Role of Tumor-informed Personalized Circulating Tumor DNA Assay in Informing Recurrence in Patients With Peritoneal Metastases From Colorectal and High-grade Appendix Cancer Undergoing Curative-intent Surgery

Objective: To investigate the role of a personalized, tumor-informed circulating tumor DNA (ctDNA) assay in informing recurrence in patients with peritoneal metastases (PM) from colorectal (CRC) and high-grade appendix (HGA) cancer after curative cytoreductive surgery with hyperthermic intraperitoneal chemotherapy (CRS-HIPEC). Background: Over 50% of patients with CRC/HGA-PM recur after optimal CRS-HIPEC. The limited sensitivity of axial imaging and diagnostic biomarkers is a significant cause of delay in the detection of recurrence and initiation of further therapies. Plasma ctDNA has a promising role in monitoring response to treatment and/or recurrence after primary cancer resection. Methods: Patients with CRC/HGA-PM who underwent curative CRS-HIPEC and serial postresection ctDNA assessments were included. Patients with rising postoperative ctDNA levels were compared with those with stable, undetectable ctDNA levels. Primary outcomes were the percentage of patients with recurrence and disease-free survival (DFS). Secondary outcomes were overall survival, ctDNA sensitivity, lead time, and performance of ctDNA compared with carcinoembryonic antigen. Results: One hundred thirty serial postresection ctDNA assessments [median 4, interquartile range (IQR), 3 to 5] were performed in 33 patients (n = 13 CRC, n = 20 HGA) who underwent completeness of cytoreduction-0/1 CRS with a median follow-up of 13 months. Of the 19 patients with rising ctDNA levels, 90% recurred versus 21% in the stable ctDNA group (n = 14, < 0.001). Median DFS in the rising ctDNA cohort was 11 months (IQR, 6 to 12) and not reached in the stable (P = 0.01). A rising ctDNA level was the most significant factor associated with DFS (hazard ratio: 3.67, 95% CI: 1.06-12.66, P = 0.03). The sensitivity and specificity of rising ctDNA levels in predicting recurrence were 85% and 84.6%, respectively. The median ctDNA lead time was 3 months (IQR, 1 to 4). Carcinoembryonic antigen was less sensitive (50%) than ctDNA. Conclusions: This study supports the clinical validity of serial ctDNA assessment as a strong prognostic biomarker in informing recurrence in patients with CRC/HGA-PM undergoing curative resection. It also holds promises for informing future clinical trial designs and further research.