BackgroundThe majority of pancreatic ductal adenocarcinomas (PDACs) are driven by mutant (mt) KRAS. How mt KRAS and co-driver mutations affect the immune cell (IC) landscape of PDAC remains uncertain. Herein, we characterize the types of IC in the PDAC tumor microenvironment (TME) and the prevalence of immuno-oncologic (IO) biomarkers by genomic and transcriptomic analysis in the context of KRAS status.Materials and methods4142 PDAC and 3727 colorectal cancer (CRC) cases with KRAS mt were analyzed using next-generation DNA sequencing, immunohistochemistry, and whole-transcriptome RNA sequencing. Microsatellite instability and deficiency in mismatch repair (MSI-H/dMMR) and tumor mutational burden (TMB) were also assessed.ResultsWe found KRAS mt in 81% of PDAC, with the most common variant being G12D in PDAC, and fewer cases of KRAS mt were co-expressed with the predictive IO marker MSI-H/dMMR than KRAS-wild-type (wt). However, KRASG12D, KRASG12V, and KRASQ61 mutations had significantly lower TMB than KRAS wt tumors in PDAC. The IC environment of KRAS mt PDAC showed significant differences in nearly all IC types; a similar pattern was observed in CRC but was less pronounced.ConclusionsTherapeutic IO targets like programmed death-ligand 1 are enriched in pancreatic adenocarcinoma cases harboring specific targetable variants of KRAS mt PDAC. Better understanding of the TME could lead to tailored immunotherapeutic strategies to overcome these barriers in KRAS mt PDAC, possibly in combination with molecularly targeted treatment strategies.

With the emergence of novel KRAS-targeted drugs, molecular analysis of unique KRAS mutations in non-small cell lung cancer (NSCLC) has become increasingly relevant. Acquired KRAS mutations are a known resistance mechanism in driver mutation-positive (DM+) NSCLC. The impact of acquired vs. de novo KRAS mutations on the tumor-immune microenvironment (TIME) is unknown. Mutation status was assessed using next-generation sequencing (Caris Life Sciences) with paired whole-exome and whole-transcriptome sequencing (Illumina NovaSeq). KRAS-mutated (KRASmt) subgroups were defined as de novo KRASmt NSCLC (KRAS only identified driver – DN) and DM+ NSCLC with acquired KRAS mutations (concurrent KRASmt with other known drivers – ACQ). Immune pathway enrichment was assessed with single sample gene set enrichment analyses (SSGSEA). Hierarchical agglomerative clustering (HAC) was performed on expression of a panel of immune checkpoints to define immune low/medium/high clusters. Fisher’s exact, Chi-square, and Mann-Whitney U tests were used, as appropriate, to assess statistical significance and corrected for multiple hypothesis testing (q<0.05/FDR <0.25). 6240 KRASmt NSCLC cases were identified - 6186 DN, 54 ACQ. A trend (p<0.05,q>0.05) toward increased expression of immune checkpoints PD-L1, TIM-3, and CD86, as well as a significant enrichment of TNFα, IFNα, IFNλ, JAK-STAT, and inflammatory response signaling pathways were observed in ACQ compared to DN. HAC analysis to delineate the immune profile associated with specific KRAS mutations revealed that the distribution of KRAS mutations was significantly different among immune clusters in the DN but not the ACQ subgroup (see table).Table: 1399PKRAS subgroupDNACQImmune checkpoint expression clusterLow (n=2708)Medium (n=2738)High (n=760)Low (n=17)Medium (n=31)High (n=6)G12C (%)38.641.641.841.225.850.0G12V (%)19.418.720.811.822.60.0G12D (%)15.714.110.317.719.416.7G12A (%)6.56.37.911.86.50.0G13C (%)3.83.34.35.93.216.7Q61H (%)4.54.34.95.99.716.7Other (%)11.411.710.05.912.90.0 Open table in a new tab . Differences in the TIME were observed between DN and ACQ KRASmt NSCLC. KRAS mutation subtype frequency appeared to differ according to TIME phenotype in DN KRASmt NSCLC. The clinical and therapeutic implications of these findings warrant further investigation.

Uveal melanoma (UM) is a rare form of melanoma having poor responses to currently available systemic agents. Our goal is to understand differences in chemokine expression in relation to tumor genetics and response to immunotherapy. UM patient tumors (N=278, 41 primary, 174 liver mets and 63 other mets) were profiled by NGS DNA/RNA at Caris Life Sciences (Phoenix, AZ). Chemokine expression, high/Low (H/L) defined as samples with >75th- or <25th-percentile of transcripts per million (TPM). Tumor microenvironment (TME) cell fraction estimated by RNA, with median fold changes (H/L) or proportion of samples with non-zero fraction reported (Table). PDL1+ (SP142) tested by IHC. Survival data obtained from insurance claims. Primary tumors had increased expression of CXCR4, CXCR1, CXCR2, CCL27 and CXCL13 compared to liver mets (FC range 1.2-4.2), while CXCL2 expression was increased in liver mets (2.3, p<0.01). In liver mets, increased infiltration of immunosuppressive cells was observed for CXCR4-H and CXCL12-H tumors(Table). Only M1 macrophages, CD8+ T cells and B cells were increased for CXCR4-H and CXCL12-H primary tumors. PDL1+ rates were increased in CXCR4-H tumors overall (H 36% vs L 13%, p<0.05). In liver mets, SF3B1 mutation was associated with lower CXCL1 and CXCL2 expression compared to WT (0.35- and 0.47-fold, respectively, p-<0.01). BAP1-mutated liver mets showed increased CXCL1 expression (2.0-fold, p-0.04), whereas CCR10 expression was increased in BAP1-mutated primary tumors (2.6-fold, p-0.02). Among immunotherapy treated patients with liver mets, there was a trend for improved survival for CXCL12 H (n=13) vs L (n=12) (HR 0.51 (0.21-1.3), p=0.14) and CXCR4 H (n=13) vs L (n=12) (HR 0.49 (0.20-1.2), p=0.12) though not significant.Table: 1132PTME (liver mets): CXCR4/CXCL12 median fold change (H/L) (where median is 0, % non-zero H vs L).Immune CellCXCR4 Median FC (H/L) or non-zero%p-valueCXCL12 Median FC (H/L) or non-zero%p-valueMonocyte13% vs 0%0.0137% vs 11%0.31Treg2.9<0.012.9<0.01T cell CD870% vs 28%<0.0167% vs 34%<0.01T cell CD432% vs 13%0.0237% vs 11%0.01NK cell1.4<0.0011.5<0.001Macrophage M21.30.0021.3<0.001Macrophage M192% vs 50%<0.0196% vs 49%<0.001B cell1.6<0.011.30.02 Open table in a new tab Our results suggest chemokines are differentially expressed in tumors harboring the common alterations associated with medium risk of distant metastases.

Patients with recurrent head and neck squamous cell carcinoma (HNSCC) have dismal outcomes despite advances in salvage definitive therapies. Aiming to expand our understanding of recurrent HNSCC, we characterize the molecular and immune landscape of HNSCC patients enrolled in a Phase II trial of adjuvant nivolumab after salvage resection (NCT03355560). Tissue-based DNA and RNA sequencing utilizing a commercially available CLIA-certified assay (Caris Life Sciences) were used to analyze tissue samples from HNSCC patients enrolled in the trial. NGS, WES and WTS was performed on 37 tumors. TMB cutoff point of ≥10 was used. PD-L1 expression was tested by IHC (22c3 antibody, ≥1 CPS being positive). Statistical significance was determined using Fisher's-Exact or X2 test with Benjamini-Hochberg correction. Gene expression was analyzed using the Limma R package. Immune cell abundance was measured using Microenvironment Cell Populations-counter (MCP) method. In patients with recurrent H&N cancers treated with Nivolumab, distinct immune cell abundance and molecular alterations were not found between patients who developed a relapsed disease (n= 11) and those who did not (n= 26). Molecular alterations in TP53 (86%, 100%, 0%; q < 0.01), P1K3CA (0%, 0%, 57%; q< 0.01) and p16 IHC (8%, 0%, 86%; q < 0.01) were detected in tumors from the larynx, oral cavity, and oropharynx. Myeloid dendritic cell (FC= 1.3; p < 0.05) and endothelial cell (FC = 1.2; p < 0.05) abundance were increased in patients who received RT only vs ChemRT. ABCA10 gene expression decreased in patients treated previously with ChemRT (logFC = -1.4; p < 0.05), compared to RT only. TMB-High tumors exhibited increased SNRPFP1 gene expression (logFC = 1.2; p < 0.01). Well differentiated tumors showed increased gene expression in SEMA3G (logFC = 1.4; p < 0.05) and HS3ST6 (logFC = 1.4; p < 0.05) genes compared to poorly/moderately differentiated tumors. Recurrent HNSCC cancers manifest heterogeneous molecular and immune profiles on subgroups analysis with no clear marker to predict response to Nivolumab in the adjuvant setting. Larger prospective trials are needed to explore this area further.

RAS pathway alterations in NSCLC have been linked to worse prognosis and remain a challenging therapeutic target. HRAS, a RAS family members, is under-investigated and its activation depends on farnesylation, making it an appealing target for farnesyltransferase inhibitors. Although KRAS has been associated with immune modulation in NSCLC, the role of HRAS remains unclear. We aimed to investigate the relationship of HRAS gene mutations and the TME. Molecular profiling of 29,767 NSCLC tumor samples was performed at Caris Life Sciences to obtain whole exom sequencing, whole transcriptome sequencing and immunohistochemistry. MAPK signaling activity (MPAS) and immune cell fraction were assessed by mRNA analysis. Wilcoxon, Fisher's exact were used for statistical significance. Overall survival was calculated using the Kaplan-Meier method. Comparisons were conducted between HRAS mutated (mt) tumors and the entire NSCLC general cohort (GC). HRAS mt were detected in 0.4% of NSCLC tumors (n=128), with a majority among smokers. HRAS mt displayed less B cell, macrophage M2, NK cell, CD4+ and CD8+ T cells infiltrates compared to GC (q<0.05 ). In addition, HRAS Q61 was found to be correlated with worse prognosis in pts treated with pembro (HR = 2.779, 95% CI [1.04-7.423], p =0.03) but not G12 or G13, and displayed the highest MPAS score (p = 0.05), which had previously been demonstrated to indicate immune evasion. The median difference for survival curves was -246 days between HRAS mt squamous cell carcinomas (SCC, all smokers) and GC SCC in pts treated with pembro (HR 1.72, 95% CI 0.85-3.64, p =0.126). HRAS mt SCC showed less infiltration with M1, B cells, myeloid dendritic cells (mDC) and also displayed a significantly higher MPAS score and LAG3 expression as compared to GC (0.53 vs -0.31, p = 0.03; 2.10 vs 1.03 TPM, p = 0.04). No similar trends were observed in adenocarcinoma histology. HRAS mt NSCLC displayed a relatively immune-cold pattern with less CD4+ and CD8+ T cells infiltrates compared to GC. HRAS mt SCC also showed an immune-cold TME associated with high MAPK pathway activation and high LAG3 expression. This warrants further investigation, in particular in HRAS Q61 or HRAS mt SCC, with combination therapies targeting MAPK pathway or LAG3 protein.

Comprehensive molecular profiling to identify genomic driver mutations is inconsistently performed for squamous (sq) non-small cell lung cancer (NSCLC). Thorough profiling of this, by smoking status, will help guide diagnostic and therapeutic decision-making. Analyses included advanced sq-NSCLC tumors molecularly profiled by next-generation sequencing (NGS) using a 592 gene panel (NextSeq) with paired whole-exome and whole-transcriptome sequencing (NovaSeq). Smoking status was obtained from medical records. Genomic alterations and tumor mutation burden (TMB) were compared by smoking status. Gene set enrichment analyses (GSEA) were also assessed. Fisher's exact, Chi-square, and Mann-Whitney U tests were used, where appropriate, to assess statistical significance. Significance = p < 0.05/q < 0.05 for all comparisons except GSEA, where significance = p<0.05/FDR < 0.25. 2413 sq-NSCLC tumors with reported smoking status were assessed, of which 2347 (97.3%) and 66 (2.7%) were from ever- and never-smokers, respectively. A significantly higher median TMB (8.0 vs 5.0 mut/Mb) was seen in ever- vs never-smokers. Actionable mutations were detected in both groups as detailed in the table. A significantly higher prevalence of TP53 alterations was seen in ever-smokers (90.9% vs 61.4%), while never-smokers had a significantly higher prevalence of METex14 alterations (18.8% vs 0.3%) with a trend toward higher prevalence of actionable EGFR mutations (6.9% vs 0.6%; p<0.05, q = 0.052). GSEA identified significantly enriched expression of JAK-STAT signaling in never-smokers, with trend toward enriched interferon-gamma and interferon-alpha signaling in never-smokers (FDR 0.27).Table: 1052PMolecular AlterationNever-smokers % (n)Ever-smokers % (n)METex1418.8 (3)0.3 (4)EGFR exon19del/L858R6.9 (4)0.6 (13)BRAF V600E3.5 (2)0.2 (4)ALK fusion3.2 (1)0.1 (1)KRAS G12C0 (0)1.8 (39)HER20 (0)0.2 (4)NTRK fusion0 (0)0.08 (1)NRG1 fusion0 (0)0.08 (1) Open table in a new tab Differences in the molecular and immunologic composition of sq-NSCLC by smoking status were observed, though clinically actionable mutations were seen in both groups. Relevant information can be gleaned from NGS to guide decision-making in sq-NSCLC regardless of smoking status.

HRAS gene encodes the GTPase HRAS protein, a member of the RAS superfamily, which is enriched in high-grade urothelial bladder carcinoma (UBC). We aimed to characterize genomic and clinical outcomes of patients with HRAS mutations (HRASmt) as a possible prognostic and therapeutic biomarker. A total of 3,426 UBC tissue samples underwent molecular profiling at Caris Life Sciences utilizing NGS of DNA. MAP kinase pathway activation and the likelihood of a tumor's response to anti-PD(L)1 therapy were measured via MPAS and IFN signature, respectively. Wilcoxon, Fisher's exact were used for statistical significance. Overall survival (OS) was calculated from the date of tissue collection or the start of treatment until the last contact from insurance claims. All comparisons were conducted between HRAS mutated tumors and the entire UBC general cohort (GC). HRASmt were detected in 107 patients with metastatic UBC (3.12%). Of those, 70% were detected in primary and 30% from metastatic sites. HRASmt were associated with lower TMB (5 vs 8 mut/Mb, q<0.01) and higher PD-L1 expression (55.3% vs 39.1%, q<0.01). HRASmt tumors harbored less TP53 (25.5% vs 60.1%, q<0.05), RB1 (5.3% vs 22.1%, q<0.05), FGFR3 (1.9% vs 13.9%, q<0.05), and KMT2D (14.0% vs 25.4%, q<0.05), but more BRAF mutations (14.0% vs 2.1%, q<0.05). The most frequent HRAS point mutation involved Q61 (47.6%). HRAS-Q61 mutant tumors showed higher MPAS scores (2.21 vs 0.6, q<0.001), enrichment in pathways (all q<0.05) such as EMT (0.86 vs 0.83), TGFb (0.91 vs 0.89), angiogenesis (0.78 vs 0.75) and notch signaling (0.82 vs 0.81). Patients with HRAS-Q61 point mutations experienced worse OS and prognosis compared to HRASwt tumors when treated with checkpoint inhibitors (HR 1.45, 95% 1.01-2.10; p <0.05). HRAS-Q61 point mutations represent a clinically relevant mutation in UBC, as evidenced by its association with worse clinical outcomes compared to HRASwt tumors. Our transcriptomic analysis suggests HRAS-Q61 mutations were associated with lower response to immunotherapy and leads to MAPK activation, as well as modulation of important TME pathways, possibly indicating a distinct biological and clinical phenotype. Our study provides rationale for therapeutic HRAS targeting in UBC.

MET exon 14 skipping mutations (METex14) are a heterogeneous family of oncogenic mutations (mt) found in NSCLC that can be effectively treated with approved targeted agents. While more commonly found in lung adenocarcinoma, METex14 is also known to occur in squamous NSCLC and in patients with a smoking history, unlike many other actionable drivers. We evaluated the association of histology and smoking history with the mutational landscape among patients with NSCLC harboring METex14 mutations.

Extra-pulmonary small cell carcinomas (EPSCC) are aggressive neuroendocrine tumors that are clinicopathologically distinct from small cell lung cancer (SCLC). Recent data suggest that transcriptionally-defined SCLC subtypes exhibit different underlying biology and therapeutic vulnerabilities. With limited data available on the EPSCC transcriptomic landscape, we analyzed gene expression profiles and its correlation with clinical outcomes across EPSCC anatomic sites. DNA (592 genes/whole-exome) and RNA (whole transcriptome) sequencing were performed for 1070 small cell carcinoma (SCC) patient (pt) samples that underwent molecular profiling at Caris Life Sciences (Phoenix, AZ). Samples were stratified into 5 subtypes based on the relative expression of key transcription factors (TFs): ASCL1, NEUROD1, YAP1, POU2F3, and Mixed. Real-world survival information was available for 111 pts treated with etoposide+platinum (EP) therapy. Overall survival (OS) was obtained from insurance claims & Kaplan-Meier estimates. Statistical significance is determined by Chi-square & Wilcoxon rank sum tests. EPSCC comprised 28.8% (n=308) of SCCs evaluated; most common primary sites included gynecologic (GYN, 21.8%, n=67), prostate (16.6%, n=51), bladder (15.9%, n=49) and colorectal (10.1%, n=31). Each primary site had a unique distribution of molecular subtypes relative to SCLC. ASCL1 was less frequent in GYN and bladder SCC (13.4% and 16.3%) compared to SCLC (35.7%), with bladder SCC enriched in NEUROD1 and POU2F3 (30.6% and 22.4% vs SCLC 17.5% and 6.4%, P<0.05). YAP1 was most common in GYN SCC (35.8% vs SCLC 20.7%, P<0.05). For both SCLC and EPSCC, low YAP1 was associated with improved OS (SCLC HR 2.1, P=0.05; EPSCC HR 4.3, P=0.02), and high NEUROD1 trended towards improved OS (SCLC HR 0.6, P=0.12; EPSCC HR 0.4, P=0.10) from start of EP therapy. Our analysis revealed differential expression of key lineage-defining TFs in EPSCCs from various anatomic sites that looked distinct from SCLC. EPSCC and SCLC OS was similarly associated with TF expression, suggesting the underlying biology of SCLC and EPSCC subtypes might predict comparable therapeutic vulnerabilities.