Development of a Prognostic Model for Stage IV Colorectal Cancer Using Metastatic Patterns

Abstract

Background/aim: Metastatic patterns are the most convenient and common prediction models for the prognosis of patients with stage IV colorectal cancer. However, current prediction models do not include the severity of metastases in organs and exclude certain types of metastatic patterns. The aim of this study was to develop a prediction model that included several metastatic organs as well as the severity of liver and lung metastases, based on the Japanese Classification of Colorectal, Appendiceal, and Anal Carcinoma: the 3^rd English Edition.

Patients and methods: We performed a state-wide cohort study and developed a prediction model using Cox proportional hazard regression analysis, utilizing data on patients with stage IV colorectal cancer in hospital-based cancer registries of all nine designated cancer hospitals across Fukushima Prefecture, Japan.

Results: The study included 1,230 patients with stage IV colorectal cancer. The prediction score consisted of the severity of liver and lung metastases, peritoneal dissemination, non-regional lymph node metastases, and other organ metastases (scale: 0-9 on a 10-point scale; divided into a 2-point scale, grade: I-V). The study found that the model had good discrimination properties, with a Harrell's concordance index of 0.64 (95% confidence interval: 0.62-0.66), and the grade was an independent prognostic factor [hazard ratio (HR)=1.83; 95% confidence interval=1.68-2.00; p<0.001].

Conclusion: We created a practical prediction model for stage IV colorectal cancer that can be applied at the time of diagnosis, using only metastatic patterns. Further external validation studies are required to ensure the accuracy of this model.

Keywords: Prediction model; liver metastasis; lung metastasis; metastatic colorectal cancer; overall survival.