Deep learning using histological images for gene mutation prediction in lung cancer: a multicentre retrospective study

Yu Zhao ¹, Shan Xiong ², Qin Ren ³

¹Department of Thoracic Oncology and Surgery, The First Affiliated Hospital of Guangzhou Medical University, State Key Laboratory of Respiratory Disease & National Clinical Research Center for Respiratory Disease, Guangzhou, China; AI Lab, Tencent, Shenzhen, China.
²Department of Thoracic Oncology and Surgery, The First Affiliated Hospital of Guangzhou Medical University, State Key Laboratory of Respiratory Disease & National Clinical Research Center for Respiratory Disease, Guangzhou, China; Department of Thoracic Oncology and Surgery, Hengqin Hospital, The First Affiliated Hospital of Guangzhou Medical University, Hengqin, China.
³AI Lab, Tencent, Shenzhen, China.
⁴Department of Respiratory Medicine, National Key Clinical Specialty, Branch of National Clinical Research Center for Respiratory Disease, Xiangya Hospital, Central South University, Changsha, China.
⁵Department of Thoracic Surgery, Shenzhen People's Hospital, 2nd Clinical Medical College of Jinan University, Shenzhen, China.
⁶Department of Respiratory Medicine, Shenzhen People's Hospital, 2nd Clinical Medical College of Jinan University, Shenzhen, China.
⁷Division of Pulmonary and Critical Care Medicine, The First Affiliated Hospital of Sun Yat-sen University, Guangzhou, China.
⁸Department of Thoracic Surgery, Shengli Clinical Medical College of Fujian Medical University, Fujian Provincial Hospital, Fuzhou, China.
⁹Department of Thoracic Surgery, Hainan General Hospital, Haikou, China.
¹⁰Thoracic Surgery Department 2, Hunan Cancer Hospital and The Affiliated Cancer Hospital of Xiangya School of Medicine, Central South University, Changsha, China.
¹¹Department of Oncology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital & Shenzhen Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Shenzhen, China.
¹²Department of Thoracic Oncology and Surgery, The First Affiliated Hospital of Guangzhou Medical University, State Key Laboratory of Respiratory Disease & National Clinical Research Center for Respiratory Disease, Guangzhou, China.
¹³Department of Thoracic Oncology, Clinical Oncology School of Fujian Medical University, Fujian Cancer Hospital, Fujian Key Laboratory of Advanced Technology for Cancer Screening and Early Diagnosis, Fuzhou, China.
¹⁴Department of Oncology, University of Electronic Science and Technology of China, Sichuan Cancer Hospital and Institute & Cancer, The Second People's Hospital of Sichuan Province, Chengdu, China.
¹⁵Department of Oncology, Guangdong Sanjiu Brain Hospital, Guangzhou, China.
¹⁶Department of Thoracic Surgery, Chengdu Third People's Hospital, Affiliated Hospital of Southwest Jiaotong University, Chengdu, China.
¹⁷Department of Thoracic Surgery, Peking University Shenzhen Hospital, Shenzhen Peking University-The Hong Kong University of Science and Technology Medical Center, Shenzhen, China.
¹⁸Department of Medical Oncology, The First Affiliated Hospital of Xiamen University, Xiamen, China.
¹⁹Department of Chest Surgery, Affiliated Cancer Hospital & Institute of Guangzhou Medical University, Guangzhou, China.
²⁰Department of Oncology, The Second Xiangya Hospital of Central South University, Changsha, China.
²¹Department of Pathology, The First Affiliated Hospital of Guangzhou Medical University, Guangzhou, China.
²²Burning Rock Biotech, Guangzhou, China.
²³Biomedicine Discovery Institute and Monash Data Futures Institute, Monash University, Melbourne, VIC, Australia.
²⁴Department of Thoracic Medical Oncology, Hunan Cancer Hospital and The Affiliated Cancer Hospital of Xiangya School of Medicine, Central South University, Changsha, China.

⁰Department of Thoracic Oncology and Surgery, The First Affiliated Hospital of Guangzhou Medical University, State Key Laboratory of Respiratory Disease & National Clinical Research Center for Respiratory Disease, Guangzhou, China; AI Lab, Tencent, Shenzhen, China.

The Lancet. Oncology +

|

December 9, 2024

View abstract on PubMed

Summary

An AI tool called DeepGEM can predict lung cancer gene mutations from standard histology slides, offering a fast and affordable alternative to traditional genomic testing for better treatment planning.

Area Of Science

Oncology
Computational Biology
Medical Imaging

Background

Accurate driver gene mutation detection is vital for lung cancer prognosis and treatment.
Conventional genomic testing is limited by tissue requirements and resource intensity.
An AI-driven approach is needed for accessible mutation prediction.

Purpose Of The Study

To develop an annotation-free AI method (DeepGEM) for predicting gene mutations from histological slides.
To evaluate DeepGEM's performance on internal, external, and public datasets.
To assess the model's generalizability to lymph node metastasis biopsies.

Main Methods

A multicentre retrospective study involving 3637 lung cancer patients with paired pathology images and mutation data.
Development of a co-supervised multiple instance learning model with label disambiguation.
Validation on internal, external (15 hospitals), and The Cancer Genome Atlas (TCGA) datasets.

Main Results

DeepGEM demonstrated robust performance across all datasets, with AUC values ranging from 0.76 to 0.97.
The model accurately predicted mutations in both excisional and aspiration biopsy samples.
Generalization to lymph node metastases was confirmed, with high AUC for EGFR and KRAS mutations.

Conclusions

An AI-based method (DeepGEM) provides accurate, timely, and economical prediction of gene mutations and their spatial distribution.
DeepGEM shows significant potential as an assistive tool for guiding lung cancer clinical treatment.
The AI method overcomes limitations of conventional genomic testing, especially in resource-limited settings.