Circulating tumor DNA in plasma as an early predictor for pathological lymph node involvement and (neo-)adjuvant chemotherapy in patients with colon cancer.

Abstract

3532 Background: Postoperative chemotherapy is standard for colon cancer patients with pathological lymph node involvement (pN+). If the presence of pN+ could be accurately predicted prior to surgery, the initiation of neoadjuvant chemotherapy could potentially reduce the presence of micrometastasis and enhance survival. The objective of this study was to assess the utility of preoperative methylated circulating tumor DNA (meth-ctDNA) as a predictive marker for pN+ and its clinical value as a marker for the initiation of neoadjuvant chemotherapy. Methods: The clinical data of 203 colon cancer patients (stage I-III) was collected from the Colorectal Cancer Database established at Danish Colorectal Cancer Center South, Vejle Hospital, Denmark. Patients were randomized into a discovery and a validation cohort. Plasma collected preoperatively was analyzed for three different meth-ctDNAs (Neuropeptide Y, Galactose-3-O-Sulfotransferase 3, and KN Motif and Ankyrin Repeat Domains 1) by droplet digital PCR (ddPCR). Results: Based on analysis of the discovery cohort, samples were considered positive if two or more of the three methylation markers had at least three positive droplets in the ddPCR analysis. 40% of patients in the discovery cohort and 46% of patients in the validation cohort were considered positive. When using meth-ctDNA as a diagnostic tool to predict pN+, the sensitivity and specificity in the discovery cohort were 42% and 61% respectively. In the validation cohort the values were 54% and 59%, respectively. AUC was < 60% in both cohorts. Meth-ctDNA was associated to clinical tumor (T)- and node (N)-category in both cohorts (p≤0.01). Four-year disease free survival (DFS) in patients with and without detectable meth-ctDNA was 60% and 90% in the discovery cohort (p = 0.01), and 62% and 78% in the validation cohort (p = 0.04). Meth-ctDNA remained the strongest prognostic factor of DFS in the multivariate analysis compared to clinical T- or N-category, gender and age (p = 0.065, HR 2.46, 95% CI 1.03-5.88), however not statistically significant. Four-year overall survival (OS) in patients with and without detectable meth-ctDNA was 75% and 95% in the discovery cohort (p = 0.006), and 78% and 91% in the validation cohort (p = 0.02). In the multivariate analysis, meth-ctDNA was the strongest prognostic factor of OS compared to clinical T- or N-category, gender and age (p = 0.016, HR 4.54, 95% CI 1.32-15.6). Conclusions: The preoperative measurement of meth-ctDNA was not found to be a clinically significant predictor of pN+ or a reliable indicator for the initiation of neoadjuvant chemotherapy. Nonetheless, the results showed that preoperative ctDNA is a powerful prognostic indicator, suggesting that randomized controlled trials should be conducted to determine the efficacy of neoadjuvant chemotherapy in colon cancer patients with preoperative detectable ctDNA.