Saved in:
Bibliographic Details
Main Authors: André, Pascaline, Heitz, Charles, Christodoulou, Evangelia, Reinke, Annika, Sudre, Carole H., Antonelli, Michela, Godau, Patrick, Cardoso, M. Jorge, Gilson, Antoine, Montcel, Sophie Tezenas du, Varoquaux, Gaël, Maier-Hein, Lena, Colliot, Olivier
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.17103
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911395708141568
author André, Pascaline
Heitz, Charles
Christodoulou, Evangelia
Reinke, Annika
Sudre, Carole H.
Antonelli, Michela
Godau, Patrick
Cardoso, M. Jorge
Gilson, Antoine
Montcel, Sophie Tezenas du
Varoquaux, Gaël
Maier-Hein, Lena
Colliot, Olivier
author_facet André, Pascaline
Heitz, Charles
Christodoulou, Evangelia
Reinke, Annika
Sudre, Carole H.
Antonelli, Michela
Godau, Patrick
Cardoso, M. Jorge
Gilson, Antoine
Montcel, Sophie Tezenas du
Varoquaux, Gaël
Maier-Hein, Lena
Colliot, Olivier
contents Performance uncertainty quantification is essential for reliable validation and eventual clinical translation of medical imaging artificial intelligence (AI). Confidence intervals (CIs) play a central role in this process by indicating how precise a reported performance estimate is. Yet, due to the limited amount of work examining CI behavior in medical imaging, the community remains largely unaware of how many diverse CI methods exist and how they behave in specific settings. The purpose of this study is to close this gap. To this end, we conducted a large-scale empirical analysis across a total of 24 segmentation and classification tasks, using 19 trained models per task group, a broad spectrum of commonly used performance metrics, multiple aggregation strategies, and several widely adopted CI methods. Reliability (coverage) and precision (width) of each CI method were estimated across all settings to characterize their dependence on study characteristics. Our analysis revealed five principal findings: 1) the sample size required for reliable CIs varies from a few dozens to several thousands of cases depending on study parameters; 2) CI behavior is strongly affected by the choice of performance metric; 3) aggregation strategy substantially influences the reliability of CIs, e.g. they require more observations for macro than for micro; 4) the machine learning problem (segmentation versus classification) modulates these effects; 5) different CI methods are not equally reliable and precise depending on the use case. These results form key components for the development of future guidelines on reporting performance uncertainty in medical imaging AI.
format Preprint
id arxiv_https___arxiv_org_abs_2601_17103
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Performance uncertainty in medical image analysis: a large-scale investigation of confidence intervals
André, Pascaline
Heitz, Charles
Christodoulou, Evangelia
Reinke, Annika
Sudre, Carole H.
Antonelli, Michela
Godau, Patrick
Cardoso, M. Jorge
Gilson, Antoine
Montcel, Sophie Tezenas du
Varoquaux, Gaël
Maier-Hein, Lena
Colliot, Olivier
Computer Vision and Pattern Recognition
Machine Learning
Performance uncertainty quantification is essential for reliable validation and eventual clinical translation of medical imaging artificial intelligence (AI). Confidence intervals (CIs) play a central role in this process by indicating how precise a reported performance estimate is. Yet, due to the limited amount of work examining CI behavior in medical imaging, the community remains largely unaware of how many diverse CI methods exist and how they behave in specific settings. The purpose of this study is to close this gap. To this end, we conducted a large-scale empirical analysis across a total of 24 segmentation and classification tasks, using 19 trained models per task group, a broad spectrum of commonly used performance metrics, multiple aggregation strategies, and several widely adopted CI methods. Reliability (coverage) and precision (width) of each CI method were estimated across all settings to characterize their dependence on study characteristics. Our analysis revealed five principal findings: 1) the sample size required for reliable CIs varies from a few dozens to several thousands of cases depending on study parameters; 2) CI behavior is strongly affected by the choice of performance metric; 3) aggregation strategy substantially influences the reliability of CIs, e.g. they require more observations for macro than for micro; 4) the machine learning problem (segmentation versus classification) modulates these effects; 5) different CI methods are not equally reliable and precise depending on the use case. These results form key components for the development of future guidelines on reporting performance uncertainty in medical imaging AI.
title Performance uncertainty in medical image analysis: a large-scale investigation of confidence intervals
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2601.17103