Dataset Information

Prognosis Individualized: Survival predictions for WHO grade II and III gliomas with a machine learning-based web application.

ABSTRACT: WHO grade II and III gliomas demonstrate diverse biological behaviors resulting in variable survival outcomes. In the context of glioma prognosis, machine learning (ML) approaches could facilitate the navigation through the maze of factors influencing survival, aiding clinicians in generating more precise and personalized survival predictions. Here we report the utilization of ML models in predicting survival at 12, 24, 36, and 60 months following grade II and III glioma diagnosis. From the National Cancer Database, we analyze 10,001 WHO grade II and 11,456 grade III cranial gliomas. Using the area under the receiver operating characteristic (AUROC) values, we deploy the top-performing models in a web application for individualized predictions. SHapley Additive exPlanations (SHAP) enhance the interpretability of the models. Top-performing predictive models are the ones built with LightGBM and Random Forest algorithms. For grade II gliomas, the models yield AUROC values ranging from 0.813 to 0.896 for predicting mortality across different timeframes, and for grade III gliomas, the models yield AUROCs ranging from 0.855 to 0.878. ML models provide individualized survival forecasts for grade II and III glioma patients across multiple clinically relevant time points. The user-friendly web application represents a pioneering digital tool to potentially integrate predictive analytics into neuro-oncology clinical practice, to empower prognostication and personalize clinical decision-making.

SUBMITTER: Karabacak M

PROVIDER: S-EPMC10603035 | biostudies-literature | 2023 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Prognosis Individualized: Survival predictions for WHO grade II and III gliomas with a machine learning-based web application.

Karabacak Mert M Jagtiani Pemla P Carrasquilla Alejandro A Germano Isabelle M IM Margetis Konstantinos K

NPJ digital medicine 20231026 1

WHO grade II and III gliomas demonstrate diverse biological behaviors resulting in variable survival outcomes. In the context of glioma prognosis, machine learning (ML) approaches could facilitate the navigation through the maze of factors influencing survival, aiding clinicians in generating more precise and personalized survival predictions. Here we report the utilization of ML models in predicting survival at 12, 24, 36, and 60 months following grade II and III glioma diagnosis. From the Nati ...[more]

PMID: 37884599

Similar Datasets

Project description:PurposeMolecular markers, such as FOXO1 fusion genes and TP53 and MYOD1 mutations, increasingly influence risk-stratified treatment selection for pediatric rhabdomyosarcoma (RMS). This study aims to integrate molecular and clinical data to produce individualized prognosis predictions that can further improve treatment selection.Patients and methodsClinical variables and somatic mutation data for 20 genes from 641 RMS patients in the United Kingdom and the United States were used to develop three Cox proportional hazard models for predicting event-free survival (EFS). The 'Baseline Clinical' (BC) model included treatment location, age, fusion status, and risk group. The 'Gene Enhanced 2' (GE2) model added TP53 and MYOD1 mutations to the BC predictors. The 'Gene Enhanced 6' (GE6) model further included NF1, MET, CDKN2A, and MYCN mutations, selected through LASSO regression. Model performance was assessed using likelihood ratio (LR) tests and optimism-adjusted, bootstrapped validation and calibration metrics.ResultsThe GE6 model demonstrated superior predictive performance, offering 39% more predictive information than the BC model (LR p<0.001) and 15% more than the GE2 model (LR p<0.001). The GE6 model achieved the highest discrimination with a C-index of 0.7087, a Nagalkerke R2 of 0.205, and appropriate calibration. Mutations in TP53, MYOD1, CDKN2A, MET, and MYCN were associated with higher hazards, while NF1 mutation correlated with lower hazard. Individual prognosis predictions varied between models in ways that may suggest different treatments for the same patient. For example, the 5-year EFS for a 10-year-old patient with high-risk, fusion-negative, NF1-positive disease was 50.0% (95% confidence interval: 39-64%) from BC but 76% (64-90%) from GE6.ConclusionIncorporating molecular markers into RMS prognosis models improves prognosis predictions. Individualized prognosis predictions may suggest alternative treatment regimens compared to traditional risk-classification schemas. Improved clinical variables and external validation are required prior to implementing these models into clinical practice.

Dataset Information

Prognosis Individualized: Survival predictions for WHO grade II and III gliomas with a machine learning-based web application.

Publications

Prognosis Individualized: Survival predictions for WHO grade II and III gliomas with a machine learning-based web application.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets