Evaluation of reproducibility in MRI quantitative volumetric assessment and its role in the prediction of overall survival and progression-free survival in glioblastoma.

ABSTRACT

BACKGROUND: Residual tumor volume (RTV) and extent of resection (EOR) have previously been shown to affect survival in glioblastoma (GB) patients. Quantitative radiological assessment (QRA) of these factors could potentially affect clinical decision-making in the postoperative period. PURPOSE: The first aim was to evaluate the reproducibility of different volume estimation methods of RTV and EOR by comparing QRA with subjective visual estimation and with objective volume estimations. The second aim was to clarify whether QRA of RTV and EOR would provide accuracy in predicting progression-free survival (PFS) and overall survival (OS) in GB patients. MATERIAL AND METHODS: Seventy GB patients were studied retrospectively. Reproducibility of QRA was compared to conventional visual analysis. Intra-rater agreement between two repeated measurements of 25 patients was calculated. QRA for RTV and EOR was made for the entire study population. Survival analysis was performed by multivariate cox-regression analysis. RESULTS: QRA of RTV and EOR gave superior intra-rater agreement compared to subjective evaluation. Multivariate survival analysis showed prognostic significance on 18 months PFS (hazard ratio [HR] = 0.44, P = 0.003) and OS (HR = 0.42, P = 0.012) at RTV < 1.6 mL and with EOR > 96% on PFS (HR = 2.152, P = 0.005) but not on OS (HR = 1.92, P = 0.053). CONCLUSION: QRA of tumor volumes is more robust compared to standard evaluation methods. Since EOR and RTV are correlated to the prognosis in GB, quantitative analysis of tumor volumes could aid decision-making and patient management postoperatively.