Abstract Background Currently, no international guidelines recommend carbohydrate antigen 19-9 (CA19-9) as a biomarker in colorectal cancer in clinical practice, although it has been reported that tumour cells with high affinity for peritoneal cells strongly express CA19-9. The aim of this study was to determine the utility of preoperative serum CA19-9 levels in predicting prognosis in patients with colorectal cancer and synchronous peritoneal metastasis. Methods Prognostic analyses were retrospectively performed in patients with colorectal cancer and synchronous peritoneal metastases who received treatment in 16 hospitals between 1991 and 2007 (Cohort 1). The categorization for CA19-9 and carcinoembryonic antigen were established based on the thresholds that minimized Akaike information criterion values for overall survival. A separate validation cohort of patients with synchronous peritoneal metastasis was prospectively registered across 28 institutions between 2012 and 2016 (Cohort 2) to validate these findings. Results A total of 689 patients were analysed in Cohort 1 and 144 in the validation Cohort 2. The categorized serum CA19-9 and carcinoembryonic antigen levels were significantly correlated with overall survival (P 0.001 and P 0.001, respectively). The median postoperative overall survival was 19.0 months for patients with a preoperative CA19-9 level 100 U/ml compared with 4.1 months for those with a preoperative CA19-9 level of ≥ 10 000 U/ml. Using multivariable analysis, serum CA19-9, but not carcinoembryonic antigen, was an independent prognostic factor for overall survival. These results were confirmed in Cohort 2, revealing that the prognostic power of CA19-9 was superior to that of carcinoembryonic antigen. Conclusions Preoperative serum CA19-9 outperformed carcinoembryonic antigen as an independent prognostic biomarker in patients with colorectal cancer and peritoneal metastasis.
Tashiro et al. (Thu,) studied this question.