Development and validation of a clinical prognostic model for BRAF V600E-mutated colorectal cancer patients based on pathological stage, microsatellite status, and primary tumor site

Objective: To develop and validate a prognostic model for patients with BRAF V600E-mutated colorectal cancer.

Methods: The clinical and pathological information of 206 patients with BRAF V600E-mutated colorectal cancer diagnosed in Cancer Hospital, Chinese Academy of Medical Sciences, and Peking Union Medical College from 2014 to 2021 was retrospectively collected. Least absolute shrinkage and selection operator (LASSO) regression, Cox regression, and nomograms were used to develop clinical prognostic models. The differentiation was measured using C-statistic, and the predicted variability was evaluated using the calibration curve. The prognostic model was externally validated using validation set data from 164 patients pooled from five studies.

Results: Our clinical prognostic model included three variables: pathological stage, microsatellite status, and primary tumor site. In internal validation, the model had a concordant index of 0.785 (95% CI [0.732-0.839]) and a concordant index of 0.754 (95% CI [0.698-0.810]) using pathological staging. External validation confirmed the robustness of the model with a consistency index of 0.670 (95% CI [0.617-0.724]) and a consistency index of 0.584 (95% CI [0.546-0.622]) using pathological staging. Likelihood ratio test results show that our model is better (internal validation, p = 5.141e-03; external validation, p = 2.728e-05). The calibration graph drawn based on the prediction and the actual situation is close to the 45° diagonal.

Conclusion: By adding microsatellite status and primary tumor site on the basis of pathological stage, we improved the discriminability and prediction accuracy of the model and successfully established a prognosis model for patients with BRAF V600E mutation of colorectal cancer.