Multi-diagnosis multi-instance learning for auxiliary gene mutation diagnosis in whole slide images
Pathology - Research and Practice · 2026-10 · 286 · 156576
DOI: 10.1016/j.prp.2026.156576 ↗Kurz gefasst
Fachartikel in der Zeitschrift Pathology - Research and Practice, 2026-10. Verfasst von Ao Liu und 8 weiteren Autor:innen. Beteiligte Einrichtung: School of Software Engineering, Xi'an Jiaotong University, Xi'an, China. Hauptthemen laut MEDLINE: Mutation, Image Interpretation, Computer-Assisted, Neoplasms, Lung Neoplasms. Die Arbeit nennt 49 Literaturangaben. Gefördert von: Fujian Provincial Natural Science Foundation.
Abstract (Originalsprache)
BACKGROUND AND OBJECTIVE: Multi-Instance Learning (MIL) is widely applied in the representation of high-resolution whole slide images (WSIs) in pathology. However, there are still challenges in associating WSIs with genetic representations. The connection between the types of genetic mutations in WSI and their histological images is not intuitive, making it difficult for traditional models to effectively capture the key feature representations of genetic mutations. METHODS: This study proposes a novel cancer-gene multi-diagnosis multi-instance learning approach for WSI-level genetic mutation diagnosis. This method integrates cancer diagnostic data with pixel-level annotations and genetic mutation diagnostic data with WSI-level annotations as multi-diagnosis data inputs, systematically learning the representations of genetic mutations. By leveraging the high correlation between cancerous regions and genetic mutation areas, we filter out representative instances containing lesions and delve into the morphological and histological feature representations of genetic mutations through a focused sampling strategy. RESULTS: Our proposed framework demonstrates advanced performance on gene mutation datasets of lung cancer, bladder cancer and breast cancer. Moreover, this method generates a series of visualizations that serve as interpretability tools reflecting model attention. CONCLUSION: Our framework achieves deep coupling of cross-diagnostic pathological representations by deconstructing latent correlation maps between cancer regions and gene mutation regions in WSIs. Our method can, to some extent, provide decision support for pathologists' clinical diagnosis, and is potentially useful for triage or prioritization.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- School of Software Engineering, Xi'an Jiaotong University, Xi'an, China.
- School of Cyber Science and Engineering, MoE KLINNS Lab, Xi'an Jiaotong University, Xi'an, China.
- School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China.
- Department of Pathology, Shengli Clinical Medical College, Fujian Medical University, Fuzhou, Fujian, China.
- Department of Pathology, Quanzhou First Hospital affiliated to Fujian Medical University, Quanzhou, China.
- Department of Breast Disease Diagnosis and Treatment Center, The Second Affiliated Hospital of Xi'an Jiaotong University, Xi'an, China. Electronic address: xuebaoc@mail.xjtu.edu.cn.
Förderung
- Fujian Provincial Natural Science Foundation
Zitieren
APA
Liu, A., Liu, Y., Li, W., Liu, C., Qin, X., Li, Z., Chen, Y., Jin, L., & Liang, B. (2026). Multi-diagnosis multi-instance learning for auxiliary gene mutation diagnosis in whole slide images. Pathology - Research and Practice, 286, 156576. https://doi.org/10.1016/j.prp.2026.156576
Vancouver
Liu A, Liu Y, Li W, Liu C, Qin X, Li Z, et al. Multi-diagnosis multi-instance learning for auxiliary gene mutation diagnosis in whole slide images. Pathology - Research and Practice. 2026;286:156576. doi:10.1016/j.prp.2026.156576
BibTeX
@article{Liu2026,
author = {Liu, Ao and Liu, Yang and Li, Wentao and Liu, Chen and Qin, Xiaoying and Li, Zhongyu and Chen, Yifeng and Jin, Long and Liang, Baobao},
title = {Multi-diagnosis multi-instance learning for auxiliary gene mutation diagnosis in whole slide images},
journal = {Pathology - Research and Practice},
year = {2026},
volume = {286},
pages = {156576},
doi = {10.1016/j.prp.2026.156576}
}
EUVIMED