Long non-coding RNAs (lncRNAs) are involved in innate and adaptive immunities in cancers by mediating the functional states of immunologic cells, pathways, and genes. However, whether lncRNAs could serve as effective biomarkers for molecular classification and prediction of cancer immunotherapy efficacy are largely unknown.
This study analyzed lncRNA and genomic data of 348 atezolizumab-treated bladder cancer patients from phase II IMvigor 210 trial and 3,021 patients in lung cancer, breast cancer, bladder cancer, and melanoma cohorts from The Cancer Genome Atlas. We investigated lncRNA-based immune subtypes associated with cancer immunotherapy efficacy. We built a novel lncRNA score using computational algorithms and integrated it with programmed-death ligand 1 (PD-L1) expression and tumor mutation burden (TMB) to achieve accurate immunotherapeutic prediction.
The results from IMvigor 210 trial showed that four distinct microenvironment-based subtypes characterized by lncRNA expression and tumor specific cytotoxic T lymphocytes (CTLs) infiltration had significant difference in overall survival (OS) (hazard ratio [HR] 0.77, 95%CI 0.67–0.88; P = 0.0002), with the greatest benefits in Immune-Active Class, followed by Immune-Exclusion Class, Immune-Dysfunctional Class, and Immune-Desert Class. High NKILA lncRNA expression was identified as a negative predictor of immunotherapeutic OS benefits and a critical regulator of dysfunctional immune response. Patients with low- versus high- lncRNA scores were associated with significantly longer OS in IMvigor 210 trial (HR 0.32, 95% CI 0.24–0.42; P < 0.0001) and across various cancer types. Multiomics comprising lncRNA score, PD-L1 expression and TMB led to more precise OS prediction in IMvigor 210 trial (20-month AUC=0.80) and in melanoma immunotherapy cohort (24-month AUC=0.89) compared with variable alone.
We suggested immunotherapy for patients in Immune-Functional Class, especially for those in high CTL Immune-Active Class. Multiomics comprising lncRNA score, PD-L1 expression and TMB could accurately predict immunotherapy efficacy.
Herui Yao.
Grants from the National Natural Science Foundation of China (81372819, 81572596, U1601223), the Natural Science Foundation of Guangdong Province (2017A030313828), the Guangzhou Science and Technology Program (201704020131), the Sun Yat-Sen University Clinical Research 5010 Program (2018007).
All authors have declared no conflicts of interest.