Predicting colorectal cancer microsatellite instability with a self-attention-enabled convolutional neural network

Xiaona Chang; Jianchao Wang; Guanjun Zhang; Ming Yang; Yanfeng Xi; Chenghang Xi; Gang Chen; Xiu Nie; Bin Meng; Xueping Quan

doi:10.1016/j.xcrm.2022.100914

Predicting colorectal cancer microsatellite instability with a self-attention-enabled convolutional neural network

Cell Rep Med. 2023 Feb 21;4(2):100914. doi: 10.1016/j.xcrm.2022.100914. Epub 2023 Jan 30.

Authors

Xiaona Chang¹, Jianchao Wang², Guanjun Zhang³, Ming Yang¹, Yanfeng Xi⁴, Chenghang Xi⁵, Gang Chen², Xiu Nie⁶, Bin Meng⁷, Xueping Quan⁸

Affiliations

¹ Department of Pathology, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430022, China.
² Department of Pathology, Clinical Oncology School of Fujian Medical University, Fujian Cancer Hospital, Fuzhou, China.
³ Department of Pathology, The First Affiliated Hospital of Xi'an Jiaotong University, Xi'an 710061, China.
⁴ Department of Pathology, Shanxi Provincial Cancer Hospital, Taiyuan 030013, China.
⁵ Tongshu Biotechnology Co. Ltd, Shanghai, China.
⁶ Department of Pathology, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430022, China. Electronic address: [email protected].
⁷ Department of Pathology, National Clinical Research Center of Cancer, Key Laboratory of Cancer Prevention and Therapy, Tianjin's Clinical Research Center for Cancer, Tianjin Medical University Cancer Institute and Hospital, Tianjin, China. Electronic address: [email protected].
⁸ Tongshu Biotechnology Co. Ltd, Shanghai, China. Electronic address: [email protected].

Abstract

This study develops a method combining a convolutional neural network model, INSIGHT, with a self-attention model, WiseMSI, to predict microsatellite instability (MSI) based on the tiles in colorectal cancer patients from a multicenter Chinese cohort. After INSIGHT differentiates tumor tiles from normal tissue tiles in a whole slide image, features of tumor tiles are extracted with a ResNet model pre-trained on ImageNet. Attention-based pooling is adopted to aggregate tile-level features into slide-level representation. INSIGHT has an area under the curve (AUC) of 0.985 for tumor patch classification. The Spearman correlation coefficient of tumor cell fraction given by expert pathologist and INSIGHT is 0.7909. WiseMSI achieves a specificity of 94.7% (95% confidence interval [CI] 93.7%-95.7%), a sensitivity of 84.7% (95% CI 82.6%-86.9%), and an AUC of 0.954 (95% CI 0.948-0.960). Comparative analysis shows that this method has better performance than the other five classic deep learning methods.

Keywords: colorectal cancer; convoluted neural network; machine learning; microsatellite instability; self-attention; tumor purity; whole slide images.

Publication types

Multicenter Study
Research Support, Non-U.S. Gov't

MeSH terms

Colorectal Neoplasms* / pathology
Humans
Microsatellite Instability*
Neural Networks, Computer