e13642 Background: The recent exponential increase in targeted agents and immunotherapies provide new therapeutic opportunities for patients with advanced cancers who have failed first line therapy. However, lack of comprehensive precision oncology testing as part of routine pathology assessment means potential therapeutic opportunities are difficult to identify or remain undiscovered for many patients. Methods: To address this unmet clinical need we established Oncofocus, a clinically validated, ISO15189/2012 and CLIA accredited, precision oncology test optimized for analysis of small diagnostic PWET clinical biopsy samples (sample acceptance 94%). Oncofocus detects actionable genetic variants in 505 genes linked to 764 anti-cancer targeted therapy protocols, either on-market FDA and EMA approved, carrying ESMO and NCCN guideline references or currently in clinical trials. Oncofocus trending data was analyzed for a real-life cohort of 1111 patients with the aim of determining the frequency of actionable mutations in this population. This cohort represent patients with advanced stage disease who underwent Oncofocus testing having failed first line treatment protocols. Results: Analysis of trending data revealed a complex mutational landscape in which 90% of solid tumors harbored actionable mutations. The majority of patients harbored one or more actionable mutations (median 2, range 0-13) affecting key cancer related regulatory networks including the PI3K/AKT/MTOR and RAS/RAF/MEK/MAPK signaling pathways, DNA-damage repair pathways and cell cycle checkpoints. Actionable genetic variants across 33 DNA Damage and Repair (DDR) genes were identified in 30% of tumors. Using a defined predictive cut point of > 10% for tumor proportion score, PD-L1 expression levels were significantly raised in 19% of cases. Abrogation of DDR function and elevated PD-L1 levels were identified in 5% of patients, a subpopulation potentially more responsive to immunotherapy. Notably, many of the actionable mutations identified did not show linkage with histological type or site of origin. Conclusions: Our data indicates that comprehensive precision oncology testing should be strongly considered as part of the diagnostic work up for all patients with advanced cancers, independent of tumor type, thereby ensuring capture of all targeted therapy opportunities and accelerating “site agnostic” molecular basket clinical trials. Comprehensive precision oncology testing was performed successfully on routine biopsy samples circumventing the requirement for fresh tissue or large sample specimens.

The standard treatment for glioblastoma involves a combination of surgery, radiation and chemotherapy but have limited impact on survival. The exponential increase in targeted agents directed at pivotal oncogenic pathways now provide new therapeutic opportunities for this tumour type. However, lack of comprehensive precision oncology testing at diagnosis means such therapeutic opportunities are potentially overlooked. To investigate the role of semiconductor sequencing for detection of predictive biomarkers in routine glioblastoma samples we have undertaken analysis of test trending data generated by a clinically validated next generation sequencing platform designed to capture actionable genomic variants distributed across 505 genes. Analysis was performed across a cohort of 55 glioblastoma patients. Analysis of trending data has revealed a complex and rich actionable mutational landscape in which 166 actionable mutations were detected across 36 genes linked to 17 off label targeted therapy protocols and 111 clinical trials. The majority of patients harboured three or more actionable mutations affecting key cancer related regulatory networks including the PI3K/AKT/MTOR and RAS/RAF/MEK/MAPK signalling pathways, DNA-damage repair pathways and cell cycle checkpoints. Linkage with immunotherapy and PARP inhibitors was identified in 44% of glioblastoma patients as a consequence of alterations in DNA-damage repair genes. Taken together our data indicates that precision oncology testing utilising semiconductor sequencing can be used to identify a broad therapeutic armamentarium of targeted therapies and immunotherapies that can be potentially employed for the improved clinical management of glioblastoma patients.

e13521 Background: Immunohistochemical (IHC) assays are presently used as the gold standard predictive tests for immunotherapy but are compromised due to a number of potential variables. Comparative studies have demonstrated differing levels of PD-L1 staining between assays which appears independent of the antibody binding epitope. Secondly, inter-reader reliability even between expert pathologists is problematic particularly for assessment of PD-L1 positive immune cell populations. Methods: To improve predictive testing for anti PD-L1/PD1 immunotherapies we have developed and validated a Next Generation Sequencing Platform, Immunofocus, able to perform high-throughput quantitative PD-L1 gene expression levels in routine diagnostic PWET biopsies. We applied Immunofocus to a cohort of 130 NSCLCs and compared PD-L1 gene expression levels with PD-L1 IHC scores generated using the VENTANA PD-L1 (SP142) Assay. The PD-L1 IHC assessment was carried out double blinded by an independent laboratory. PD-L1 IHC scores were calculated using an algorithm combining tumour proportion score (TPS) with a PD-L1 positive immune cell (IC) score and immune cell area. Results: An exceptionally high degree of correlation was observed between the NGS PD-L1 levels with the combined PD-L1 IHC scores (P < 0.001). Therapeutic cut points for NGS PD-L1 levels were identified corresponding to PD-L1 IHC defined clinical cut points. Notably, ~20% of patients with negative PD-L1 IHC scores showed high NGS PD-L1 expression levels. We hypothesize that these cases represent false negatives and identify a cohort of patients who have shown significant response rates to anti-PD-L1/PD-directed immunotherapies. Conclusions: The Immunofocus NGS PD-L1 assay has potential to greatly improve patient selection for immunotherapy by removing the IHC assay variables and inter-reader variability which compromise current PD-L1 IHC tests while also providing standardized high throughput in the clinical setting. Immunofocus is able to integrate gene expression with somatic mutation analysis allowing capture of networks regulating the immune-checkpoint including for example adaptive and innate resistance pathways, JAK1/2 pathways, differential MHC expression, TEFF gene signature, neoantigen surrogates such as DDR defects and TMB. The integration of NGS PD-L1 expression with other putative biomarkers of response is presently ongoing to further improve prediction of response.

Oncogenic fusions represent compelling druggable targets in solid tumours highlighted by the recent site agnostic FDA approval of larotrectinib for NTRK rearrangements. However screening for fusions in routinely processed tissue samples is constrained due to degradation of nucleic acid as a result of formalin fixation., To investigate the clinical utility of semiconductor sequencing optimised for detection of actionable fusion transcripts in formalin fixed samples, we have undertaken an analysis of test trending data generated by a clinically validated next generation sequencing platform designed to capture 867 of the most clinically relevant druggable driver-partner oncogenic fusions. Here we show across a real-life cohort of 1112 patients with solid tumours that actionable fusions occur at high frequency (7.4%) with linkage to a wide range of targeted therapy protocols including seven fusion-drug matches with FDA/EMA approval and/or NCCN/ESMO recommendations and 80 clinical trials. The more prevalent actionable fusions identified were independent of tumour type in keeping with signalling via evolutionary conserved RAS/RAF/MEK/ERK, PI3K/AKT/MTOR, PLCy/PKC and JAK/STAT pathways. Taken together our data indicates that semiconductor sequencing for detection of actionable fusions can be integrated into routine diagnostic pathology workflows enabling the identification of personalised treatment options that have potential to improve clinical cancer management across many tumour types.

The standard treatment for glioblastoma involves a combination of surgery, radiation and chemotherapy but have limited impact on survival. The exponential increase in targeted agents directed at pivotal oncogenic pathways now provide new therapeutic opportunities for this tumour type. However, lack of comprehensive precision oncology testing at diagnosis means such therapeutic opportunities are potentially overlooked.To investigate the role of semiconductor sequencing for detection of predictive biomarkers in routine glioblastoma samples we have undertaken analysis of test trending data generated by a clinically validated next generation sequencing platform designed to capture 764 of the leading anti-cancer targeted agents/combinations and immunotherapies via analysis of actionable genomic variants distributed across 505 genes. Analysis was performed across a cohort of 55 glioblastoma patients.Analysis of trending data has revealed a complex and rich actionable mutational landscape in which 166 actionable mutations were detected across 36 genes linked to 17 off label targeted therapy protocols and 111 clinical trials. The majority of patients harboured three or more actionable mutations affecting key cancer related regulatory networks including the PI3K/AKT/MTOR and RAS/RAF/MEK/MAPK signalling pathways, DNA-damage repair pathways and cell cycle checkpoints. Linkage with immunotherapy and PARP inhibitors was identified in 44% of glioblastoma patients as a consequence of alterations in DNA-damage repair genes.Taken together our data indicates that precision oncology testing utilising semiconductor sequencing can be used to identify a broad therapeutic armamentarium of targeted therapies and immunotherapies that can be potentially employed for the improved clinical management of glioblastoma patients.