Predicting clinical benefit of immunotherapy by antigenic or functional mutations affecting tumour immunogenicity

Abstract

Neoantigen burden is regarded as a fundamental determinant of response to immunotherapy. However, its predictive value remains in question because some tumours with high neoantigen load show resistance. Here, we investigate our patient cohort together with a public cohort by our algorithms for the modelling of peptide-MHC binding and inter-cohort genomic prediction of therapeutic resistance. We first attempt to predict MHC-binding peptides at high accuracy with convolutional neural networks. Our prediction outperforms previous methods in > 70% of test cases. We then develop a classifier that can predict resistance from functional mutations. The predictive genes are involved in immune response and EGFR signalling, whereas their mutation patterns reflect positive selection. When integrated with our neoantigen profiling, these anti-immunogenic mutations reveal higher predictive power than known resistance factors. Our results suggest that the clinical benefit of immunotherapy can be determined by neoantigens that induce immunity and functional mutations that facilitate immune evasion.

Genome-wide methylation patterns predict clinical benefit of immunotherapy in lung cancer

“…The processing of DNA methylation (850K Infinium Methylation EPIC Array), exome-and RNA-seq (n = 27) data was illustrated in our previous study [45]. Neo-antigen load was calculated using DeepNeo [46]. Aneuploidy level was retrieved from our previous research [45].…”

First-line pembrolizumab, trastuzumab, and chemotherapy in advanced HER2-positive gastric cancer with sequential genomic profiling

“…7). Besides, patients with high human leukocyte antigen-B (HLA-B)-corrected neoantigen load (> median) predicted by convolutional neural network (CNN) model 8 (calculating binding of the MHC class I molecules and somatic mutated 176 peptides from NGS results as previously described) showed better prognosis in terms of PFS 177 (22.0 vs. 7.5 months, HR 0.41, 95% CI 0.17-0.99, P=0.0393) and OS (NR vs. 17.4 months, 178 HR 0.32, 95% CI 0.11-0.92, P=0.0263) compared to those with low neoantigen load (Fig. 1st-Line anti-PD-1 and anti-HER2 with chemotherapy for HER2-positive AGC Lee et al…”

“…The median number of treatment cycles was twelve (interquartile range [IQR], [8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24], and the median number of cycles of capecitabine was eight (IQR, [6][7][8][9][10][11][12] and cisplatin was six (IQR, [4][5][6][7]. Median number of maintenance cycles (beyond 6th cycle) was six (IQR, 2-12) with median duration of 3.9 months (95% CI 2.13-11.57) (Extended Data Table 2).…”

“…TMB has been regarded as an important marker to predict immunotherapy response 32 , but not 318 in this study, possibly due to the small number of TMB-high patients (n=3). HLA is 319 becoming important in predicting immunotherapy response 8,9,[33][34][35] . In this study, we estimated neoantigen load by predicting the binding of individual patient's somatic mutation peptides to the MHC class I molecules using deep learning algorithm (CNN).…”

Integration of the Tumor Mutational Burden and Tumor Heterogeneity Identify an Immunological Subtype of Melanoma With Favorable Survival

“…Putative biomarkers, such as tumor heterogeneity, have been reported to contribute to this variation in the immune response (14). Ultraviolet beta (UVB)-induced heterogeneity was confirmed to be able to diminish the immune response in melanoma, which contributes to loss of efficacy of checkpoint blockade (15), suggesting the potential of tumor heterogeneity as a candidate marker to achieve better performance in ICI response prediction, together with the TMB (16).…”