Artificial intelligence-quantified tumour-lymphocyte spatial interaction predicts disease-free survival in resected lung adenocarcinoma: A graph-based, multicentre study

Zhengyun Feng, Huan Lin, Zaiyi Liu, Lixu Yan, Yumeng Wang, Bingbing Li, Entao Liu, Chu Han, Zhenwei Shi, Cheng Lu, Zhenbing Liu, Cheng Pang, Zhenhui Li, Yanfen Cui, Xipeng Pan, Xin Chen

Abstract: Highlights•Proposed a novel biomarker based on cell graph to predict lung adenocarcinoma disease-free survival (DFS).•By integrating the novel biomarker with clinicopathologic risk factors, the integrated model (full model) improves the prediction of DFS in multiple independent cohorts.•Proposed a correction strategy incorporating tissue information into the existing model for cell nuclei segmentation and classification, which could further improve the performance of nuclei classification.•The workflow is low-cost, and applied for other cancer-related diseases easily.•This study included a complete analysis process from methodology to application.