Toward the standardization of bioinformatics methods for the accurate assessment of tumor mutational burden (TMB)

Abstract

research subjects with melanoma and non-small cell lung carcinoma (NSCLC) and peripheral blood leukocytes (PBL) from healthy donors. Methods: Total RNA from 63 melanoma and 19 NSCLC tumor biopsy research samples (non-FFPE) was extracted for use in long-amplicon TCRB chain sequencing (mean amplicon of 330bp covering CDR1, 2 and 3) via the Oncomine TCR Beta-LR Research Assay. To evaluate T cell convergence, we searched for instances where TCRB chains were identical in amino acid space but had distinct nucleotide sequences owing to N-addition and exonucleotide chewback within the V-D and D-J junctions of the CDR3. To provide context, we evaluated evidence for T cell convergence in PBL T cell repertoires derived from 16 healthy donors. Results: Sequencing of melanoma biopsy research samples typically yielded within the range of 2000 to 8000 clones per sample. Convergent T cell receptors were identified in the great majority of melanoma and NSCLC tumor infiltrating T cell repertoires having greater than 100 detected clones (92% and 100%, respectively). The frequency of con-vergent T cell rearrangements was significantly greater in melanoma and NSCLC tumor biopsies than T cell repertoires derived from healthy PBL research samples. Conclusions: These data suggest that T cell convergence may be a common feature of the melanoma and NSCLC infiltrating T cell repertoire. Convergence was more frequently observed within the TME than T cell repertoires derived from healthy PBL, consistent with elevated antigen-driven T cell selection within the TME. The finding of elevated T cell convergence in melanoma and NSCLC suggests that convergence may be a hallmark of immunogenic tumors. For research use only. Legal entity responsible for the study: Thermo Fisher Scientific. Background: Copanlisib, a pan-class I phosphatidylinositol 3-kinase (PI3K) inhibitor with predominant activity against a and d isoforms, shows immune stimulating effects and anti-tumor activity in combination with immune checkpoint blockers in tumor models. Here we explore tumor PI3K signaling and the effects of copanlisib on immune and tumor microenvironment modulation in subjects with NHL and solid tumors (ST) in a phase I pharmacodynamic study (NCT02155582). Methods: Patients received 0.4 or 0.8 mg/kg (equivalent to a flat dose of 60 mg) copan-lisib on an intermittent schedule (QW, IV, 3 wks on/1 wk off). Tumor biopsies were performed at baseline and Day 15. Tumor PI3K isoforms, PTEN, and CD3, CD4 and CD8 tumor infiltrating lymphocytes (TILs) were assessed by IHC. Plasma protein markers were measured using multiplexed immunoassay. Tumor and lymphoma responses were based on RECIST 1.1 and modified Cheson 2007 criteria, respectively. Results: A total of 61 patients were treated: 33 NHL (20 at 0.4 mg/kg and 13 at 0.8 mg/ kg) and 28 ST (14 at 0.4 mg/kg and 14 at 0.8 mg/kg). Among patients treated at 0.8 mg/ kg, 2 had CR (PTCL and DLBCL) and 5 had PR (MCL, FL, 2 DLBCL, and endometrial adenocarcinoma); 1 additional PR (DLBCL) occurred at 0.4 mg/kg. Tumor PI3K a was detected in most NHL (18/19) and ST (22/25) samples with comparable intensity. PI3K d was predominantly present in NHL (18/19). PI3K ß and c were present in a subset of NHL and ST. PTEN loss was more frequent in ST than NHL. CD3, CD4 and CD8 TIL numbers were higher in NHL than in ST. At 0.8 mg/kg copanlisib decreased tumor pAKT and pS6 in both NHL and ST, and reduced CD4 TILs in NHL (mean-81%, n ¼ 7), with little effect on CD8 TILs in both NHL and ST. In plasma, copanlisib decreased cytokine/chemokines (e.g. CCL2, CCL5, CCL17), and factors associated with macrophages (e.g. CD163, CCL4, CCL22) and Treg cells (IL-2Ra). High baseline levels of CD27 and IL2Ra were associated with tumor reduction in NHL (p < 0.05, unadjusted). Conclusions: The high prevalence of the PI3K isoforms-especially a in both NHL and ST and d in NHL-is consistent with a role for PI3K signaling in immune suppression. The immune modulation profile for copanlisib supports combination studies with immunotherapy. Clinical trial identification: NCT02155582. Legal entity responsible for the study: Bayer AG. Background: Squamous cell carcinomas (SCC) have diverse anatomic etiologies but may share common genomic biomarkers. We profiled 7,871 unique SCCs across nine anatomic sites to investigate commonality in genomic alterations (GA), tumor mutational burden (TMB), human papillomavirus (HPV) association, and mutational signatures. Methods: Tissue from over 8,100 unique SCC samples originating from nine anatomic sites (anogenital (anus, cervix, penis, vagina, vulva), esophagus, head and neck, lung, and skin) were sequenced by hybrid capture-based comprehensive genomic profiling to evaluate GA and TMB. About 3% of non-cutaneous SCC samples had UV signatures, indicative of potential primary site misdiagnoses, and were filtered from the analysis. Detection of HPV, including high-risk strains 16, 18, 31, 33, and 45, was implemented through de novo assembly of non-human sequencing reads and BLASTn comparison against all viral nucleotide sequences in the NCBI database. Results: The proportion of HPVþ patients by anatomic site varied, with the highest being anal (91%) and cervical (83%). The mutational landscape of each cohort was similar, regardless of anatomic origin, but clustered based on HPV status. The largest differences in GA frequency as stratified by HPV-vs. HPVþ were TP53 (87% vs. 12%), CDKN2A (45% vs. 6%), and PIK3CA (22% vs. 33%). The median TMB in cases originating from HPV-associated sites was similar, regardless of HPV status. Higher median TMB was observed in lung and skin cases, which exhibited significant enrichment of mutational signatures indicative of tobacco-and UV-induced DNA damage, respectively. Conclusions: HPVþ and HPV-SCC populations have distinct genomic profiles and, for the latter, anatomic site is correlated with TMB distribution, secondary to associated carcinogen exposure. As such, biomarkers such as TMB and UV signature can provide unexpected insight into site of origin misdiagnoses and may correlate with benefit from immune checkpoint inhibitors. Table: 68P Tumor Site % HPVþ Median TMB (Interquartile Range) % TMB > ¼ 10 % TMB > ¼ 20 Anogenital (n ¼ 1213) 76 5 (6) 17 5 Head and Neck (n ¼ 1843) 36 4 (5) 15 5 Esophageal (n ¼ 416) 6 5 (4) 13 2 Lung (n ¼ 3977) 5 9 (8) 43 9 Skin (n ¼ 422) 8 40 (69) 68 62 Legal entity responsible for the study: Foundation Medicine, Inc. Background: TMB has emerged as a predictive biomarker of response to immune checkpoint inhibitors. CheckMate 227 demonstrated that patients with non-small cell lung cancer (NSCLC) with TMB !10 mutations/megabase derived enhanced benefit from first-line treatment with nivolumab þ ipilimumab vs chemotherapy (Hellmann et al. NEJM 2018). Standardized approaches for the measurement and reporting of TMB are essential for the real-world implementation of TMB. This study aimed to refine a bioinformatic pipeline for mutation calling and annotation of whole exome sequencing (WES) data for TMB assessment.