Large-vocabulary forensic pathological analyses via prototypical cross-modal contrastive learning

Abstract Forensic pathology plays a vital role in determining the cause and manner of death through macroscopic and microscopic post-mortem examinations. However, the field faces challenges such as variability in outcomes, labor-intensive processes, and a shortage of skilled professionals. This pape...

Full description

Saved in:
Bibliographic Details
Main Authors: Chen Shen, Chunfeng Lian, Wanqing Zhang, Fan Wang, Jianhua Zhang, Shuanliang Fan, Xin Wei, Gongji Wang, Kehan Li, Hongshu Mu, Hao Wu, Xinggong Liang, Jianhua Ma, Zhenyuan Wang
Format: Article
Language:English
Published: Nature Portfolio 2025-07-01
Series:Nature Communications
Online Access:https://doi.org/10.1038/s41467-025-62060-x
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849764246588489728
author Chen Shen
Chunfeng Lian
Wanqing Zhang
Fan Wang
Jianhua Zhang
Shuanliang Fan
Xin Wei
Gongji Wang
Kehan Li
Hongshu Mu
Hao Wu
Xinggong Liang
Jianhua Ma
Zhenyuan Wang
author_facet Chen Shen
Chunfeng Lian
Wanqing Zhang
Fan Wang
Jianhua Zhang
Shuanliang Fan
Xin Wei
Gongji Wang
Kehan Li
Hongshu Mu
Hao Wu
Xinggong Liang
Jianhua Ma
Zhenyuan Wang
author_sort Chen Shen
collection DOAJ
description Abstract Forensic pathology plays a vital role in determining the cause and manner of death through macroscopic and microscopic post-mortem examinations. However, the field faces challenges such as variability in outcomes, labor-intensive processes, and a shortage of skilled professionals. This paper introduces SongCi, a visual-language model tailored for forensic pathology. Leveraging advanced prototypical cross-modal self-supervised contrastive learning, SongCi improves the accuracy, efficiency, and generalizability of forensic analyses. Pre-trained and validated on a large multi-center dataset comprising over 16 million high-resolution image patches, 2, 228 vision-language pairs from post-mortem whole slide images, gross key findings, and 471 unique diagnostic outcomes, SongCi demonstrates superior performance over existing multi-modal models and computational pathology foundation models in forensic tasks. It matches experienced forensic pathologists’ capabilities, significantly outperforms less experienced practitioners, and offers robust multi-modal explainability.
format Article
id doaj-art-fea06ba2e7a646ea90e22c8eb883ca59
institution DOAJ
issn 2041-1723
language English
publishDate 2025-07-01
publisher Nature Portfolio
record_format Article
series Nature Communications
spelling doaj-art-fea06ba2e7a646ea90e22c8eb883ca592025-08-20T03:05:10ZengNature PortfolioNature Communications2041-17232025-07-0116112010.1038/s41467-025-62060-xLarge-vocabulary forensic pathological analyses via prototypical cross-modal contrastive learningChen Shen0Chunfeng Lian1Wanqing Zhang2Fan Wang3Jianhua Zhang4Shuanliang Fan5Xin Wei6Gongji Wang7Kehan Li8Hongshu Mu9Hao Wu10Xinggong Liang11Jianhua Ma12Zhenyuan Wang13Key Laboratory of National Ministry of Health for Forensic Sciences, School of Medicine & Forensics, Health Science Center, Xi’an Jiaotong UniversitySchool of Mathematics and Statistics, Xi’an Jiaotong UniversityKey Laboratory of National Ministry of Health for Forensic Sciences, School of Medicine & Forensics, Health Science Center, Xi’an Jiaotong UniversityKey Laboratory of Biomedical Information Engineering of Ministry of Education, School of Life Science and Technology, Xi’an Jiaotong UniversityShanghai Key Laboratory of Forensic Medicine, Shanghai Forensic Service Platform, Academy of Forensic ScienceKey Laboratory of National Ministry of Health for Forensic Sciences, School of Medicine & Forensics, Health Science Center, Xi’an Jiaotong UniversityKey Laboratory of National Ministry of Health for Forensic Sciences, School of Medicine & Forensics, Health Science Center, Xi’an Jiaotong UniversityKey Laboratory of National Ministry of Health for Forensic Sciences, School of Medicine & Forensics, Health Science Center, Xi’an Jiaotong UniversityKey Laboratory of Biomedical Information Engineering of Ministry of Education, School of Life Science and Technology, Xi’an Jiaotong UniversityWeicheng Branch, Xian’yang Public Security BureauKey Laboratory of National Ministry of Health for Forensic Sciences, School of Medicine & Forensics, Health Science Center, Xi’an Jiaotong UniversityKey Laboratory of National Ministry of Health for Forensic Sciences, School of Medicine & Forensics, Health Science Center, Xi’an Jiaotong UniversityPazhou Lab (Huangpu)Key Laboratory of National Ministry of Health for Forensic Sciences, School of Medicine & Forensics, Health Science Center, Xi’an Jiaotong UniversityAbstract Forensic pathology plays a vital role in determining the cause and manner of death through macroscopic and microscopic post-mortem examinations. However, the field faces challenges such as variability in outcomes, labor-intensive processes, and a shortage of skilled professionals. This paper introduces SongCi, a visual-language model tailored for forensic pathology. Leveraging advanced prototypical cross-modal self-supervised contrastive learning, SongCi improves the accuracy, efficiency, and generalizability of forensic analyses. Pre-trained and validated on a large multi-center dataset comprising over 16 million high-resolution image patches, 2, 228 vision-language pairs from post-mortem whole slide images, gross key findings, and 471 unique diagnostic outcomes, SongCi demonstrates superior performance over existing multi-modal models and computational pathology foundation models in forensic tasks. It matches experienced forensic pathologists’ capabilities, significantly outperforms less experienced practitioners, and offers robust multi-modal explainability.https://doi.org/10.1038/s41467-025-62060-x
spellingShingle Chen Shen
Chunfeng Lian
Wanqing Zhang
Fan Wang
Jianhua Zhang
Shuanliang Fan
Xin Wei
Gongji Wang
Kehan Li
Hongshu Mu
Hao Wu
Xinggong Liang
Jianhua Ma
Zhenyuan Wang
Large-vocabulary forensic pathological analyses via prototypical cross-modal contrastive learning
Nature Communications
title Large-vocabulary forensic pathological analyses via prototypical cross-modal contrastive learning
title_full Large-vocabulary forensic pathological analyses via prototypical cross-modal contrastive learning
title_fullStr Large-vocabulary forensic pathological analyses via prototypical cross-modal contrastive learning
title_full_unstemmed Large-vocabulary forensic pathological analyses via prototypical cross-modal contrastive learning
title_short Large-vocabulary forensic pathological analyses via prototypical cross-modal contrastive learning
title_sort large vocabulary forensic pathological analyses via prototypical cross modal contrastive learning
url https://doi.org/10.1038/s41467-025-62060-x
work_keys_str_mv AT chenshen largevocabularyforensicpathologicalanalysesviaprototypicalcrossmodalcontrastivelearning
AT chunfenglian largevocabularyforensicpathologicalanalysesviaprototypicalcrossmodalcontrastivelearning
AT wanqingzhang largevocabularyforensicpathologicalanalysesviaprototypicalcrossmodalcontrastivelearning
AT fanwang largevocabularyforensicpathologicalanalysesviaprototypicalcrossmodalcontrastivelearning
AT jianhuazhang largevocabularyforensicpathologicalanalysesviaprototypicalcrossmodalcontrastivelearning
AT shuanliangfan largevocabularyforensicpathologicalanalysesviaprototypicalcrossmodalcontrastivelearning
AT xinwei largevocabularyforensicpathologicalanalysesviaprototypicalcrossmodalcontrastivelearning
AT gongjiwang largevocabularyforensicpathologicalanalysesviaprototypicalcrossmodalcontrastivelearning
AT kehanli largevocabularyforensicpathologicalanalysesviaprototypicalcrossmodalcontrastivelearning
AT hongshumu largevocabularyforensicpathologicalanalysesviaprototypicalcrossmodalcontrastivelearning
AT haowu largevocabularyforensicpathologicalanalysesviaprototypicalcrossmodalcontrastivelearning
AT xinggongliang largevocabularyforensicpathologicalanalysesviaprototypicalcrossmodalcontrastivelearning
AT jianhuama largevocabularyforensicpathologicalanalysesviaprototypicalcrossmodalcontrastivelearning
AT zhenyuanwang largevocabularyforensicpathologicalanalysesviaprototypicalcrossmodalcontrastivelearning