74 Novel approach for profiling immune-tumor cell interactions and mutations in the same tumor section by multiplex immunohistochemistry and NGS in immuno-oncology trials

Background Both proteins (e.g., PD-L1 IHC) and tumor mutation burden (NGS-based) are known to independently predict clinical response to anti-PD-1/PD-L1 therapies. In a meta-analysis of tumor specimens from 8135 patients treated with PD-1/PD-L1 blockers, multiplex fluorescence immunohistochemistry (mFIHC) had significantly higher diagnostic accuracy than PD-L1 IHC, tumor mutational burden (TMB), or gene expression profiling alone in predicting clinical response1 or equivalent to a multimodality approach (e.g., PD-L1IHC + TMB). While the benefits of combining mFIHC (tumor-immune interplay) and NGS approaches in selection of patients for next generation immunotherapies is appealing, tumor tissue is a key limiting factor for multimodality analyses in clinical trials. To address this critical limitation, we developed a novel approach for sequential profiling of tumor and immune cell interactions by 7-parameter mFIHC assays, followed by analyses of nucleic acid extracted from same tissue sections. Methods Formalin-fixed paraffin-embedded (FFPE) tumor tissue and cell line blocks were sectioned, and then stained using mFIHC followed by isolation of nucleic acids, or direct isolation of total nucleic acids. NanoString, qPCR, and NGS were performed on isolated nucleic acids. Nucleic acid quality, transcript abundance, and TMB scores were compared before and after mFIHC staining. Results mFIHC revealed a broad range of immune cell phenotypes and spatial interactions, including T cells, B cells, NK cells, monocytes, neutrophils, and their functional status. Isolation of testable quantities of DNA from mFIHC treated slides was achieved when using a DNA-only isolation method, and TMB scores were robust across tested conditions. Cell phenotypes identified by mFIHC were compared to TMB scores across the tested samples. Following mFIHC treatment, RNA yields were reduced relative to the non-mFIHC treated replicates, but still sufficient for optimal input into a 770-target NanoString gene expression panel. However, for mFIHC treated samples, transcript levels were not distinguishable from background for the assessed targets. Conclusions In summary, integrating mFIHC testing and TMB analysis on the same samples allows for comprehensive biomarker evaluation. The real world benefits of the combined approach will be described in upcoming clinical trials. Reference Lu, et al., Comparison of biomarker modalities for predicting response to PD-1/PD-L1 checkpoint blockade, a systematic review and meta-analysis. JAMA Oncology 2019; 5(8):1195–1204