Vacuum phenomenon: prevalence and appearance in the knee with 3 T magnetic resonance imaging

Abstract

To determine the prevalence of vacuum phenomenon (VP) in the knee on magnetic resonance (MR) images, describe the imaging features that characterize VP, and assess how often VP mimics pathological knee lesions. Consecutive knee MR studies performed on a 3 T MR system over a 9-month period were retrospectively reviewed by one radiologist who then selected studies with findings potentially indicating VP. Three experienced musculoskeletal radiologists reviewed these cases in consensus to confirm the presence of VP and to assess the shape, size, and signal of VP; the presence of magnetic susceptibility artifacts; and the ability of MR sequences to show VP. A total of 914 consecutive exams from 875 patients (524 men; mean age, 35 years) were reviewed. Vacuum phenomenon was found in 12 patients (prevalence 1.3%). In six (50%) patients, VP mimicked a meniscal tear, with four cases simulating a torn medial discoid meniscus. The VP signal was not easily differentiated from meniscal signal on most sequences in most cases (9/12). Gradient-recalled echo (GRE) localizer images proved most definitive, with 3D SPACE images the next most effective. Fast spin echo (FSE) images were only occasionally able to differentiate VP from meniscus. Rarely recognized on MR, VP can mimic meniscal pathology, potentially leading to inappropriate surgery. Because differentiation of VP from the meniscus is challenging on FSE at 3 T, radiologists should become familiar with the appearance of VP and review GRE localizer or 3D images carefully to avoid misinterpretation.