Search

a b s t r a c t Background: Lesion size in fluid attenuation inversion recovery (FLAIR) images is an important clinical parameter for patient assessment and follow-up. Although manual delineation of lesion areas consid- ered as ground truth, it is time-consuming, highly user-dependent and difficult to perform in areas of indistinct borders. In this study, an automatic methodology for FLAIR lesion segmentation is proposed, and its application in patients with brain tumors undergoing therapy; and in patients following stroke is demonstrated. Materials and methods: FLAIR lesion segmentation was performed in 57 magnetic resonance imaging (MRI) data sets obtained from 44 patients: 28 patients with primary brain tumors; 5 patients with recurrent- progressive glioblastoma (rGB) who were scanned longitudinally during anti-angiogenic therapy (18 MRI scans); and 11 patients following ischemic stroke. Results: FLAIR lesion segmentation was obtained in all patients. When compared to manual delineation, a high visual similarity was observed, with an absolute relative volume difference of 16.80% and 20.96% and a volumetric overlap error of 24.87% and 27.50% obtained for two raters: accepted values for automatic methods. Quantitative measurements of the segmented lesion volumes were in line with qualitative radiological assessment in four patients who received anti-anogiogenic drugs. In stroke patients the proposed methodology enabled identification of the ischemic lesion and differentiation from other FLAIR hyperintense areas, such as pre-existing disease. Conclusion: This study proposed a replicable methodology for FLAIR lesion detection and quantification and for discrimination between lesion of interest and pre-existing disease. Results from this study show the wide clinical applications of this methodology in research and clinical practice. © 2013 Published by Elsevier Ireland Ltd.

Current methods for evaluation of treatment response in glioblastoma are inaccurate, limited and time-consuming. This study aimed to develop a multi-modal MRI automatic classification method to improve accuracy and efficiency of treatment response assessment in patients with recurrent glioblastoma (GB).A modification of the k-Nearest-Neighbors (kNN) classification method was developed and applied to 59 longitudinal MR data sets of 13 patients with recurrent GB undergoing bevacizumab (anti-angiogenic) therapy. Changes in the enhancing tumor volume were assessed using the proposed method and compared with Macdonald's criteria and with manual volumetric measurements. The edema-like area was further subclassified into peri- and non-peri-tumoral edema, using both the kNN method and an unsupervised method, to monitor longitudinal changes.Automatic classification using the modified kNN method was applicable in all scans, even when the tumors were infiltrative with unclear borders. The enhancing tumor volume obtained using the automatic method was highly correlated with manual measurements (N=33, r=0.96, p0.0001), while standard radiographic assessment based on Macdonald's criteria matched manual delineation and automatic results in only 68% of cases. A graded pattern of tumor infiltration within the edema-like area was revealed by both automatic methods, showing high agreement. All classification results were confirmed by a senior neuro-radiologist and validated using MR spectroscopy.This study emphasizes the important role of automatic tools based on a multi-modal view of the tissue in monitoring therapy response in patients with high grade gliomas specifically under anti-angiogenic therapy.

Background Treatment with bevacizumab is associated with substantial radiologic response in patients with glioblastoma (GB). However, following this initial response, changes in T 2 -weighted MRI signal may develop, suggesting an infiltrative pattern of tumor progression. The aim of this study was to differentiate between vasogenic-edema versus tumor-infiltrative area in GB patients. Methods and materials Fourteen patients with GB were longitudinally scanned, before and during intravenous bevacizumab therapy (5/10mg/kg every 2-weeks). A total of 40 MR scans including conventional, diffusion, dynamic susceptibility contrast, dynamic contrast enhancement imaging, and MR-spectroscopy (MRS) were analyzed. Classification of non-enhancing fluid-attenuation-inversion-recovery (FLAIR) area was performed based on mean diffusivity, cerebral blood volume and flow maps, and further characterized using multiple MRI parameters. Results The non-enhancing FLAIR lesion area was classified into: vasogenic-edema , characterized by reduced perfusion and increased FLAIR values; or tumor-infiltrative area, characterized by increased perfusion. Tumor-infiltrative area demonstrated a higher malignant pattern on MRS compared to areas of vasogenic-edema. Substantial reductions of the enhanced T 1 -weighted (58±10%) and hyperintense FLAIR (53±9%) lesion volumes were detected mainly during the first weeks of therapy, with a shift to an infiltrative pattern of tumor progression thereafter, as detected by an increase in tumor-infiltrative area in the majority of patients, which correlated with progression-free survival (week 8: r =−0.86, p =0.003, week 16: r =−0.99, p =0.001). Conclusion Characterization of non-enhancing hyperintense FLAIR lesion area in GB patients can provide an MR-based biomarker, indicating a shift to an infiltrative progression pattern, and may improve therapy response assessment in patients following bevacizumab therapy.