Enhancing the fairness of AI prediction models by Quasi-Pareto improvement among heterogeneous thyroid nodule population
Abstract Artificial Intelligence (AI) models for medical diagnosis often face challenges of generalizability and fairness. We highlighted the algorithmic unfairness in a large thyroid ultrasound dataset with significant diagnostic performance disparities across subgroups linked causally to sample si...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Nature Portfolio
2024-03-01
|
Series: | Nature Communications |
Online Access: | https://doi.org/10.1038/s41467-024-44906-y |
_version_ | 1797273943207313408 |
---|---|
author | Siqiong Yao Fang Dai Peng Sun Weituo Zhang Biyun Qian Hui Lu |
author_facet | Siqiong Yao Fang Dai Peng Sun Weituo Zhang Biyun Qian Hui Lu |
author_sort | Siqiong Yao |
collection | DOAJ |
description | Abstract Artificial Intelligence (AI) models for medical diagnosis often face challenges of generalizability and fairness. We highlighted the algorithmic unfairness in a large thyroid ultrasound dataset with significant diagnostic performance disparities across subgroups linked causally to sample size imbalances. To address this, we introduced the Quasi-Pareto Improvement (QPI) approach and a deep learning implementation (QP-Net) combining multi-task learning and domain adaptation to improve model performance among disadvantaged subgroups without compromising overall population performance. On the thyroid ultrasound dataset, our method significantly mitigated the area under curve (AUC) disparity for three less-prevalent subgroups by 0.213, 0.112, and 0.173 while maintaining the AUC for dominant subgroups; we also further confirmed the generalizability of our approach on two public datasets: the ISIC2019 skin disease dataset and the CheXpert chest radiograph dataset. Here we show the QPI approach to be widely applicable in promoting AI for equitable healthcare outcomes. |
first_indexed | 2024-03-07T14:51:21Z |
format | Article |
id | doaj.art-396ab19af8774c888b8672407200a9fb |
institution | Directory Open Access Journal |
issn | 2041-1723 |
language | English |
last_indexed | 2024-03-07T14:51:21Z |
publishDate | 2024-03-01 |
publisher | Nature Portfolio |
record_format | Article |
series | Nature Communications |
spelling | doaj.art-396ab19af8774c888b8672407200a9fb2024-03-05T19:41:28ZengNature PortfolioNature Communications2041-17232024-03-0115111310.1038/s41467-024-44906-yEnhancing the fairness of AI prediction models by Quasi-Pareto improvement among heterogeneous thyroid nodule populationSiqiong Yao0Fang Dai1Peng Sun2Weituo Zhang3Biyun Qian4Hui Lu5State Key Laboratory of Microbial Metabolism, Joint International Research Laboratory of Metabolic and Developmental Sciences, School of Life Sciences and Biotechnology, Shanghai Jiao Tong UniversityState Key Laboratory of Microbial Metabolism, Joint International Research Laboratory of Metabolic and Developmental Sciences, School of Life Sciences and Biotechnology, Shanghai Jiao Tong UniversityState Key Laboratory of Microbial Metabolism, Joint International Research Laboratory of Metabolic and Developmental Sciences, School of Life Sciences and Biotechnology, Shanghai Jiao Tong UniversityHongqiao International Institute of Medicine, Shanghai Tong Ren Hospital and School of Public Health, Shanghai Jiao Tong University School of MedicineHongqiao International Institute of Medicine, Shanghai Tong Ren Hospital and School of Public Health, Shanghai Jiao Tong University School of MedicineState Key Laboratory of Microbial Metabolism, Joint International Research Laboratory of Metabolic and Developmental Sciences, School of Life Sciences and Biotechnology, Shanghai Jiao Tong UniversityAbstract Artificial Intelligence (AI) models for medical diagnosis often face challenges of generalizability and fairness. We highlighted the algorithmic unfairness in a large thyroid ultrasound dataset with significant diagnostic performance disparities across subgroups linked causally to sample size imbalances. To address this, we introduced the Quasi-Pareto Improvement (QPI) approach and a deep learning implementation (QP-Net) combining multi-task learning and domain adaptation to improve model performance among disadvantaged subgroups without compromising overall population performance. On the thyroid ultrasound dataset, our method significantly mitigated the area under curve (AUC) disparity for three less-prevalent subgroups by 0.213, 0.112, and 0.173 while maintaining the AUC for dominant subgroups; we also further confirmed the generalizability of our approach on two public datasets: the ISIC2019 skin disease dataset and the CheXpert chest radiograph dataset. Here we show the QPI approach to be widely applicable in promoting AI for equitable healthcare outcomes.https://doi.org/10.1038/s41467-024-44906-y |
spellingShingle | Siqiong Yao Fang Dai Peng Sun Weituo Zhang Biyun Qian Hui Lu Enhancing the fairness of AI prediction models by Quasi-Pareto improvement among heterogeneous thyroid nodule population Nature Communications |
title | Enhancing the fairness of AI prediction models by Quasi-Pareto improvement among heterogeneous thyroid nodule population |
title_full | Enhancing the fairness of AI prediction models by Quasi-Pareto improvement among heterogeneous thyroid nodule population |
title_fullStr | Enhancing the fairness of AI prediction models by Quasi-Pareto improvement among heterogeneous thyroid nodule population |
title_full_unstemmed | Enhancing the fairness of AI prediction models by Quasi-Pareto improvement among heterogeneous thyroid nodule population |
title_short | Enhancing the fairness of AI prediction models by Quasi-Pareto improvement among heterogeneous thyroid nodule population |
title_sort | enhancing the fairness of ai prediction models by quasi pareto improvement among heterogeneous thyroid nodule population |
url | https://doi.org/10.1038/s41467-024-44906-y |
work_keys_str_mv | AT siqiongyao enhancingthefairnessofaipredictionmodelsbyquasiparetoimprovementamongheterogeneousthyroidnodulepopulation AT fangdai enhancingthefairnessofaipredictionmodelsbyquasiparetoimprovementamongheterogeneousthyroidnodulepopulation AT pengsun enhancingthefairnessofaipredictionmodelsbyquasiparetoimprovementamongheterogeneousthyroidnodulepopulation AT weituozhang enhancingthefairnessofaipredictionmodelsbyquasiparetoimprovementamongheterogeneousthyroidnodulepopulation AT biyunqian enhancingthefairnessofaipredictionmodelsbyquasiparetoimprovementamongheterogeneousthyroidnodulepopulation AT huilu enhancingthefairnessofaipredictionmodelsbyquasiparetoimprovementamongheterogeneousthyroidnodulepopulation |