Clinical parameters outweigh diffusion- and perfusion-derived MRI parameters in predicting survival in newly diagnosed glioblastoma

Abstract

Background: The purpose of this study was to determine the relevance of clinical data, apparent diffusion coefficient (ADC), and relative cerebral blood volume (rCBV) from dynamic susceptibility contrast (DSC) perfusion and the volume transfer constant (k^trans) from dynamic contrast-enhanced (DCE) perfusion for predicting overall survival (OS) and progression-free survival (PFS) in newly diagnosed treatment-naïve glioblastoma patients.

Methods: Preoperative MR scans including standardized contrast-enhanced T1 (cT1), T2 - fluid-attenuated inversion recovery (FLAIR), ADC, DSC, and DCE of 125 patients with subsequent histopathologically confirmed glioblastoma were performed on a 3 Tesla MRI scanner. ADC, DSC, and DCE parameters were analyzed in semiautomatically segmented tumor volumes on contrast-enhanced (CE) cT1 and hyperintense signal changes on T2 FLAIR (ED). Univariate and multivariable Cox regression analyses including age, sex, extent of resection (EOR), and KPS were performed to assess the influence of each parameter on OS and PFS.

Results: Univariate Cox regression analysis demonstrated a significant association of age, KPS, and EOR with PFS and age, KPS, EOR, lower ADC, and higher rCBV with OS. Multivariable analysis showed independent significance of male sex, KPS, EOR, and increased rCBV_CE for PFS, and age, sex, KPS, and EOR for OS.

Conclusions: MRI parameters help to predict OS in a univariate Cox regression analysis, and increased rCBV_CE is associated with shorter PFS in the multivariable model. In summary, however, our findings suggest that the relevance of MRI parameters is outperformed by clinical parameters in a multivariable analysis, which limits their prognostic value for survival prediction at the time of initial diagnosis.

Keywords: MRI; diffusion; glioblastoma; perfusion; survival.

Fig. 1.

Semiautomatic segmentation of the volumes of interest in a 74-year-old female patient with a glioblastoma. A, contrast enhancement volume of interest on the cT1 image (red area); B, nonenhancing peritumoral edema volume of interest (blue area) on the FLAIR image; C, both volumes of interest displayed in 3D for the whole tumor. The segmentation masks were then transferred to D, apparent diffusion coefficient, from diffuse weighted imaging; E, relative cerebral blood flow from dynamic susceptibility contrast perfusion MRI, and F, k^trans map from dynamic contrast-enhanced perfusion MRI for the readout of histograms.