Reassessing the role of CEA in the treatment of rectal cancer in the era of TNT

Background Measurement of carcinoembryonic antigen (CEA) is an integral part of surveillance strategies in rectal cancer. We evaluated the prognostic significance of CEA level before total neoadjuvant therapy (TNT) and its role in surveillance strategies after TNT. Patients and methods This post hoc analysis of the CAO/ARO/AIO-12 trial involved 306 patients who were treated with induction chemotherapy (CT) followed by chemoradiotherapy (CRT), or CRT followed by CT, and then underwent total mesorectal excision. Results Baseline CEA levels were available for 282 patients (92%). The median baseline CEA level was 5.45 ng/ml for patients with ypN+, 3.35 ng/ml for patients with ypT+, and 2.45 ng/ml for patients with complete response (CR: pathological CR + clinical CR). Only baseline CEA levels and baseline white blood cell count remained significant parameters in a best-fit generalized linear regression model predicting CR (area under the curve 0.725). In a multivariate Fine–Gray model, CEA was significantly associated with treatment failure (TF); TNT only showed a significant reduction in the risk of TF in the subgroup of patients with baseline CEA >10 ng/ml compared with patients treated with 5-fluorouracil (5-FU) CRT and adjuvant 5-FU in the CAO/ARO/AIO-04 trial. The risk of TF was 13% in patients with no CEA elevation >3 ng/ml during follow-up, compared with 54% (63%) in patients with CEA elevation >5 ng/ml (10 ng/ml). No local recurrence after TNT was associated with an elevated CEA level >5 ng/ml. Conclusions Baseline CEA levels should be considered an integral part of risk-adapted treatment strategies and remain an essential and cost-effective part of surveillance strategies after TNT.