Perfusion MR prior to radiotherapy is a strong predictor of survival in high-grade gliomas after proton and carbon ion radiotherapy

Xianxin Qiu,Jing Gao,Jing Yang,Jiyi Hu,Weixu Hu,Xiaoyong Zhang,Jiade J. Lu,Lin Kong
DOI: https://doi.org/10.21037/atm-20-1646
IF: 3.616
2021-01-01
Annals of Translational Medicine
Abstract:Background: To assess the survival predictability of perfusion magnetic resonance imaging (MRI) by the normalized cerebral blood volume (nCBV) prior to particle beam radiotherapy (PBRT) in high-grade glioma (HGG) patients underwent particle therapy.Methods: The study retrieved dynamic susceptibility contrast MRI acquired prior to PBRT between 6/2015 and 3/2019 in 45 patients with HGG. Maximum nCBV (nCBVmax) within or adjacent to surgical/tumor bed was measured using 'hot-spot' method. The predictive values of nCBVmax for progression-free survival (PFS) and overall survival (OS) were assessed in univariate Kaplan-Meier curve and multivariate Cox proportional hazards (CPH) models. Nomograms based on CPH results were constructed to individualize the predicted probability of OS and PFS.Results: The Kaplan-Meier curves and all CPH models based on nCBVmax as continuous variable (nCBVmax-C), group by cut-off derived from median value and Youden-index method showed that nCBVmax prior to radiotherapy was a strong predictor for both PFS and OS in HGG patients who underwent PBRT. Nomograms built on CPH models showed similar excellent performance in both discrimination and calibration.Conclusions: Perfusion imaging prior to PBRT is a strong predictor of survival in HGG. Novel perfusion MR-based nomogram with prospective validation could potentially be formally used in future clinical practice to individualize survival probability.
oncology,medicine, research & experimental
What problem does this paper attempt to address?