-
Towards objective and systematic evaluation of bias in artificial intelligence for medical imaging
Authors:
Emma A. M. Stanley,
Raissa Souza,
Anthony Winder,
Vedant Gulve,
Kimberly Amador,
Matthias Wilms,
Nils D. Forkert
Abstract:
Artificial intelligence (AI) models trained using medical images for clinical tasks often exhibit bias in the form of disparities in performance between subgroups. Since not all sources of biases in real-world medical imaging data are easily identifiable, it is challenging to comprehensively assess how those biases are encoded in models, and how capable bias mitigation methods are at ameliorating…
▽ More
Artificial intelligence (AI) models trained using medical images for clinical tasks often exhibit bias in the form of disparities in performance between subgroups. Since not all sources of biases in real-world medical imaging data are easily identifiable, it is challenging to comprehensively assess how those biases are encoded in models, and how capable bias mitigation methods are at ameliorating performance disparities. In this article, we introduce a novel analysis framework for systematically and objectively investigating the impact of biases in medical images on AI models. We developed and tested this framework for conducting controlled in silico trials to assess bias in medical imaging AI using a tool for generating synthetic magnetic resonance images with known disease effects and sources of bias. The feasibility is showcased by using three counterfactual bias scenarios to measure the impact of simulated bias effects on a convolutional neural network (CNN) classifier and the efficacy of three bias mitigation strategies. The analysis revealed that the simulated biases resulted in expected subgroup performance disparities when the CNN was trained on the synthetic datasets. Moreover, reweighing was identified as the most successful bias mitigation strategy for this setup, and we demonstrated how explainable AI methods can aid in investigating the manifestation of bias in the model using this framework. Develo** fair AI models is a considerable challenge given that many and often unknown sources of biases can be present in medical imaging datasets. In this work, we present a novel methodology to objectively study the impact of biases and mitigation strategies on deep learning pipelines, which can support the development of clinical AI that is robust and responsible.
△ Less
Submitted 1 July, 2024; v1 submitted 2 November, 2023;
originally announced November 2023.
-
Bidirectional Modeling and Analysis of Brain Aging with Normalizing Flows
Authors:
Matthias Wilms,
Jordan J. Bannister,
Pauline Mouches,
M. Ethan MacDonald,
Deepthi Rajashekar,
Sönke Langner,
Nils D. Forkert
Abstract:
Brain aging is a widely studied longitudinal process throughout which the brain undergoes considerable morphological changes and various machine learning approaches have been proposed to analyze it. Within this context, brain age prediction from structural MR images and age-specific brain morphology template generation are two problems that have attracted much attention. While most approaches tack…
▽ More
Brain aging is a widely studied longitudinal process throughout which the brain undergoes considerable morphological changes and various machine learning approaches have been proposed to analyze it. Within this context, brain age prediction from structural MR images and age-specific brain morphology template generation are two problems that have attracted much attention. While most approaches tackle these tasks independently, we assume that they are inverse directions of the same functional bidirectional relationship between a brain's morphology and an age variable. In this paper, we propose to model this relationship with a single conditional normalizing flow, which unifies brain age prediction and age-conditioned generative modeling in a novel way. In an initial evaluation of this idea, we show that our normalizing flow brain aging model can accurately predict brain age while also being able to generate age-specific brain morphology templates that realistically represent the typical aging trend in a healthy population. This work is a step towards unified modeling of functional relationships between 3D brain morphology and clinical variables of interest with powerful normalizing flows.
△ Less
Submitted 26 November, 2020;
originally announced November 2020.
-
Population-based Respiratory 4D Motion Atlas Construction and its Application for VR Simulations of Liver Punctures
Authors:
Andre Mastmeyer,
Matthias Wilms,
Heinz Handels
Abstract:
Virtual reality (VR) training simulators of liver needle insertion in the hepatic area of breathing virtual patients currently need 4D data acquisitions as a prerequisite. Here, first a population-based breathing virtual patient 4D atlas can be built and second the requirement of a dose-relevant or expensive acquisition of a 4D data set for a new static 3D patient can be mitigated by war** the m…
▽ More
Virtual reality (VR) training simulators of liver needle insertion in the hepatic area of breathing virtual patients currently need 4D data acquisitions as a prerequisite. Here, first a population-based breathing virtual patient 4D atlas can be built and second the requirement of a dose-relevant or expensive acquisition of a 4D data set for a new static 3D patient can be mitigated by war** the mean atlas motion. The breakthrough contribution of this work is the construction and reuse of population-based learned 4D motion models.
△ Less
Submitted 22 December, 2017; v1 submitted 5 December, 2017;
originally announced December 2017.
-
Interpatient Respiratory Motion Model Transfer for Virtual Reality Simulations of Liver Punctures
Authors:
Andre Mastmeyer,
Matthias Wilms,
Heinz Handels
Abstract:
Current virtual reality (VR) training simulators of liver punctures often rely on static 3D patient data and use an unrealistic (sinusoidal) periodic animation of the respiratory movement. Existing methods for the animation of breathing motion support simple mathematical or patient-specific, estimated breathing models. However with personalized breathing models for each new patient, a heavily dose…
▽ More
Current virtual reality (VR) training simulators of liver punctures often rely on static 3D patient data and use an unrealistic (sinusoidal) periodic animation of the respiratory movement. Existing methods for the animation of breathing motion support simple mathematical or patient-specific, estimated breathing models. However with personalized breathing models for each new patient, a heavily dose relevant or expensive 4D data acquisition is mandatory for keyframe-based motion modeling. Given the reference 4D data, first a model building stage using linear regression motion field modeling takes place. Then the methodology shown here allows the transfer of existing reference respiratory motion models of a 4D reference patient to a new static 3D patient. This goal is achieved by using non-linear inter-patient registration to warp one personalized 4D motion field model to new 3D patient data. This cost- and dose-saving new method is shown here visually in a qualitative proof-of-concept study.
△ Less
Submitted 2 August, 2017; v1 submitted 26 July, 2017;
originally announced July 2017.
