Interstitial bone marrow edema in zygapophyseal joints of the lumbar spine as detected by histological examination was present in all eight patients with AS to a different degree, varying from 10% to 60% of the total bone marrow area. There was a clear correlation between histopathologically confirmed interstitial edema and edema as shown by MRI. However, a small amount of histopathological interstitial edema in the bone marrow (less than 30% of the surface area) is not detected by MRI. There was only a poor correlation between histopathologically observed interstitial edema and inflammatory cell infiltration, and therefore only a poor correlation between cell infiltration and MRI edema, which might explain why some patients with AS can have active disease despite a normal MRI [16
]. An earlier study comparing AS histology from computed-tomography-guided biopsies from the sacroiliac joint with MRI observed some correlation between cell infiltration and bone marrow edema by MRI but did not investigate and compare histopathological and MRI bone marrow edema [9
We are aware that this analysis has its limitations: the number of patients in this study is small, all patients had advanced disease progression, the slices of MR images were 4 mm thick or were for example available only in sagittal sections (patients 7 and 8) and because of that it might be possible that inflammation and edema was missed. Furthermore, transverse images of the zygapophyseal joints might be more sensitive [8
]. However, the protocol used for MR images is a standard protocol for daily routine and should therefore be used for comparative analysis.
In patients with AS, spinal MRI is being used to assess spinal inflammation as an indicator of disease activity. Lesions of active inflammation are depicted as areas of increased signal intensity in T2-weighted images with fat saturation (STIR sequences) and most probably represent an increased water content, probably as a correlate of bone marrow edema [17
]. The term 'bone marrow edema' in MR images was first used by Wilson and colleagues in 1988. Regional decreased signal intensity of the bone marrow in T1-weighted images and increased signal intensity on T2-weighted images represented an accumulation of 'bone marrow water', which could be confirmed by biopsy. They defined such lesions as 'bone marrow edema' [18
]. A first systematic analysis of bone marrow edema in MR images and histopathological analysis was performed in osteoarthritis [15
] in which bone marrow edema could be observed in MR images in up to 50 to 68% of patients. In this study the bone marrow edema, defined as a hyperintense zone on STIR images and hypointense on T1-weighted MR images, consisted mainly of normal tissue (53% of the area was fatty marrow, 16% was intact trabeculae, and 2% was blood vessels) and in a smaller proportion of other changes, namely interstitial bone marrow edema 4% [15
]. Taken as a whole, the study by Zanetti and colleagues [15
] revealed non-characteristic histopathological abnormalities without increased infiltrations of mononuclear cells, increased microvessel density or interstitial bone marrow edema, clearly indicating that MRI-detected 'bone marrow edema' in patients with osteoarthritis can have various underlying histomorphological alterations. In our study the presence of large numbers of cellular infiltrates in all patients with a greater percentage of histopathologically confirmed interstitial edema argues strongly for the hypothesis that the bone marrow edema in MRI in our cohort of patients with AS was caused by inflammation.