Abstract
BACKGROUND: In mCRC, CEA is used to monitor response to systemic therapy together with imaging. After the end of induction, no major improvement in tumour shrinkage is expected, and the availability of a marker able to predict progressive disease (PD) versus no-PD might allow avoiding CT scans. METHODS: We pooled data from patients with baseline CEA ≥ 10 ng/mL included in TRIBE and TRIBE2 studies with the aim of identifying a threshold for percent increase of CEA from nadir able to predict PD after the end of the induction therapy. RESULTS: In total, 1178 paired CEA and radiological assessments from 434 patients were included. According to the optimal cut-off determined by ROC, a CEA increase of at least 120% from nadir differentiated between PD and no-PD with a sensitivity of 74% and a specificity of 78%, excluding PD in the 92% of radiological assessments and allowing to avoid the 67% of CT scans. However, CEA cut-off of 120% was not able to detect radiological PD in 26% of cases. In order to mitigate this issue, a different clinically relevant threshold was evaluated based on the best sensitivity cut-off. Therefore, using any CEA increase from nadir as a threshold, the sensitivity grew to 93% and only in the 7% of cases the radiological PD was not detected. CONCLUSIONS: In mCRC with baseline CEA ≥ 10 ng/mL, CEA values can accurately predict PD versus no-PD after the end of the first-line induction therapy.