Proton magnetic resonance spectroscopy and diffusion-weighted imaging in intracranial cystic mass lesions

Background: The differential diagnosis of various intracranial cystic lesions is sometimes difficult on the basis of CT or MRI findings. Our objective was to evaluate H-1 MRS and DWI in the differential diagnosis of these lesions and in comparison with conventional MRI. Methods: Fifty patients with intracranial cystic lesions (21 pyogenic abscesses, 23 tumor cysts, 3 epidermoid cysts, and 3 arachnoid cysts) were evaluated with conventional MRI, DWI, and in vivo H-1 MRS. Preoperative diagnosis of the lesions was based on the results of DWI and in vivo MRS. All DWI and H-1 MRS studies were performed with a clinical 1.5-T system. The DWI was performed using single-shot spin-echo echo-planar pulse sequence with b = 1000 S/mm(2). The ADC value was measured. Diagnostic accuracy of conventional MRI, DWI, and in vivo H-1 MRS was calculated with respect to a final diagnosis of brain abscess vs nonabscess cystic tumor. Results: Lactate and cytosolic amino acids with/without succinate, acetate, and alanine were observed in 18 of 21 cases of abscesses on MRS. In 3 cases of epidermoid cysts, lactate was observed and could be differentiated from 3 cases of arachnoid cysts, which showed only minimal lactate. Only lactate was seen in 14 of 23 patients with tumor cysts, whereas both lipid/lactate and choline were visible in 9 patients with tumor cysts. Increased signal was seen in 20 of 21 patients with abscesses and 3 patients with epidermoid cysts on DWI. Decreased signal was observed in 22 of 23 patients with tumor cysts and 3 patients with arachnoid cyst on DWI. Diagnostic accuracy, sensitivity, specificity, positive predictive value, and negative predictive value of conventional MRI for the differentiation of brain abscess from nonabscess cystic tumor were 61.4%, 61.9%, 60.9%, 59.1%, and 63.6%, respectively, whereas they were 93.2%, 85.7%, 100%, 100%, and 88.5% with MRS; 95.5%, 95.2%, 95.7%, 95.2%, and 95.7% with DWI; and 97.7%, 95.2%, 100%, 100%, and 95.8% with MRS and DWI. Magnetic resonance imaging, when combined with in vivo MRS and DWI, accurately predicted the diagnosis in 47 (94%) of 50 and 48 (96%) of 50 of the cases, respectively. Conclusions: Proton MRS and DWI are useful as additional diagnostic modalities in differentiating intracranial cystic lesions. Combination of DWI with calculated ADC values and metabolite spectrum acquired by MRS add more information to MRI in the differentiation of intracranial cystic mass lesions. (c) 2007 Elsevier Inc. All rights reserved.