Data from Multifactorial Deep Learning Reveals Pan-Cancer Genomic Tumor Clusters with Distinct Immunogenomic Landscape and Response to Immunotherapy

• DOI: 10.1158/1078-0432.c.6528338.v1
• Publication Date: 2023-04-01T01:33:34Z
• Authors: Feng Xie; Jianjun Zhang; Jiayin Wang; Alexandre Reuben; Wei Xu; Xin Yi; Frederick S. Varn; Yongsheng Ye; Junwen Cheng; Miao Yu; Yue Wang; Yufeng Liu; Mingchao Xie; Peng Du; Ke Ma; Xin Ma; Penghui Zhou; Shengli Yang; Yaobing Chen; Guoping Wang; Xuefeng Xia; Zhongxing Liao; John V. Heymach; Ignacio I. Wistuba; P. Andrew Futreal; Kai Ye; Chao Cheng; Tian Xia
• Keywords: Microsatellite Instability; Cancer Immunotherapy

<div>AbstractPurpose:<p>Tumor genomic features have been of particular interest because their potential impact on the tumor immune microenvironment and response to immunotherapy. Due substantial heterogeneity, an integrative approach incorporating diverse molecular is needed characterize immunologic underlying primary resistance immunotherapy for establishment novel predictive biomarkers.</p>Experimental Design:<p>We developed a pan-cancer deep machine learning model integrating mutation burden, microsatellite instability, somatic copy-number alterations classify tumors different types into clusters, assessed in each cluster association with immunotherapy.</p>Results:<p>Our grouped 8,646 29 cancer from The Cancer Genome Atlas four clusters. Analysis RNA-sequencing data revealed distinct class. Furthermore, applying this two melanoma clinical cohorts demonstrated that patients classes achieved benefit Interestingly, 4 cold lack despite high instability burden.</p>Conclusions:<p>Our study provides proof principle modeling may discover intrinsic statistical cross-modality correlations multifactorial input dissect mechanisms immunotherapy, which likely involves multiple factors both host at levels.</p></div>