Treatment response assessment in glioblastoma with GRASP DCE-MRI.

PURPOSE: Following radiochemotherapy, glioblastoma patients often develop new or enlarging contrast-enhancing lesions, posing a critical challenge in distinguishing treatment-related changes (TRC) from tumor progression (TP). This study investigates the diagnostic accuracy of Golden Angle Radial Sparse Parallel Dynamic Contrast-Enhanced (GRASP DCE) MRI parameters derived from the signal intensity curve over time to discriminate TRC from TP.

METHODS

Sixty-six glioblastoma patients, between 01/2017 and 12/2021, who underwent radiochemotherapy and developed new or enlarging contrast-enhancing lesions suspicious for TP, were analyzed. A diagnostic accuracy analysis using the area under the receiver operating characteristic curve (AUROC) assessed the effectiveness of GRASP DCE-MRI kinetic parameters in distinguishing TRC from TP.

RESULTS

138 DCE-MRI scans were classified into TRC (n = 63) vs. TP group (n = 75) according to the lesion outcome on serial MRI follow-up and the pathology results. The peak enhancement percentage (PE%), the initial area under the curve 30 s (iAUC 30) and 60 s (iAUC 60) were consistently lower in the TRC than in the TP group. Both PE% (Sensitivity 89 % (95 % CI [83, 96]), Specificity 87 % (95 % CI [78, 95])) and iAUC30 (Sensitivity 80 % [71, 88]), Specificity 81 % [71, 91]) showed very high accuracy in discriminating TRC from TP. The best threshold for PE% to differentiate both groups was 40.7 %.

CONCLUSION

Peak enhancement derived from GRASP DCE-MRI demonstrated the highest accuracy in distinguishing treatment-related changes from tumor progression in glioblastoma. This reproducible, easily interpretable metric may aid neuroradiologists in evaluating new or enlarging lesions post-radiochemotherapy, improving clinical decision-making.