Abstract B49: 'Triple wild-type' co-mutational profile in early-stage KRAS-mutant lung cancer

Introduction: Kirsten rat sarcoma viral oncogene (KRAS) is the most frequently mutated oncogene in non-small cell lung cancer (NSCLC), occurring in approximately 30% of cases. Recent work suggests KRAS -mutant (KRASm) NSCLC is a microcosm for diverse immune checkpoint inhibitor responses, partly dictated by its co-mutation with TP53 (“KP” group=~1/3 cases, “immune hot”), STK11 (“KL” group=~1/3 cases, “immune cold”) or CDKN2A (“KC” group=~1/3 cases). Here we identify and describe genomic and clinical associations of a further subset of KRASm NSCLC defined by “triple WT” status for common KRASm co-mutations. Method: Sequencing results from 364 early-stage NSCLC cases from the Cancer Research UK TRACERx program (whole-exome sequencing) and the Genomics England 100,000 genome program (whole-genome sequencing) were analyzed for KRASm, associated co-mutations in TP53, STK11 and CDKN2A, copy number changes in the RAS-RAF-MEK-ERK axis, tumor mutational burden (TMB) and mutational signatures. Clinical demographics and factors including tumor size, nodal status and stage were ascertained and assessed for statistical associations with sequencing data using the Mann-Whitney and Fisher’s exact tests. Results: Overall, 143/364 lung cancer cases (39%) were KRASm with 65/143 (45%) tumors identified as “KP,” 26/143 (18%) as “KL,” and 3/143 (2%) as “KC.” There were 49/143 KRASm cases (34%) with no co-mutation in TP53, STK11 or CDKN2A, defined as “triple wild-type” (triple WT). Relative to the KP, KL and KC genotypes, there was a positive association of the KRAS G12D mutation with triple WT status: 10/49 cases, 20%; KP/KL/KC: 6/94 cases, 6%; p=0.022). In the 100,000-genome cohort, TMB was similar for the triple WT and KL groups but significantly lower than that observed for KP (KP: median 10.55 mut/Mb, 95% CI 8.8-15.33; triple WT: median 6.96, 95% CI 4.87-12.53; p=0.036). Despite this lower median TMB, smoking-associated mutational signature 4 was common in triple WT tumors (triple WT: 13/14 pts, 93%; overall 76/114 pts, 67%). In the triple WT group overall, there was a reduction of >50% in cancers with copy number changes of NF1 and NRAS. There was an inverse association between pathologic stage III tumors and KRASm triple WT genotype (p=0.0014). Conclusion: There are a significant proportion of KRASm NSCLC patients whose tumors are triple WT for TP53, STK11 and CDKN2A. Despite their high frequency of smoking-associated mutational signatures, these tumors are characterized by low TMB and are more common in early-stage disease. They also associate more commonly with KRAS G12D. Our observations suggest a discrete KRASm subset that may have implications for stratification in trials of immune checkpoint inhibitors and/or targeted therapeutics of KRASm tumors. Citation Format: Colin R. Lindsay, Pantelis Nicola, Mariam Jamal-Hanjani, Andrew Wallace, Gareth Wilson, George Burghel, Helene Schlecht, Katie Baker, Eleanour Baker, Lynsey Priest, Jane Rogan, Sharzad Moghadam, Mathew Carter, Caroline Dive, Robert G. Bristow, Charles Swanton, William Newman, Fiona Blackhall. “Triple wild-type” co-mutational profile in early-stage KRAS-mutant lung cancer [abstract]. In: Proceedings of the AACR Special Conference on Targeting RAS-Driven Cancers; 2018 Dec 9-12; San Diego, CA. Philadelphia (PA): AACR; Mol Cancer Res 2020;18(5_Suppl):Abstract nr B49.