Evaluating Sex and Age Biases in Multimodal Large Language Models for Skin Disease Identification from Dermatoscopic Images.

Health data science Pub Date : 2025-04-01 eCollection Date: 2025-01-01 DOI:10.34133/hds.0256
Zhiyu Wan, Yuhang Guo, Shunxing Bao, Qian Wang, Bradley A Malin
{"title":"Evaluating Sex and Age Biases in Multimodal Large Language Models for Skin Disease Identification from Dermatoscopic Images.","authors":"Zhiyu Wan, Yuhang Guo, Shunxing Bao, Qian Wang, Bradley A Malin","doi":"10.34133/hds.0256","DOIUrl":null,"url":null,"abstract":"<p><p><b>Background:</b> Multimodal large language models (LLMs) have shown potential in various health-related fields. However, many healthcare studies have raised concerns about the reliability and biases of LLMs in healthcare applications. <b>Methods:</b> To explore the practical application of multimodal LLMs in skin disease identification, and to evaluate sex and age biases, we tested the performance of 2 popular multimodal LLMs, ChatGPT-4 and LLaVA-1.6, across diverse sex and age groups using a subset of a large dermatoscopic dataset containing around 10,000 images and 3 skin diseases (melanoma, melanocytic nevi, and benign keratosis-like lesions). <b>Results:</b> In comparison to 3 deep learning models (VGG16, ResNet50, and Model Derm) based on convolutional neural network (CNN), one vision transformer model (Swin-B), we found that ChatGPT-4 and LLaVA-1.6 demonstrated overall accuracies that were 3% and 23% higher (and F1-scores that were 4% and 34% higher), respectively, than the best performing CNN-based baseline while maintaining accuracies that were 38% and 26% lower (and F1-scores that were 38% and 19% lower), respectively, than Swin-B. Meanwhile, ChatGPT-4 is generally unbiased in identifying these skin diseases across sex and age groups, while LLaVA-1.6 is generally unbiased across age groups, in contrast to Swin-B, which is biased in identifying melanocytic nevi. <b>Conclusions:</b> This study suggests the usefulness and fairness of LLMs in dermatological applications, aiding physicians and practitioners with diagnostic recommendations and patient screening. To further verify and evaluate the reliability and fairness of LLMs in healthcare, experiments using larger and more diverse datasets need to be performed in the future.</p>","PeriodicalId":73207,"journal":{"name":"Health data science","volume":"5 ","pages":"0256"},"PeriodicalIF":0.0000,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11961048/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Health data science","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.34133/hds.0256","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2025/1/1 0:00:00","PubModel":"eCollection","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

Abstract

Background: Multimodal large language models (LLMs) have shown potential in various health-related fields. However, many healthcare studies have raised concerns about the reliability and biases of LLMs in healthcare applications. Methods: To explore the practical application of multimodal LLMs in skin disease identification, and to evaluate sex and age biases, we tested the performance of 2 popular multimodal LLMs, ChatGPT-4 and LLaVA-1.6, across diverse sex and age groups using a subset of a large dermatoscopic dataset containing around 10,000 images and 3 skin diseases (melanoma, melanocytic nevi, and benign keratosis-like lesions). Results: In comparison to 3 deep learning models (VGG16, ResNet50, and Model Derm) based on convolutional neural network (CNN), one vision transformer model (Swin-B), we found that ChatGPT-4 and LLaVA-1.6 demonstrated overall accuracies that were 3% and 23% higher (and F1-scores that were 4% and 34% higher), respectively, than the best performing CNN-based baseline while maintaining accuracies that were 38% and 26% lower (and F1-scores that were 38% and 19% lower), respectively, than Swin-B. Meanwhile, ChatGPT-4 is generally unbiased in identifying these skin diseases across sex and age groups, while LLaVA-1.6 is generally unbiased across age groups, in contrast to Swin-B, which is biased in identifying melanocytic nevi. Conclusions: This study suggests the usefulness and fairness of LLMs in dermatological applications, aiding physicians and practitioners with diagnostic recommendations and patient screening. To further verify and evaluate the reliability and fairness of LLMs in healthcare, experiments using larger and more diverse datasets need to be performed in the future.

背景:多模态大语言模型(LLM)已在多个健康相关领域显示出潜力。然而,许多医疗保健研究对 LLM 在医疗保健应用中的可靠性和偏差表示担忧。研究方法为了探索多模态 LLM 在皮肤病识别中的实际应用,并评估性别和年龄偏差,我们使用包含约 10,000 张图像和 3 种皮肤病(黑色素瘤、黑素细胞痣和良性角化病样病变)的大型皮肤镜数据集的子集,测试了 2 种流行的多模态 LLM(ChatGPT-4 和 LLaVA-1.6)在不同性别和年龄组中的性能。结果与 3 个基于卷积神经网络(CNN)的深度学习模型(VGG16、ResNet50 和 Model Derm)和 1 个视觉转换器模型(Swin-B)相比,我们发现 ChatGPT-4 和 LLaVA-1.6 的总体准确率分别比基于 CNN 的最佳基线高出 3% 和 23%(F1 分数分别高出 4% 和 34%),而准确率则分别比 Swin-B 低 38% 和 26%(F1 分数分别低 38% 和 19%)。同时,ChatGPT-4 在跨性别和年龄组识别这些皮肤病方面基本无偏见,而 LLaVA-1.6 在跨年龄组识别这些皮肤病方面基本无偏见,这与 Swin-B 形成鲜明对比,后者在识别黑素细胞痣方面存在偏差。结论:这项研究表明,LLMs 在皮肤科应用中是有用和公平的,可以帮助医生和从业人员提出诊断建议和筛查病人。为了进一步验证和评估 LLM 在医疗保健领域的可靠性和公平性,今后需要使用更大、更多样化的数据集进行实验。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
CiteScore
3.70
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信