Endplate lesions in the lumbar spine: a novel MRI-based classification scheme and epidemiology in low back pain patients.

PURPOSE: The aims of the study were to introduce a classification scheme for endplate lesions based on T2-weighted magnetic resonance imaging and to detect possible associations between endplate lesions and other variables such as age, sex, disc degeneration and Modic changes in a large population. METHODS: MRI images of 996 low back pain patients were collected. All intervertebral spaces were classified as "normal", "wavy/irregular", "notched", "Schmorl's node" and "fracture". The associations between endplate lesions and age, sex, disc degeneration and Modic changes were determined in the considered population. RESULTS: The most common endplate lesions were "notched" and "Schmorl's nodes". The prevalence was higher among the male subjects. In most patients (62.8%), no endplate lesions were detected, with a significant difference between male (57.5%) and female subjects (67.9%) (p < 0.001). Lesions were found to be associated with intervertebral disc degeneration (relative risk 2.49) and signal alterations (relative risk 3.08). Fleiss kappas of 0.73 and 0.89 were, respectively, assessed for the inter- and intra-observer reliabilities of the new classification system. CONCLUSIONS: Endplate lesions were detected, classified with a novel scheme and analysed in a large population of patients suffering from low back pain based on MRI images. The reliability of the novel classification system was demonstrated. These slides can be retrieved under Electronic Supplementary Material.

The role of imaging specialists as authors of systematic reviews on diagnostic and interventional imaging and its impact on scientific quality: report from the EuroAIM Evidence-based Radiology Working Group.

PURPOSE: To evaluate the inclusion of radiologists or nuclear medicine physicians (imaging specialists) as authors of systematic reviews (SRs) on imaging and imaging-guided diagnostic procedures and to determine the impact of imaging specialists' presence as authors on the overall quality of the reviews. MATERIALS AND METHODS: A MEDLINE and EMBASE search was performed for SRs of diagnostic and interventional image-guided procedures that were published from January 2001 to December 2010. SRs about procedures primarily performed by nonimaging specialists were excluded. The inclusion of imaging specialists among the SR authors and the frequency of publication in imaging journals were evaluated. The quality of a subset of 200 SRs (100 most recent SRs with imaging specialists as authors and 100 most recent SRs without imaging specialists as authors) was rated by using a 12-item modified assessment of multiple SRs (AMSTAR) evaluation tool. Spearman, χ(2), and Mann-Whitney statistics were used. RESULTS: From among 3258 retrieved citations, 867 SRs were included in the study. Neuroimaging had the largest number of SRs (28% [241 of 867]), 41% (354 of 867) of SRs concerned diagnostic performance, and 26% (228 of 867) of SRs were published in imaging journals. Imaging specialists were authors (in any position) in 330 (38%) of 867 SRs; they were first authors of 176 SRs and last authors of 161 SRs. SRs with imaging specialists as authors were more often published in imaging journals than in nonimaging journals (54% [179 of 330] vs 9% [49 of 537]; P < .001). The median number of modified AMSTAR quality indicators was nine in SRs with imaging specialists as authors, while that in SRs without imaging specialists as authors was seven (P = .003). CONCLUSION: Only 38% (330 of 867) of SRs on radiology or nuclear medicine-related imaging published from January 2001 to December 2010 included imaging specialists as authors. However, the inclusion of imaging specialists as authors was associated with a significant increase in the scientific quality (as judged by using a modified AMSTAR scale) of the SR.