Performance comparison of the methylated BCAT1/IKZF1 ctDNA test (COLVERA) with the CEA assay for detection of recurrent colorectal cancer

Research output: Contribution to journalMeeting Abstractpeer-review


Background: Early detection of recurrent colorectal cancer (CRC) will improve treatment options, but the current standard blood test of carcinoembryonic-antigen (CEA) has suboptimal sensitivity for recurrence. This study compared performance of a quantitative circulating tumor DNA (ctDNA) assay for methylated BCAT1 and IKZF1 (COLVERA) with that of CEA. Methods: 301 patients were monitored for recurrence after clearance of primary CRC. Blood was collected at scheduled intervals and concentrations of CEA and ctDNA were measured using the LIAISON CEA test (Diasorin) and the COLVERA ctDNA test (Clinical Genomics). Surveillance for recurrent disease was examined using regular CT scans. Sensitivity of each blood test for recurrence was assessed in the sample collected closest to the time of imaging confirming recurrence status. Absence of recurrence was defined as at least two consecutive clear CT scans. Receiver operator characteristic (ROC) analyses were used to determine optimal positivity threshold for Colvera. Results: 131 patients underwent satisfactory assessment for recurrence and had blood testing performed within 12 months of determining recurrence status (61.8% male, mean age 62.6 ±.12.2(SD)y). Of the 47 recurrence cases, 37 (74%) were distant. The areas under the ROC curves were 0.7761 and 0.8188 for CEA and COLVERA, respectively (each p < 0.001). An optimal cut-off of 12.8pg/sample was determined for COLVERA and the standard 5ng/mL cut-off was selected for CEA. COLVERA had a significantly higher sensitivity for detecting recurrence as compared to CEA (68.1% vs 31.9%, p < 0.001) with a similar specificity (97.6% Vs 96.4%, p = 0.6547). A multivariate analysis determined COLVERA to be a predictor of recurrence independent of CEA with positive COLVERA samples being 66.4 times (95%CI 14.0-315.8) more likely to have recurrence confirmed within the study timeframe, whereas CEA was not a significant predictor of recurrence (p = 0.228). Conclusions: These findings indicate that COLVERA, reporting in quantitative mode, is a more sensitive test than CEA. It provides a viable alternative for sensitive and early detection of recurrent CRC
Original languageEnglish
Pages (from-to)3589
Number of pages1
JournalJournal of Clinical Oncology
Issue number15 Suppl
Publication statusPublished - 20 May 2019

Cite this