Evaluation of Comparative Surveillance Strategies of Circulating Tumor DNA, Imaging, and Carcinoembryonic Antigen Levels in Patients With Resected Colorectal Cancer

IMPORTANCE: A circulating tumor DNA (ctDNA) assay (Signatera; Natera) has been marketed for use in the surveillance of resected colorectal cancer despite limited data supporting such practice. OBJECTIVE: To compare a ctDNA assay with standard radiographic imaging and measurement of carcinoembryonic antigen (CEA) levels, per National Comprehensive Cancer Network guidelines, in the surveillance of resected colorectal cancer. DESIGN, SETTING, AND PARTICIPANTS: This retrospective, single-center cohort study evaluated surveillance strategies of ctDNA, imaging, and measurement of CEA levels in patients with resected colorectal cancer from September 1, 2019, to November 30, 2021. MAIN OUTCOMES AND MEASURES: The sensitivity and specificity of ctDNA, imaging, measurement of CEA levels, and combination of imaging plus measurement of CEA levels in detecting a confirmed recurrence of colorectal disease. A confirmed recurrence was defined as a positive ctDNA finding or a finding on imaging confirmed by biopsy, CEA level elevation, or subsequent tumor radiographic dynamics. RESULTS: A total of 48 patients with curatively resected colorectal cancer satisfied the inclusion criteria for this study (28 men [58.3%]; median age, 60 [IQR, 34-85] years) and underwent surveillance by ctDNA, imaging, and measurement of CEA levels. Fifteen patients had disease recurrence during surveillance. Positive ctDNA findings confirmed disease recurrence in 8 patients; imaging, in 9 patients; CEA levels, in 3 patients; and combined imaging plus CEA levels, in 11 patients. Numerically, ctDNA did not perform better than imaging in detecting recurrence, with sensitivities of 53.3% (95% CI, 27.4%-77.7%) and 60.0% (95% CI, 32.9%-82.5%), respectively (P > .99). The combination of imaging plus measurement of CEA levels (sensitivity, 73.3% [95% CI, 44.8%-91.1%]) had a numerical advantage compared with ctDNA in identifying recurrence (P = .55). In addition, no significant difference was noted among ctDNA (median, 14.3 months), imaging (median, 15.0 months), or imaging plus measurement of CEA levels (median, 15.0 months) in the time to identify disease recurrence. CONCLUSIONS AND RELEVANCE: The findings of this cohort study suggest that ctDNA assay may not provide advantages as a surveillance strategy compared with standard imaging combined with CEA levels when performed per National Comprehensive Cancer Network guidelines.