Sensitivity and specificity of SWI venography for detection of cerebral venous alterations in multiple sclerosis

Abstract

Objectives: To determine the sensitivity and specificity of decreased venous vasculature visibility (VVV) on susceptibility-weighted imaging (SWI) venography in multiple sclerosis (MS) patients versus controls, and to compare this with assessment of whole brain atrophy.Methods: Forty MS patients and 22 controls without known central nervous system (CNS) disease who had non-specific white-matter (WM) lesions were imaged on a 3T GE scanner using SWI venography. Apparent total venous volume (ATVV) and increased average distance from vein (DFV) were calculated for various vein mean diameter categories: <0·3, 0·3–0·6, 0·6–0·9, and >0·9 mm. Principal component analysis (PCA) was used to identify potential discriminatory metrics. Receiver operating characteristics (ROC) of these metrics, along with normalized brain volume (NBV), were calculated to determine sensitivity and specificity values between the groups. The efficacy of the metrics was validated against blinded data from 14 MS patients and 8 controls who had non-specific WM lesions.Results: PCA identified 0·3–0·6 mm venous relative fraction (VRF) and DFV as useful metrics. ROC analysis results in initial sample of 40 MS patients and 22 controls were (sensitivity, specificity): 0·3–0·6 mm VRF (95·0%, 100·0%); DFV (100·0%, 100·0%); and NBV (82·5%, 68·2%). The results in validation sample were: 0·3–0·6 mm VRF (92·9%, 75·0%); DFV (100·0%, 100·0%); and NBV (78·6%, 75·0%).Discussion: Altered VVV indices on SWI venography showed high sensitivity and specificity for MS. The value of SWI venography for diagnosis of MS has to be further tested at early disease stages and against patients with other neurologic diseases.