Development and validation of a novel survival prediction model for newly diagnosed lower-grade gliomas

Diffuse gliomas are the most common primary gliomas with a poor prognosis. This study aimed to develop and validate prognostic models for predicting the survival probability in newly diagnosed lower-grade glioma (LGG) patients.

Detailed data were obtained for newly diagnosed LGG from The Cancer Genome Atlas (TCGA) and the Chinese Glioma Genome Atlas (CGGA) cohorts. Survival was assessed using Cox proportional hazards regression with adjustment for known prognostic factors. The model was established using the TCGA cohort, and independently validated using the CGGA cohort, to predict the 3-, 5-, and 10-year survival probabilities of patients.

Data from 293 patients with newly diagnosed LGG from the TCGA cohort were used to establish a prognostic model, and from 232 patients with primary LGG in the CGGA cohort to validate the model. Age, tumor grade, molecular subtype, tumor resection, and preoperative neurological deficits were included in the prediction model. The Cox regression model had a satisfactory corrected concordance index of 0.8508, 0.8510, and 0.8516 in the internal bootstrap validation at 3, 5, and 10 years, respectively. The calibration plots demonstrated high consistency of the predicted and observed outcomes. The CGGA cohort was used for external validation and showed satisfactory discrimination of 0.7776, 0.7682, and 0.7051 at 3, 5, and 10 years, respectively. The calibration plots demonstrated an acceptable calibration capability in the external validation.

This study established and validated a prognostic model to predict the survival probability of patients with newly diagnosed LGG. The model performed well in discrimination and calibration with ease of use, speed, accessibility, interpretability, and generalizability. An easily used nomogram based on the Cox model was established for clinical application. Moreover, a free, easy-to-use software interface based on the nomogram is provided online.