ABSTRACT
Colorectal cancer patients with synchronous liver metastases (CRSLM) can be treated by simultaneous surgery, that is the primary tumor and liver metastasis are removed at the same time. However, criteria for simultaneous surgery are underwent continuously modified and expanded. An appropriate selection of adequate candidates for simultaneous surgery is vital to get best benefits. A retrospective study including CRSLM patients underwent simultaneous surgical treatment was conducted. CRSLM patients from SEER database were screened as development set, while CRSLM patients in Harbin (China) were enrolled as validation set. Overall survival (OS) and cancer-specific survival (CSS) were applied as end-point. Variables were screen by LASSO-Cox regression, then Cox regression was applied to construct 1-, 3- and 5-year OS, and CSS nomograms. Nomograms were compared to TMN stage for survival prediction and evaluated by concordance indexes (C-indexes), Time-dependent receiver operating characteristic (ROC) curves, Decision Curve Analysis (DCA). 1347 and 112 CRSLM patients were included in the development set and validation set respectively. Nine factors were found associated with OS and CSS, i.e., Age, Primary Site, Differentiation grade, Histology type, T stage, N stage, Tumor size, Chemotherapy, CEA. Compared to the TNM stage, OS nomogram in development set and validation set got C-indexes values of 0.701 vs 0.641, 0.670 vs 0.557 respectively. Meanwhile, compared to the TNM stage, CSS nomogram in development set and validation set got C-indexes values of 0.704 vs 0.649, 0.677 vs 0.569 respectively. AUC values of the OS and CSS nomograms were higher than the TNM stage, DCA showed the OS and CSS nomograms got more clinical net benefit than the TNM stage, in both the development set and validation set. Our nomograms for predicting survival might be helpful to identify the right CRSLM patients who can get most benefit from simultaneous surgery.
