging confidence scores to estimate the target accuracy. Despite being a promising direction, the established methods mainly estimate the model's accuracy and are rarely evaluated in a clinical domain, where strong class imbalances and dataset shifts are common. Our contributions are twofold: First, we introduce generalisations of existing performance prediction methods that directly estimate the full confusion matrix. Then, we benchmark their performance on chest x-ray data in real-world distribution shifts as well as simulated covariate and prevalence shifts. The proposed confusion matrix estimation methods reliably predicted clinically relevant counting metrics on medical images under distribution shifts. However, our simulated shift scenarios exposed important failure modes of current performance estimation techniques, calling for a better understanding of real-world deployment contexts when implementing these performance monitoring techniques for postmarket surveillance of medical AI models.

Problem

• Estimating clinical performance metrics without labels under distribution shifts
• Extending accuracy estimation to full confusion matrix prediction
• Evaluating methods on medical images with real and simulated shifts

Innovation

• Estimates full confusion matrix under shifts
• Benchmarks on chest x-ray distribution shifts
• Predicts clinical metrics reliably on medical images