Which Cells Correspond To The Typical Signals For Fatty And Inflammatory Lesions Seen In Magnetic Resonance Imaging In Ankylosing Spondylitis ? -A Prospective Study Using Biopsy Material Obtained During Spinal Surgery-

Background The occurrence of bone marrow edema (BME) and fat metaplasia detected by magnetic resonance imaging (MRI) were shown to be significantly associated with syndesmophyte formation in patients with ankylosing spondylitis (AS). The cell type responsible for the fat signal seen in MRI has not been defined to date. Objectives To histologically analyze the cells seen in fatty lesions (FL) as detected by MRI in spinal biopsies of AS patients and compare them with controls. Methods Spinal biopsies from vertebral edges of patients with AS and controls who underwent surgery for spinal deformity or spinal stenosis were prospectively collected. All patients had spinal STIR- and T1-weighted MRIs available exactly from the area of biopsy, and all biopsies were taken from areas that had a fat signal on MRI. The biopsies were analysed blinded to patients9 diagnosis. Histomorphological analyses were performed to detect normal bone marrow, fat cells, inflammatory cells and fibroblasts. Histologic results were compared with MRI findings. Results Biopsies mostly obtained from the lower thoracic and the lumbar spine of 13 AS patients (mean age 56.3 years, mean disease duration 26 years) and 6 controls (mean age 53.4 years) were available. Similar proportions of AS patients, (12/16, 75%) and non-AS patients (4/6, 67%) had vital bone marrow. Fat cells were found in all 13 biopsies obtained from AS patients from the area of the fat signal vs. only 2 non-AS patients (33%), while inflammatory cells were found in 9 AS patients (56.3%), all of which also had BME on MRI, vs. 3 non-AS patients (50%). Fibroblasts were seen in 3 AS (18.9%) and 2 non-AS patients (33%). Conclusions The underlying cell types of FL and BME as detected by MRI in these long standing AS patients were fatty and inflammatory cells. The main difference between AS and non-AS patients was the proportion of biopsies containing fat cells. This suggests that fat cells are responsible for the MRI signal, at least in patients with longstanding ankylosing spondylitis. Disclosure of Interest None declared