Abstract 457P
Background
Comprehensive genomic profiling (CGP) of plasma ctDNA provides a minimally invasive method to profile CRC. The landscape of genomic alterations in CRC using ctDNA based CGP was compared to tissue based CGP.
Methods
Plasma from 837 pts with CRC was analyzed using FoundationOneLiquid CDx and compared with 24,584 CRC FoundationOne CDx tissue biopsy results. In pts with both specimens, positive percent agreement (PPA) was calculated at the variant level, with tissue as reference. Tumor fraction (TF) was estimated using a measure of aneuploidy; variant germline status was computationally predicted.
Results
790 pts (94.4%) had detectable ctDNA variants. Compared to tissue, CGP of ctDNA was enriched for mutations in genes mediating resistance to EGFR inhibitors (EGFR, MAP2K1, p < 1E-5), likely reflecting a higher proportion of patients who have received targeted therapy. EGFR mutations were detected in 40 liquid biopsies (5%), 12 with multiple (up to 6). 53 of 64 EGFR mutations mapped to the ECD. 31 liquid biopsies (3.7%) had MAP2K1 mutations; 6 with multiple). Mutations in clonal hematopoiesis genes were also enriched in ctDNA. CGP of ctDNA otherwise resembled CGP of tissue: commonly altered genes included TP53 (66%), APC (62%), KRAS (43%), PIK3CA (14%), SMAD4 (9%), BRAF (8%), and SOX9 (6%). 30 liquid biopsies (4%) detected actionable rearrangements (8 BRAF, 7 EGFR, 5 FGFR2, 5 FGFR3, 4 ALK, 3 ROS1, 2 RET, 1 MET, and 1 NTRK1). Germline variants detected in ctDNA included MUTYH (1.4%, including 2 with biallelic inactivation), APC (1.1%), and MLH1/MSH2/MSH6/PMS2 (0.84%). Liquid and tissue CGP results from 147 pts were used for concordance analysis (median time difference: 15 months). PPA for ctDNA detection of KRAS and PIK3CA mutations was 75% (59/79) and 62% (16/26), increasing to 100% (38/38) and 82% (9/11) in 69 liquid biopsies with TF >10%.
Conclusions
CGP of plasma from pts with CRC contains rich ctDNA signal and recapitulates the genomic landscape detected in tissue biopsies. ctDNA-based detection of KRAS and BRAF alterations, targetable fusions, and possible resistance mutations, suggests this may be a compelling alternative to tissue CGP when a tissue specimen is inadequate or unavailable.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
Foundation Medicine.
Funding
Foundation Medicine.
Disclosure
H. Tukachinsky: Financial Interests, Personal, Full or part-time Employment: Foundation Medicine; Financial Interests, Personal, Stocks/Shares: Roche. L. Zhang: Financial Interests, Personal, Full or part-time Employment: Foundation Medicine; Financial Interests, Personal, Stocks/Shares: Roche. A.B. Schrock: Financial Interests, Personal, Full or part-time Employment: Foundation Medicine; Financial Interests, Personal, Stocks/Shares: Roche. B. Decker: Financial Interests, Personal, Full or part-time Employment: Foundation Medicine; Financial Interests, Personal, Stocks/Shares: Roche; Financial Interests, Personal, Stocks/Shares: Avidea Technologies. D.C. Pavlick: Financial Interests, Personal, Full or part-time Employment: Foundation Medicine; Financial Interests, Personal, Stocks/Shares: Roche. J. Venstrom: Financial Interests, Personal, Full or part-time Employment: Foundation Medicine; Financial Interests, Personal, Stocks/Shares: Roche. H. Nimeiri: Financial Interests, Personal, Full or part-time Employment: Foundation Medicine; Financial Interests, Personal, Stocks/Shares: Roche. G. Oxnard: Financial Interests, Personal, Full or part-time Employment: Foundation Medicine; Financial Interests, Personal, Stocks/Shares: Roche. All other authors have declared no conflicts of interest.