comparison study of F Choline PET CT with MRI in intracranial space occupying lesions and evaluation of diagnostic accuracy for differentiating neoplastic versus non neoplastic lesions

Objectives: To correlate the diagnostic performance of F-18 Choline PET/CT with MRI in ICSOLs for differentiating neoplastic versus non neoplastic lesions

Methods: Prospective analysis of ICSOL patients (n=37) fulfilling the inclusion/exclusion criteria was performed with MRI and F-18 Choline PET/CT and then followed with biopsy/histopathology/pooled data from adequate clinical follow up after the treatment of the lesions with either surgery, radiotherapy or chemotherapy alone or in combination or other forms of treatments. Two experienced nuclear medicine physicians who were blinded to the previous MRI and clinical data analyzed all studies independently and any abnormal tracer uptake within the corresponding lesion on MRI was interpreted as positive for neoplastic lesion. Histopathological examination and pooled data from adequate clinical or imaging follow up were taken as gold standard.

Results: 37 patients (mean age 37.8±17.1 years within age range of 8-72 years) selected for the final analysis underwent FCH PET/CT (within a mean time interval of 17±13.2 days from the previous MRI study). MR imaging was positive for neoplastic tumor in 30/37 (81.1%) patients with ICSOL evaluated with 100% true positives compared to the gold standard. In the 7/37 patients in whom MRI was negative, only two patients were false negative. Thus MRI yielded an estimated sensitivity of 93.8% (90 % CI) with 100% specificity, 100% positive predictive value (PPV) and 71.4% negative predictive value (NPV). There was significant agreement between the MRI results and the gold standard as given by the Kappa (k) value of <0.001. In the FCH PET/CT, 34/37 (91.9%) patients were reported positive for neoplastic tumor in which 2 were false positive. All the three patients negative on FCH PET/CT were true negatives in gold standard. The sensitivity, specificity, PPV and NPV of FCH PET/CT study were 100%, 60%, 94.1% and 100% respectively. A substantial measure of agreement was also noted between FCH PET/CT and the gold standard (kappa value 0.72; p <0.001). ROC curve was obtained for both FCH PET/CT and MRI. All the 30 patients with positive result for a neoplastic lesion in the MR finding showed similar result in the FCH PET/CT with concordant sensitivity. FCH PET/CT showed positive result in additional 4 patients with discordant results on MRI, out of which 2 had positivity in the gold standard. The two independent imaging modalities (FCH and MRI) yielded statistically significant agreement in the results (kappa value 0.55; p <0.001)

Conclusions: F-18 Choline PET/CT is a highly sensitive tracer but has a fairly low specificity in differentiating neoplastic versus non neoplastic lesions of the brain. MRI is a sensitive and a highly specific modality for evaluation of neoplastic versus non neoplastic lesions of the brain. F-Choline PET/CT and MRI are complementary imaging modalities each providing separate details of the ICSOLs critical in patient management. There is significant agreement between these two modalities as diagnostic tumor imaging modality/agents.