Bone Abstracts

Background: Dual energy X-ray absorptiometry measurement of lumbar spine bone mineral density helps diagnose osteoporosis and risk-stratify fracture risk. Vertebral anomalies can disproportionately impact T scores leading to diagnostic errors. The International Society for Clinical Densitometry (1) recommend excluding vertebra which look abnormal and are either deemed non-assessable within the system, or have a T-score differing by 1 standard deviation (SD) from adjacent vertebra(e). We audited scans with incorrect vertebral exclusions, identifying common errors to facilitate training as part of a new community-based DXA service rollout.

Methods: We reviewed a random sample of 29 scans where vertebral exclusion errors prompted re-analysis. Impact on diagnostic outcomes was assessed using T-score thresholds of -2.5 (osteoporosis), -3.5 (very high risk), and -4 (teriparatide funding).

Results: In 10/29 cases, vertebrae were excluded which did not fulfil exclusion criteria. 5/10 of these looked abnormal but the SD difference was 1 and they were considered assessable. In 3/10 cases, the incorrect vertebra(e) was excluded; all had lower T-scores, looked different from adjacent vertebrae which had higher BMD and were more sclerotic/degenerative. In 4/10 cases excluded vertebrae were deemed non-assessable due to previously reported fractures, but review of imaging revealed no significant fracture in 2/10 and/or fracture level discrepancy in 3/10. In 19/29 cases, vertebrae fulfilling the criteria were not excluded. 3/19 cases had confirmed significant vertebral fracture. 2/19 cases had overlying artifact. The remainder were degenerative. Overall, re-analysis resulted in a T-score change that crossed a diagnostic threshold in 14/29 cases. Feedback discussions with radiographers highlighted difficulties with SD subtractions involving negative numbers.

Conclusion: This study highlights the importance of excluding abnormal vertebrae correctly. Reasons for incorrect application of exclusion criteria included SD calculation issues, uncertainty regarding abnormal vertebra(e) and fracture misidentification. Following this, we introduced SD calculation guidance and recommended that vertebra(e) considered abnormal and/or with reported fractures are only excluded if they have a higher BMD/T-score and fulfil the 1SD criteria. If not, they are escalated for senior review, before being sent for reporting.

Reference: 1. International Society for Clinical Densitometry 2023 - Official Positions Adult