Optimization of diagnostic performance for differentiation of recurrence from radiation necrosis in patients with metastatic brain tumors using tumor volume-corrected C-methionine uptake.

BACKGROUND

Tumor to normal tissue ratio (T/N ratio) on C-methionine (C-MET) positron emission tomography/computed tomography (PET/CT) is affected by variable factors. We investigated whether T/N ratio cutoff values corrected according to metabolic tumor volume (MTV) could improve the diagnostic performance of C-MET PET/CT for diagnosis of recurrence in patients with metastatic brain tumor. Forty-eight patients with metastatic brain tumors underwent C-MET PET/CT for differential diagnosis between recurrence and radiation necrosis after gamma knife radiosurgery (GKR). Both T/N ratio and MTV were estimated in each lesion on C-MET PET/CT. The lesions were classified into three groups based on MTV criteria (≤ 0.5 cm; > 0.5, ≤ 4.0 cm; and > 4.0 cm). The optimal cutoff values of the T/N ratio from receiver operating characteristic (ROC) curve were determined in each group (MTV-corrected) as well as total lesions (non-corrected). Finally, diagnostic performance of C-MET PET/CT was compared with the MTV-corrected cutoff values.

RESULTS

Among 77 lesions, 51 were diagnosed with recurrence. The mean T/N ratio was 2.25 (± 1.12) for recurrent lesions and 1.44 (± 0.22) for radiation necrosis (P < 0.001). T/N ratio of 1.61 (non-corrected) provided the best sensitivity, specificity, and diagnostic accuracy (70.6, 80.8, and 74.0%, respectively). Using the MTV criteria, optimal cutoff values of the T/N ratios in each group were 1.23 (MTV ≤ 0.5 cm), 1.54 (0.5 cm < MTV ≤ 4.0 cm), and 1.85 (MTV > 4.0 cm). In small-sized lesions (MTV ≤ 0.5 cm), MTV-corrected cutoff values (1.23) could maintain favorable diagnostic performance with sensitivity, specificity, and diagnostic accuracy (70.0, 80.0, and 73.3%, respectively), compared to non-corrected cutoff values.

CONCLUSIONS

MTV-corrected cutoff values of T/N ratio could maintain the diagnostic performance of C-MET PET/CT in small sized, metastatic brain tumors. We expect our results to contribute to reproducible and standardized interpretation of C-MET PET/CT.