Construction and Evaluation of a Prognostic Columnar Graphic Model for Adult Patients with Diffuse Midline Gliomas

To explore the prognostic factors of adult patients with diffuse midline glioma (DMG), and to further construct and evaluate prognostic columnar graphic models to provide some reference for the clinical management of this group of patients.

We included adult patients with histologically confirmed DMG from the SEER (Surveillance Epidemiology and End Results) database (2004-2015), dividing them into training and validation sets (7:3 ratio). Using Kaplan-Meier and Cox regression analyses, we determined independent prognostic factors for overall survival (OS) and cancer-specific survival (CSS). Prognostic column-line graphic models were developed for OS and CSS, incorporating patient demographics and clinical characteristics. The models underwent internal and external validation, with performance assessed using the Concordance Index, area under the curve values, and calibration plots.

The study encompassed 226 patients, showing age, tumor extension, and World Health Organization grades as significant prognostic factors. The constructed models for OS and CSS showed moderate reliability and predictive accuracy, with Concordance Index values of 0.786 (OS) and 0.79 (CSS) in the training set and 0.743 (OS) and 0.787 (CSS) in the validation set. Calibration plots and decision curve analysis confirmed the clinical usefulness of the models.

The column-line graphic prediction models for OS and CSS have moderately reliable predictive efficacy and help clinicians to better assess the prognosis and provide individualized treatment options for adults with DMG.