Brain metastasis in lung cancer. Comparison of cerebral MRI and 18F-FDG-PET/CT for diagnosis in the initial staging

FDG-PET/CT is increasingly used in staging of lung cancer as single "one stop shop" method. AIM, PATIENTS, METHODS: We prospectively included 104 neurological asymptomatic patients (65 years, 26% women) with primary diagnosis of lung cancer. In all patients PET/CT including cerebral imaging and cerebral MRI were performed. RESULTS: Diagnosis of brain metastases (BM) was made by PET/CT in 8 patients only (7.7%), by MRI in 22 (21.2%). In 80 patients both PET/CT and MRI showed no BM. In 6 patients (5.8%) BM were detectable on PET/CT as well as on MRI. Exclusive diagnosis of BM by MRI with negative finding on PET/CT was present in 16 patients (15.4%). 2 patients (1.9%) had findings typical for BM on PET/CT but were negative on MRI. With MRI overall 100 BM were detected, with PET/CT only 17 BM (p < 0.01). For the diagnosis of BM PET/CT showed a sensitivity of 27.3%, specificity of 97.6%, positive predictive value of 75% and negative predictive value of 83.3%. BM diameter on PET/CT and MRI were consistent in 43%, in 57% BM were measured larger on MRI. DISCUSSION: Compared to the gold standard of MRI for cerebral staging a considerable number of patients are falsely diagnosed as free from BM by PET/CT. MRI is more accurate than PET/CT for detecting multiple and smaller BM. CONCLUSION: In patients with a curative option MRI should be performed additionally to PET/CT for definitive exclusion of brain metastases