LGG

LGG predicts individualized survival probabilities for patients with lower-grade gliomas, including diffuse low-grade and intermediate-grade tumors classified as World Health Organization (WHO) grades II and III, using clinical and molecular variables.

Key Features:

• Development and validation: Developed on data from The Cancer Genome Atlas (TCGA) and externally validated on the Ohio Brain Tumor Study (OBTS).
• Variables considered: Incorporates tumor grade (II or III), age at diagnosis, sex, Karnofsky Performance Status (KPS), and molecular subtype including IDH mutant with 1p/19q codeletion (IDHmut-codel), IDH mutant without 1p/19q codeletion, and IDH wild-type.
• Predictive factors: Identifies Grade II tumor, younger age at diagnosis, high KPS, and IDHmut-codel molecular subtype as factors associated with increased survival.
• Modeling approaches: Employs Cox proportional hazards regression, random survival forests, and recursive partitioning analysis adjusted for known prognostic factors.
• Output: Produces individualized survival probability estimates presented as a nomogram.
• Validation procedures: Uses internal validation with 10-fold cross-validation and external validation assessed with calibration curves.

Scientific Applications:

• Individualized survival prediction: Provides patient-specific survival probability estimates for lower-grade glioma cohorts.
• Prognostic stratification: Enables stratification of patients by prognostic variables such as grade, age, KPS, and molecular subtype.
• Informing therapeutic decisions: Supplies prognostic information to inform evaluation of treatment strategies and clinical decision-making.

Methodology:

Model development used Cox proportional hazards regression, random survival forests, and recursive partitioning analysis adjusted for known prognostic factors; internal validation employed 10-fold cross-validation and external validation used calibration curves.