AI-driven multimodal algorithm predicts immunotherapy and targeted therapy outcomes in clear cell renal cell carcinoma
Danil Stupichev, Natalia Miheecheva, Ekaterina Postovalova, Yang Lyu, Akshaya Ramachandran, Ilya Galkin, Gleb Khegai, Kristina Perevoshchikova, Anna Love, Sofia Menshikova, А. В. Тарасов, Viktor Svekolkin, Maria Bruttan, Arina Varlamova, Kirill Kriukov, Ravshan Ataullakhanov, Nathan Fowler, Emily H. Cheng, Alexander Bagaev, James J. Hsieh
Abstract
Treatment for metastatic clear cell renal cell carcinoma (ccRCC) has dramatically advanced with tyrosine kinase inhibitor (TKI) and immune checkpoint inhibitor (ICI) administration. However, most patients eventually succumb to their disease, and toxicities associated with individual treatment modalities are significant. Multiple single-modality transcriptomic signatures have been developed to predict treatment response, yielding insightful yet inconsistent results when applied to independent cohorts. By unifying transcriptomic data from 14 cohorts (total n = 3,621), we present harmonized immune tumor microenvironment (HiTME) ccRCC subtypes validated with spatial proteomics. This AI-based multimodal approach integrates genomic, transcriptomic, and tumor microenvironment (TME) features for ICI and TKI therapy response prediction.