Objective: Accurate identification of primary central nervous system lymphoma (PCNSL) and its differentiation from other brain tumors remain difficult but are essential for treatment. In this study, we investigated whether (13)N-ammonia combined with (18)F-FDG could distinguish PCNSL from solid gliomas effectively.
Methods: Ten consecutive patients with final diagnosis of PCNSL (5 female and 5 male patients; mean [SD] age, 59.10 [12.47] years; range, 43-74 years) and another fifteen consecutive patients with solid glioma lesions (5 female and 10 male patients; mean [SD] age, 46.73 [19.61] years; range, 14-72 years) were included in this study. PET/CT imaging was performed for all of them with both (18)F-FDG and (13)N-ammonia as tracers. Tumor-to-gray matter (T/G) ratios were calculated for the evaluation of tumor uptake. Both Student t test and discriminant analysis were recruited to assess the differential efficacy of these 2 tracers.
Results: The T/G ratios of (18)F-FDG in PCNSL lesions were higher than in solid gliomas (3.26 [1.18] vs 1.56 [0.41], P < 0.001), whereas the T/G ratios of (13)N-ammonia in PCNSL lesions were lower than in solid gliomas significantly (1.38 [0.20] vs 2.11 [0.69], P < 0.001). All the lesions of PCNSL displayed higher T/G ratios of (18)F-FDG than (13)N-ammonia, whereas 14 (77.8%) of 18 glioma lesions showed contrary results. Tumor classification by means of canonical discriminant analysis yielded an overall accuracy of 96.9%, and only one glioma lesion was misclassified into the PCNSL group.
Conclusions: PCNSLs and solid gliomas have different metabolic profiles on N-ammonia and F-FDG imaging. The combination of these 2 tracers can distinguish these 2 clinical entities effectively and make an accurate prediction of PCNSL.