MRI assessment of relapsed glioblastoma during treatment with bevacizumab: volumetric measurement of enhanced and FLAIR lesions for evaluation of response and progression--a pilot study.

Purpose: To develop a magnetic resonance imaging (MRI) metric that is useful for therapy monitoring in patients with relapsed glioblastoma (GBM) during treatment with the antiangiogenic monoclonal antibody bevacizumab (Bev). We evaluated the feasibility of tumour volume measurement with our software tool in clinical routine and tried to establish reproducible and quantitative parameters for surveillance of patients on treatment with antiangiogenic drugs.

Materials And Methods: In this retrospective institutional pilot study, 18 patients (11 men, 7 women; mean age 53.5) with recurrent GBM received bevacizumab and irinotecan every two weeks as second line therapy. Follow up scans were assessed every two to four months. Data were collected on a 1.5 T MR System (Siemens, Symphony) with the standard head coil using our standardized tumour protocol. Volumetric measurement was performed with a commercial available software stroketool in FLAIR and T1-c imaging with following procedure: Pre-processing involved cutting noise and electing a Gaussian of 3 × 3 to smooth images, selecting a ROI (region of interest) in healthy brain area of the contra lateral side with quantifying the intensity value, adding 20% to this value to define the threshold level. Only values above this threshold are left corresponding to the tumour lesion. For the volumetric measurement the detected tumour area was circuited in all slices and finally summing up all values and multiplied by slice thickness to get the whole volume.

Results: With McDonalds criteria progression was indicated in 14 out of 18 patients. In contrast, volumetric measurement showed an increase of contrast enhancement of >25%, defined as threshold for progression, in 11 patients (78%) and in 12 patients (85%) in FLAIR volume, respectively. 6 patients revealed that volumes in MRI increased earlier than the last scan, which was primarily defined as the date of progression with McDonald criteria, changing PFS after re-evaluation of the tumour volumes from 6.8 to 5.6 months.

Conclusion: In this pilot study the applied imaging estimates objectively tumour response and progression compared to the bi-dimensional measurement. The quantitative parameters are reproducible and also applicable for the diffuse infiltrating lesions.