Dynamic monitoring of carcinoembryonic antigen, CA19-9 and inflammation-based indices in patients with advanced colorectal cancer undergoing chemotherapy

BACKGROUND: The roles of carcinoembryonic antigen (CEA) and carbohydrate antigen (CA19-9) in monitoring the patient response to chemotherapy for metastatic colorectal cancer (mCRC) are not clearly defined, and inflammatory indices, including the neutrophil-to-lymphocyte ratio (NLR), lymphocyte-to-monocyte ratio (LMR), platelet-to-lymphocyte ratio (PLR) and systemic immune-inflammation index (SII), have been sparsely investigated for this purpose. AIM: To aim of this study was to evaluate the relationship between the kinetics of CEA, CA19-9, NLR, LMR, PLR and SII in serum and patient response to chemotherapy estimated by computed tomography (CT) in patients with unresectable mCRC. METHODS: Patients with mCRC treated with a 1(st)-line and 2(nd)-line chemotherapy underwent at least 3 whole-body spiral CT scans during response monitoring according to the Response Evaluation Criteria in Solid Tumour 1.1 (RECIST 1.1), and simultaneous determination of CEA, CA19-9, neutrophil, lymphocyte, platelet and monocyte levels was performed. The kinetics of changes in the tumour markers and inflammatory indices were calculated as the percentage change from baseline or nadir, while receiver operating characteristic curves were drawn to select the thresholds to define patients with progressive or responsive disease with the highest sensitivity (Se) and specificity (Sp). The correlation of tumour marker kinetics with inflammatory index changes and RECIST response was determined by univariate and multivariate logistic regression analysis and the clinical utility index (CUI). RESULTS: A total of 102 patients with mCRC treated with chemotherapy were included. Progressive disease (PD), defined as a CEA increase of 25.52%, resulted in an Se of 80.3%, an Sp of 84%, a good CUI negative [CUI (Ve-)] value of 0.75 and a good fraction correct (FC) value of 81.2; at a CEA cut-off of -60.85% with an Se of 100% and an Sp of 35.7% for PD, CT could be avoided in 25.49% of patients. The 21.49% CA19-9 cut-off for PD had an Se of 66.5%, an Sp of 87.4%, an acceptable CUI (Ve-) value of 0.65 and an acceptable FC value of 75. An NLR increase of 11.5% for PD had an Se of 67% and an Sp of 66%; a PLR increase of 5.9% had an Se of 53% and an Sp of 69%; an SII increase above -6.04% had an Se of 72% and an Sp of 63%; and all had acceptable CUI (Ve-) values at 0.55. In the univariate logistic regression analysis, CEA (P < 0.001), CA19-9 (P < 0.05), NLR (P < 0.05), PLR (P < 0.05) and SII (P < 0.05) were important predictors of tumour progression, but in the multivariate logistic regression analysis, CEA was the only independent predictor of PD (P < 0.05). CONCLUSION: CEA is a useful marker for monitoring the chemotherapy response of patients with unresectable mCRC and could replace a quarter of CT examinations. CA19-9 has poorer diagnostic characteristics than CEA but could be useful in some clinical circumstances, particularly when CEA is not increased. Dynamic changes in the inflammatory indices NLR, PLR and SII could be promising for further investigation as markers of the chemotherapy response.