A Comparison Study of Machine Learning (Random Survival Forest) and Classic Statistic (Cox Proportional Hazards) for Predicting Progression in High-Grade Glioma after Proton and Carbon Ion Radiotherapy

Abstract

Background: Machine learning (ML) algorithms are increasingly explored in glioma prognostication. Random survival forest (RSF) is a common ML approach in analyzing time-to-event survival data. However, it is controversial which method between RSF and traditional cornerstone method Cox proportional hazards (CPH) is better fitted. The purpose of this study was to compare RSF and CPH in predicting tumor progression of high-grade glioma (HGG) after particle beam radiotherapy (PBRT). Methods: The study enrolled 82 consecutive HGG patients who were treated with PBRT at Shanghai Proton and Heavy Ion Center between 6/2015 and 11/2019. The entire cohort was split into the training and testing set in an 80/20 ratio. Ten variables from patient-related, tumor-related and treatment-related information were utilized for developing CPH and RSF for predicting progression-free survival (PFS). The model performance was compared in concordance index (C-index) for discrimination (accuracy), brier score (BS) for calibration (precision) and variable importance for interpretability. Results: The CPH model demonstrated a better performance in terms of integrated Cindex (62.9%) and BS (0.159) compared to RSF model (C-index = 61.1%, BS = 0.174). In the context of variable importance, CPH model indicated that age (P = 0.024), WHO grade (P = 0.020), IDH gene (P = 0.019), and MGMT promoter status (P = 0.040) were significantly correlated with PFS in the univariate analysis; multivariate analysis showed that age (P = 0.041), surgical completeness (P = 0.084), IDH gene (P = 0.057), and MGMT promoter (P = 0.092) had a significant or trend toward the relation with PFS. RSF showed

Artificial intelligence and machine learning for medical imaging: A technology review

“…This goes in line with the findings from more fundamental ML research studies, which have reported RFs as one of the best classical learning algorithms [133]. However, many other works in the medical field have also compared the accuracy of RFs against more complex or simpler ML classifiers, and it is well known that their performance may vary for different applications [103,113,132,[134][135][136][137][138][139] and even for different datasets within the same application [131,132]. This makes it hard to conclude on the absolute superiority of RFs algorithm over other ML classifiers.…”

Survival Prediction in Gliomas: Current State and Novel Approaches

“…They found that prediction accuracy was significantly improved with the combined use of clinical and DVH features compared with the separate use of each feature (66). Qiu et al compared RSF and traditional CPH to predict tumor progression after particle beam radiotherapy in 82 HGG patients and found that CPH demonstrated a better performance in terms of integrated C-index as compared to the RSF model (18).…”

“…Efforts have been made to adapt the SVM model for time-to-event analysis to predict survival time and improve its performance on right censored data byKhan et al (40) and Van Belle et al(41) by integrating regression constraints.Random forestRandom forest (RF) is a non-parametric ML algorithm that constructs multiple decision trees based on training features and uses the consensus or average of their output to get a more accurate prediction. Similar to SVM, RF algorithm can be used to model a large number of predictors with a limited number of observations(18,42). For survival analysis, RF has been adapted by Ishwaran et al to create a Random Survival Forest (RSF) capable of time-to-event analysis, taking into account right censored data (42).…”

The Application and Comparison of Machine Learning Models for the Prediction of Breast Cancer Prognosis: Retrospective Cohort Study

“…Compared with the Cox proportional hazard regression, machine learning methods do not make any parametric or semiparametric assumptions and have the ability to detect and account for higher-order interactions as well as nonlinear relationships [8]. While there have been some attempts to use machine learning to build cancer prognosis prediction models [6,[9][10][11][12][13], currently, there is no consensus on whether traditional or machine learning-based prognostic prediction models have a better predictive performance.…”