Predict Patient Self-reported Race from Skin Histological Images

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chen, Shengjia, Verma, Ruchika, Clare, Kevin, Jegminat, Jannes, Alleva, Eugenia, Huang, Kuan-lin, Veremis, Brandon, Fuchs, Thomas, Campanella, Gabriele
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912510592942080
author Chen, Shengjia
Verma, Ruchika
Clare, Kevin
Jegminat, Jannes
Alleva, Eugenia
Huang, Kuan-lin
Veremis, Brandon
Fuchs, Thomas
Campanella, Gabriele
author_facet Chen, Shengjia
Verma, Ruchika
Clare, Kevin
Jegminat, Jannes
Alleva, Eugenia
Huang, Kuan-lin
Veremis, Brandon
Fuchs, Thomas
Campanella, Gabriele
contents Artificial Intelligence (AI) has demonstrated success in computational pathology (CPath) for disease detection, biomarker classification, and prognosis prediction. However, its potential to learn unintended demographic biases, particularly those related to social determinants of health, remains understudied. This study investigates whether deep learning models can predict self-reported race from digitized dermatopathology slides and identifies potential morphological shortcuts. Using a multisite dataset with a racially diverse population, we apply an attention-based mechanism to uncover race-associated morphological features. After evaluating three dataset curation strategies to control for confounding factors, the final experiment showed that White and Black demographic groups retained high prediction performance (AUC: 0.799, 0.762), while overall performance dropped to 0.663. Attention analysis revealed the epidermis as a key predictive feature, with significant performance declines when these regions were removed. These findings highlight the need for careful data curation and bias mitigation to ensure equitable AI deployment in pathology. Code available at: https://github.com/sinai-computational-pathology/CPath_SAIF.
format Preprint
id arxiv_https___arxiv_org_abs_2507_21912
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Predict Patient Self-reported Race from Skin Histological Images
Chen, Shengjia
Verma, Ruchika
Clare, Kevin
Jegminat, Jannes
Alleva, Eugenia
Huang, Kuan-lin
Veremis, Brandon
Fuchs, Thomas
Campanella, Gabriele
Computer Vision and Pattern Recognition
Computational Engineering, Finance, and Science
Artificial Intelligence (AI) has demonstrated success in computational pathology (CPath) for disease detection, biomarker classification, and prognosis prediction. However, its potential to learn unintended demographic biases, particularly those related to social determinants of health, remains understudied. This study investigates whether deep learning models can predict self-reported race from digitized dermatopathology slides and identifies potential morphological shortcuts. Using a multisite dataset with a racially diverse population, we apply an attention-based mechanism to uncover race-associated morphological features. After evaluating three dataset curation strategies to control for confounding factors, the final experiment showed that White and Black demographic groups retained high prediction performance (AUC: 0.799, 0.762), while overall performance dropped to 0.663. Attention analysis revealed the epidermis as a key predictive feature, with significant performance declines when these regions were removed. These findings highlight the need for careful data curation and bias mitigation to ensure equitable AI deployment in pathology. Code available at: https://github.com/sinai-computational-pathology/CPath_SAIF.
title Predict Patient Self-reported Race from Skin Histological Images
topic Computer Vision and Pattern Recognition
Computational Engineering, Finance, and Science
url https://arxiv.org/abs/2507.21912