Purpose: After stereotactic radiation therapy for brain metastases, one of the complications is radionecrosis. Differential diagnosis with tumour recurrence can be helped by different methods of imaging, although histology remains the gold standard. According to the treatment centres, practice diverges. The objective of this single-centre retrospective study was to define the power of MRI, PET scan and NMR spectroscopy to establish a decision tree.
Material and methods: This study included patients who underwent stereotactic radiation therapy for brain metastases, and required, during follow-up, both a PET scan and NMR spectroscopy in order to differentiate a radiation necrosis and tumour recurrence. From 2010 to 2015, 25 patients were consistent with these criteria.
Results: Conventional MRI technique, with the T1/T2 mismatch criterion, had a specificity of 75% and a sensitivity of only 44%. A lesion quotient greater than 0.3 diagnosed a recurrence with a sensitivity of 92%. PET scan combined the best sensitivity and specificity, respectively of 92% and 69%. Whatever the thresholds used in NMR spectroscopy for choline/N-acetylaspartate and choline/creatin ratios, the power of this imaging modality did not exceed that of PET scan.
Conclusion: The criteria described in conventional MRI cannot by themselves establish the differential diagnosis. In first intention in case of doubt, PET scan should be done, combining the best sensitivity and specificity, whereas NMR spectroscopy used in combination does not improve these factors.
Keywords: Arbre décisionnel; Brain metastasis; Decision tree; IRM; Imagerie spectroscopique par résonance magnétique; MRI; Métastase cérébrale; NMR spectroscopy; PET scan; Radionecrosis; Radionécrose; TEP/tomodensitométrie.
Copyright © 2017 Société française de radiothérapie oncologique (SFRO). Published by Elsevier SAS. All rights reserved.