384P Concordance of genomic alterations in tumour tissue and ctDNA in patients with colorectal cancer referred for early phase trials

Abstract

Plasma-based circulating tumour DNA (ctDNA) provides a minimally invasive, contemporaneous assessment of genomic alterations (GAs) and captures tumour heterogeneity. We compare GAs identified in ctDNA and archival tissue in patients (pts) with metastatic colorectal cancer (CRC) recruited to the Tumour Characterisation to Guide Experimental Targeted Therapy (TARGET) trial, aiming to match pts to early-phase cancer clinical trials (EPCCT). Archival tumour tissue was analysed using next generation sequencing (NGS) panels (24 genes), and plasma-based ctDNA was analysed using a local research NGS panel (641 genes). Results were discussed in a molecular tumour board (MTB) to define actionability. Factors associated with the likelihood of having additional GAs in ctDNA were explored using logistic regression analysis and odds ratio (OR). 101 pts with CRC were recruited to TARGET. The median age was 56 years and 61% were male. Pts had NGS testing panels from primary (70%) or metastatic (30%) archival tumour tissue. The median age of archival tissue tested was 25 months (range 0-117 months) and 98% of pts received active treatment in the intervening time. Tissue and ctDNA NGS identified clinically relevant GAs in 90% and 79% of pts respectively, with a median of 2 GAs per pt. The most common GAs in ctDNA were TP53 (53%), KRAS (47%), APC (44%), and PIK3CA (12%). There was poor concordance of identified GAs between tissue and ctDNA (agreement=66.3%, Cohen's k<0, p=0.91). 23% of pts had clinically relevant GAs in ctDNA that were not identified in tissue, the commonest being KRAS (29%). There was no association between the presence of additional GAs in ctDNA with number of treatment lines (OR:1.23, p= 0.32), age of archival tissue (OR: 1.01, p= 0.91) or primary vs. metastatic archival tissue site (OR: 0.34, p= 0.11). The presence of liver metastases was associated with a higher likelihood of concordance between tissue and ctDNA NGS (OR: 3.5, p= 0.03). Overall, ctDNA-derived GAs were considered actionable in 51% of pts. Of those, 28% were a potential match for an EPCCT. Plasma-based ctDNA can be used as a contemporaneous tool to identify actionable GAs and guide treatment options for pts with metastatic CRC.