A concept-based interpretable model for the diagnosis of choroid neoplasias using multimodal data

Abstract Diagnosing rare diseases remains a critical challenge in clinical practice, often requiring specialist expertise. Despite the promising potential of machine learning, the scarcity of data on rare diseases and the need for interpretable, reliable artificial intelligence (AI) models complicat...

Full description

Saved in:
Bibliographic Details
Main Authors: Yifan Wu, Yang Liu, Yue Yang, Michael S. Yao, Wenli Yang, Xuehui Shi, Lihong Yang, Dongjun Li, Yueming Liu, Shiyi Yin, Chunyan Lei, Meixia Zhang, James C. Gee, Xuan Yang, Wenbin Wei, Shi Gu
Format: Article
Language:English
Published: Nature Portfolio 2025-04-01
Series:Nature Communications
Online Access:https://doi.org/10.1038/s41467-025-58801-7
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849310321257218048
author Yifan Wu
Yang Liu
Yue Yang
Michael S. Yao
Wenli Yang
Xuehui Shi
Lihong Yang
Dongjun Li
Yueming Liu
Shiyi Yin
Chunyan Lei
Meixia Zhang
James C. Gee
Xuan Yang
Wenbin Wei
Shi Gu
author_facet Yifan Wu
Yang Liu
Yue Yang
Michael S. Yao
Wenli Yang
Xuehui Shi
Lihong Yang
Dongjun Li
Yueming Liu
Shiyi Yin
Chunyan Lei
Meixia Zhang
James C. Gee
Xuan Yang
Wenbin Wei
Shi Gu
author_sort Yifan Wu
collection DOAJ
description Abstract Diagnosing rare diseases remains a critical challenge in clinical practice, often requiring specialist expertise. Despite the promising potential of machine learning, the scarcity of data on rare diseases and the need for interpretable, reliable artificial intelligence (AI) models complicates development. This study introduces a multimodal concept-based interpretable model tailored to distinguish uveal melanoma (0.4-0.6 per million in Asians) from hemangioma and metastatic carcinoma following the clinical practice. We collected a comprehensive dataset on Asians to date on choroid neoplasm imaging with radiological reports, encompassing over 750 patients from 2013 to 2019. Our model integrates domain expert insights from radiological reports and differentiates between three types of choroidal tumors, achieving an F 1 score of 0.91. This performance not only matches senior ophthalmologists but also improves the diagnostic accuracy of less experienced clinicians by 42%. The results underscore the potential of interpretable AI to enhance rare disease diagnosis and pave the way for future advancements in medical AI.
format Article
id doaj-art-e69f1e69bf9c4bafa8acd37e6ee6be86
institution Kabale University
issn 2041-1723
language English
publishDate 2025-04-01
publisher Nature Portfolio
record_format Article
series Nature Communications
spelling doaj-art-e69f1e69bf9c4bafa8acd37e6ee6be862025-08-20T03:53:46ZengNature PortfolioNature Communications2041-17232025-04-0116111410.1038/s41467-025-58801-7A concept-based interpretable model for the diagnosis of choroid neoplasias using multimodal dataYifan Wu0Yang Liu1Yue Yang2Michael S. Yao3Wenli Yang4Xuehui Shi5Lihong Yang6Dongjun Li7Yueming Liu8Shiyi Yin9Chunyan Lei10Meixia Zhang11James C. Gee12Xuan Yang13Wenbin Wei14Shi Gu15University of PennsylvaniaUniversity of Electronic Science and Technology of ChinaUniversity of PennsylvaniaUniversity of PennsylvaniaBeijing Tongren Eye Center, Beijing Tongren Hospital, Capital Medical UniversityBeijing Tongren Eye Center, Beijing Tongren Hospital, Capital Medical UniversityBeijing Tongren Eye Center, Beijing Tongren Hospital, Capital Medical UniversityBeijing Tongren Eye Center, Beijing Tongren Hospital, Capital Medical UniversityBeijing Tongren Eye Center, Beijing Tongren Hospital, Capital Medical UniversityBeijing Tongren Eye Center, Beijing Tongren Hospital, Capital Medical UniversityDepartment of Ophthalmology and Research Laboratory of Macular Disease, West China Hospital, Sichuan UniversityDepartment of Ophthalmology and Research Laboratory of Macular Disease, West China Hospital, Sichuan UniversityUniversity of PennsylvaniaBeijing Tongren Eye Center, Beijing Tongren Hospital, Capital Medical UniversityBeijing Tongren Eye Center, Beijing Tongren Hospital, Capital Medical UniversityUniversity of Electronic Science and Technology of ChinaAbstract Diagnosing rare diseases remains a critical challenge in clinical practice, often requiring specialist expertise. Despite the promising potential of machine learning, the scarcity of data on rare diseases and the need for interpretable, reliable artificial intelligence (AI) models complicates development. This study introduces a multimodal concept-based interpretable model tailored to distinguish uveal melanoma (0.4-0.6 per million in Asians) from hemangioma and metastatic carcinoma following the clinical practice. We collected a comprehensive dataset on Asians to date on choroid neoplasm imaging with radiological reports, encompassing over 750 patients from 2013 to 2019. Our model integrates domain expert insights from radiological reports and differentiates between three types of choroidal tumors, achieving an F 1 score of 0.91. This performance not only matches senior ophthalmologists but also improves the diagnostic accuracy of less experienced clinicians by 42%. The results underscore the potential of interpretable AI to enhance rare disease diagnosis and pave the way for future advancements in medical AI.https://doi.org/10.1038/s41467-025-58801-7
spellingShingle Yifan Wu
Yang Liu
Yue Yang
Michael S. Yao
Wenli Yang
Xuehui Shi
Lihong Yang
Dongjun Li
Yueming Liu
Shiyi Yin
Chunyan Lei
Meixia Zhang
James C. Gee
Xuan Yang
Wenbin Wei
Shi Gu
A concept-based interpretable model for the diagnosis of choroid neoplasias using multimodal data
Nature Communications
title A concept-based interpretable model for the diagnosis of choroid neoplasias using multimodal data
title_full A concept-based interpretable model for the diagnosis of choroid neoplasias using multimodal data
title_fullStr A concept-based interpretable model for the diagnosis of choroid neoplasias using multimodal data
title_full_unstemmed A concept-based interpretable model for the diagnosis of choroid neoplasias using multimodal data
title_short A concept-based interpretable model for the diagnosis of choroid neoplasias using multimodal data
title_sort concept based interpretable model for the diagnosis of choroid neoplasias using multimodal data
url https://doi.org/10.1038/s41467-025-58801-7
work_keys_str_mv AT yifanwu aconceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT yangliu aconceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT yueyang aconceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT michaelsyao aconceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT wenliyang aconceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT xuehuishi aconceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT lihongyang aconceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT dongjunli aconceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT yuemingliu aconceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT shiyiyin aconceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT chunyanlei aconceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT meixiazhang aconceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT jamescgee aconceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT xuanyang aconceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT wenbinwei aconceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT shigu aconceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT yifanwu conceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT yangliu conceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT yueyang conceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT michaelsyao conceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT wenliyang conceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT xuehuishi conceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT lihongyang conceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT dongjunli conceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT yuemingliu conceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT shiyiyin conceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT chunyanlei conceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT meixiazhang conceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT jamescgee conceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT xuanyang conceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT wenbinwei conceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata
AT shigu conceptbasedinterpretablemodelforthediagnosisofchoroidneoplasiasusingmultimodaldata