Circulating tumor DNA (ctDNA) dynamics and clinical outcome in metastatic colorectal cancer (mCRC) patients (pts) undergoing front-line chemotherapy

175 Background: The potential of monitoring circulating tumor DNA (ctDNA) dynamics to guide clinical decisions in metastatic colorectal cancer (mCRC) patients (pts) treated with I and II line systemic anti-cancer therapy (SACT) has not been widely tested. Methods: 862 serial plasmas were collected 4-weekly from baseline (BL) until disease progression in 75 mCRC pts undergoing SACT. ctDNA was tested using a custom (RMH GI; 20 genes) or a commercial (Roche Avenio; 77 genes) ctDNA next generation sequencing (NGS) panel. White blood cells were sequenced to rule out clonal hematopoiesis. Whole exome sequencing (WES) was performed on tissue biopsies. ctDNA normalization was defined as ≥99% clearance after 1 month of therapy (Mo1) in the 3 variants with the highest allele frequency in BL ctDNA. Results: 83 paired samples from 75 pts were available for analysis (for 8 pts, I and II line bloods were available). 12 pairs (14.4%) showed no variants in either BL or Mo1. In the remaining 71 comparisons (65 pts), 37 (52.1%) showed ctDNA normalization at Mo1. Among normalized pts there was a higher proportion of cases with a baseline ECOG performance status (PS) 0-1 (97.3% vs 82.4%, p = 0.0362) in comparison to non-normalized pts, whilst no other clinic-pathologic characteristics, including age, sex, prior primary tumor resection, sidedness, RAS/RAF genotype, type of regimen and number of metastatic sites were significantly associated with ctDNA dynamics. Pts with normalized ctDNA had significantly longer overall survival (OS), 45.6 months (95% confidence interval [CI]: 30.0 - not reached, 14 events) and progression-free survival (PFS), 13.9 months (95%CI: 11.2-18.3; 30 events) compared to non-normalized pts [OS = 22.6 months (95%CI: 16.6-31.2, 24 events) (Log-rank p = 0.01) and PFS = 10.7 months (95%CI: 7.53-13.8; 32 events) (Log-rank p = 0.036) respectively]. In addition, pts with normalized ctDNA had higher overall response rate (ORR) of 72.9% (27/37 responses; 95%CI: 53.0-84.1) compared to 38.2% (13/34 responses; 95%CI: 22.1 – 56.4) in non-normalized pts. In a multivariate model, ctDNA normalization was confirmed as an independent predictor of decreased risk of death (hazard ratio [HR] 0.47, 95%CI: 0.23-0.96; p = 0.04) and higher probability of achieving an objective response from front-line treatment (odds ratio [OR] 3.03, 95%CI: 1.08-8.49; p = 0.0351). Only 23/50 (46%) of variants detected in ctDNA were detected by WES in paired tissues in 12 pts for whom liquid/solid biopsy was available. Conclusions: ctDNA monitoring represents an early indicator of benefit from systemic therapy in mCRC pts. A significant fraction of variants detected in ctDNA was not detected in paired tissues.