2255P - Clinicogenomic landscapes and hallmarks of KRAS amplification in human cancers

Background

The recent development of direct KRAS inhibitors has changed the treatment landscape of patients (pts) with oncogenic KRAS mutations. However, whether KRAS amplification (KRAS^AMP) also defines a unique subset of cancers and represents a potential target for drug development is unknown.

Methods

Clinicogenomic data and outcomes were abstracted from independent cohorts of human cancers with genomic profiling at the Dana-Farber Cancer Institute, AACR-GENIE v.13, The Cancer Genome Atlas (TCGA), and 212 other studies, and examined according to KRAS^AMP status. We also correlated KRAS^AMP with KRAS RNA and protein expression using whole transcriptome and proteomics data from the TCGA and the Cancer Cell Lines Encyclopedia.

Results

Among 103,523 pts with solid and hematologic cancers, KRAS^AMP was identified in 1,744 (1.7%) of cases. KRAS^AMP was most prevalent in germ cell tumors (16.3%), esophagogastric (7.5%), ovarian (4.2%), pancreatic (2.2%), breast (2%), colorectal (CRC, 1.2%), and non-small cell lung cancer (NSCLC, 1.2%). Among these pts, median age was 61, 61.5% were ever smokers, 85.4% were white, and 71.8% had poorly differentiated cancers. Compared to KRAS diploid cases, KRAS^AMP cancers had a higher fraction of the genome altered (p<0.01). Microsatellite instability scores were also higher in KRAS^AMP tumors from patients with breast, esophagogastric, germ cell, pancreatic and NSCLC (p<0.01), but not in CRC (p=0.22). Tumor mutational burden was higher (p<0.01) among CRC, NSCLC, and esophagogastric cancers with KRAS^AMP vs KRAS diploid. Compared to KRAS diploid samples, KRAS^AMP samples had increased KRAS mRNA (p<0.01) and protein expression (p<0.01). Among pts with clinical outcome data (N=41,900), median overall survival (OS) was shorter in all comers with KRAS^AMP vs KRAS diploid genotype (adjusted HR 1.68, p<0.01), and in each cancer type analyzed separately. KRAS^AMP conferred worse survival outcomes in pts with both KRAS wild-type (N=35,599, adjusted HR 1.38, p=0.01) and KRAS-mutant (N=6,301, adjusted HR 1.70, p<0.01) cancers.

Conclusions

KRAS amplification defines a novel subset of human cancers characterized by distinct genomic features, and worse survival. Novel therapeutic strategies to target KRAS^AMP may improve outcomes in pts with cancer.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

Biagio Ricciuti.

Funding

Has not received any funding.

Disclosure

B. Ricciuti: Financial Interests, Personal, Advisory Board: Regeneron. P.A. Jänne: Financial Interests, Personal, Advisory Board, Consulting fees for advice on drug development: AstraZeneca, Boehringer Ingelheim, Pfizer, Roche/Genentech, Chugai Pharmaceuticals, Eli Lilly, Voronoi, Daiichi Sankyo, Novartis, Sanofi, Takeda Oncology, Mirati Therapeutics, Trasncenta, Silicon Therapeutics, Syndax, Nuvalent, Bayer, Eisai, Allorion Therapeutics, Accutar Biotech, AbbVie, Duality Biologics; Financial Interests, Personal, Advisory Board, Consulting fees for advice on diagnostic development: Biocartis; Financial Interests, Personal, Advisory Board, Consulting fee for advice on drug development: Merus, Frontier Medicines; Financial Interests, Personal, Advisory Board, Consulting fees for advice on drug development: Hongyun Biotechnology; Financial Interests, Personal, Stocks/Shares: Gatekeeper Pharmaceuticals, Allorion Therapeutics; Financial Interests, Personal, Royalties, I receive post-marketing royalties from being an inventor on a DFCI owned patent on EGFR mutations licensed to Lab Corp: Lab Corp; Financial Interests, Institutional, Research Grant, Sponsored research agreement with my institution: AstraZeneca, Daiichi Sankyo, PUMA, Eli Lilly, Boehringer Ingelheim, Revolution Medicines, Takeda Oncology. M.M. Awad: Financial Interests, Personal, Advisory Board: Merk, BMS, AZ, Genentech, Nekter, Maverick, Blueprint Medicine, AbbVie, Gritstone, ArcherDX, Mirati; Financial Interests, Advisory Board: NextCure. All other authors have declared no conflicts of interest.