Diagnostic accuracy of 18F-FDG-PET/CT, contrast enhanced MRI and diffusion weighted MRI in the differentiating brain tumor recurrence from radiation necrosis

Objectives To compare the diagnostic accuracy of FDG-PET/CT, conventional contrast enhanced MRI (cMR) and diffusion weighted MRI with apparent diffusion coefficient images (DW) in differentiating brain tumor recurrence (TR) from radiation necrosis (RN) in postoperative patients treated with radiation therapy.

Methods We reviewed the PET/CT and MRI reports of 70 cases with brain tumors. Postoperative cases treated with radiation therapy having suspected TR versus RN were selected. TR was diagnosed as new areas with hypermetabolism on PET/CT, enhancement with/without mass effect on cMR, and restricted diffusion on DW. On each modality, suspected areas were graded as positive=TR, negative=RN, or uncertain. The final diagnosis of tumor recurrence was proved by pathology or all available follow-up.

Results Fifteen/70 cases were included in the study (Glioblastoma=5, oligoastrocytoma=3, metastasis and intracranial spread of head and neck tumor=2 each, astrocytoma, anaplastic meningioma, and ependymoma= 1 each). Eight/15 cases were TR (3 pathologically proven) and 7 cases were RN (mean follow-up 230 days). There was 1, 4 and 0 uncertain cases by PET/CT, cMR and DW respectively. DW was not available in 1 case. Excluding the uncertain cases, PET/CT, cMR, and DW correctly diagnosed 7, 8 and 3 cases as TR, and 6, 2 and 6 cases as RN, respectively. After excluding the uncertain cases in each modality, the diagnostic accuracy for PET/CT, cMR, and DW was 92%, 91%, and 64% respectively (table-1).

Conclusions We continue to acquire data; however preliminary results show that PET/CT has the highest accuracy of differentiating TR from RN (92%), compared to cMR (91%) and DW (64%). Certainty for making the diagnosis was also higher in PET/CT, compared to other modalities.