• diffusion
• perfusion
• progression
• response
• chemoradiotherapy
• glioblastoma

Background and purpose
For patients with glioblastomas the standard therapy is represented by maximal safe tumor resection followed by concurrent chemoradiotherapy with temozolomide (TMZ) and adjuvant TMZ. Determination of the response to therapy is entirely dependent on the interpretation of magnetic resonance (MR) imaging findings and clinical manifestations.
The purpose of this study was to assess the utility of MR advanced techiques, as diffusion (DWI) and perfusion (PWI) imaging in predicting response or progression of gliomas to chemoradiotherapy.
Materials and methods
We retrospectively selected 32 patients with high grade glioma treated by surgical resection followed by radiation therapy and temozolomide who underwent to conventional MR imaging and advanced MR techniques. The patients were divided into two groups, progression and response, on the basis of RANO criteria. ROI analysis was performed on pre and postcontrastographic T1-weighted images and on the Apparent Diffusion Coefficient (ADC) and PWI maps coregistrated, selecting the section with the maximum diameter of the enhancing lesion. In that section we calculated Contrast Enhancement, permeability (PSR), minimum value of ADC and maximum value of relative Cerebral Blood Volume (rCBV) before surgery (T0), after surgery before RT (T1) and after RT (T2). In conventional MRI we calculated the incremental ratio of the size of the pre and post operative tumor.
Unpaired t-test was used to test difference between groups for above mentioned variable for each time step. Significative data were utilized in the univariate logistic regressions in order to determine odds ratios for assessing risk factors. A multivariate logistic regression was performed on univariate significative variables.
Results
At unpaired t-test ADC, PSR and rCBV at T1 presented significant differences between the two groups. For the group progression the univariate logistic regressions with these variables, adjusted for gender and age , showed statistical significance for ADC and rCBV (predictors of progression) while in the subsequent multivariate logistic regression only ADC showed statistical significance. For the group response the univariate logistic regressions with these variables, adjusted for gender and age , showed statistical significance for ADC, rCBV and PSR (predictors of response) while in the subsequent multivariate logistic regression only PSR showed statistical significance.
Conclusions
MR advanced techniques, in particular diffusion with ADC and perfusion with PSR, have proved a valuable aid in predicting response or progression of glioma to chemoradiotherapy.