Abstract: |
Background The aim of this study was to assess the effectiveness of fluorine-18 fluorodeoxyglucose (FDG) PET-CT and dynamic contrast-enhanced (DCE) MRI in differentiating tumor progression and radiation injury in patients with indeterminate enhancing lesions after radiation therapy (RT) for brain malignancies. Methods Patients with indeterminate enhancing brain lesions on conventional MRI after RT underwent brain DCE-MRI and PET-CT in a prospective trial. Informed consent was obtained. Lesion outcomes were determined by histopathology and/or clinical and imaging follow-up. Metrics obtained included plasma volume (Vp) and volume transfer coefficient (Ktrans) from DCE-MRI, and maximum standardized uptake value (SUVmax) from PET-CT; lesion-to-normal brain ratios of all metrics were calculated. The Wilcoxon rank sum test and receiver operating characteristic analysis were performed. Results The study included 53 patients (29 treated for 29 gliomas and 24 treated for 26 brain metastases). Progression was determined in 38/55 (69%) indeterminate lesions and radiation injury in 17 (31%). Vpratio (VP lesion/VP normal brain, P <. 001), Ktrans ratio (P =. 002), and SUVratio (P =. 002) correlated significantly with diagnosis of progression versus radiation injury. Progressing lesions exhibited higher values of all 3 metrics compared with radiation injury. Vpratio had the highest accuracy in determining progression (area under the curve = 0.87), with 92% sensitivity and 77% specificity using the optimal, retrospectively determined threshold of 2.1. When Vpratio was combined with Ktrans ratio (optimal threshold 3.6), accuracy increased to 94%. Conclusions Vpratio was the most effective metric for distinguishing progression from radiation injury. Adding Ktrans ratio to Vpratio further improved accuracy. DCE-MRI is an effective imaging technique for evaluating nonspecific enhancing intracranial lesions after RT. © 2015 The Author(s) 2015. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. |