Is bone marrow edema syndrome a precursor of hip or knee osteonecrosis? Results of 49 patients and review of the literature.

PURPOSE: Diagnosis of bone marrow edema syndrome (BMES) can be challenging. There is sometimes uncertainty about the correct diagnosis of BMES on morphologic magnetic resonance imaging (MRI), since subchondral findings like lines and spots can be misinterpreted as "beginning" or "possible" avascular osteonecrosis (AVN). The aim of our study was to systematically assess the temporal course of BMES from first diagnosis on MRI until the end of clinical symptoms and the full disappearance of bone marrow edema (BME) to determine whether subchondral lines and spots detected in these patients can develop into osteonecrosis.
METHODS: In a combined retrospective and prospective study, we retrieved serial MRI scans of hips and knees with BME from the hospital database. According to clinical and imaging data, all patients with degenerative, infectious/inflammatory, rheumatic, neoplastic conditions and those showing typical osteonecrosis were excluded. We collected all available MRI examinations from first detection of BME until its disappearance. In case edema had not fully resolved in the last available MRI scan, we performed an MRI with an additional dynamic contrast-enhanced (DCE-MRI) sequence. For each MRI scan, we recorded the severity of edema, the presence of subchondral hypointense lines and the presence of subchondral focal hypointense zones on T1-weighted images by two independent readers. The DCE-MRI scans were used to calculate parameter maps to assess the perfusion characteristics.
RESULTS: The study comprised 49 patients aged 22-71 years. In total, 171 morphologic and 5 DCE-MRI scans were evaluated. In 44 patients (89.8%), the BMES completely healed without remnants. In 18 of 49 patients (36.7%), a subchondral line was present in the first MRI exam. Nine patients (18.4%) developed a subchondral line within 1-5 months after the first MRI. In total, 27 out of 49 patients (55.1%) had subchondral lines (12 knees, 15 hips) during the timeframe of the study. All subchondral lines disappeared in the timeframe of the study. Subchondral focal hypointense zones were present in 14 out of 49 patients (28.6%): in 9 cases, subchondral focal hypointense zones disappeared after a median of 5.5 months (range, 1-85 months), while in 5 cases, subchondral focal lesions persisted until the end of the study (up to more than 85 months) without edema in the surrounding bone. All persisting subchondral focal lesions were hyperperfused. These 5 patients had associated meniscal lesions.
CONCLUSION: Our study shows that subchondral lines and spots found in patients with BMES do not develop into AVN. Subchondral lines, which resemble subchondral insufficiency fractures, are associated with BMES. Subchondral focal T1-hypointense zones do not represent AVN; most probably these areas represent reparative processes within the subchondral bone, where tensile and shear force overload is present due to altered biomechanics.