POS1110 RELIABILITY OF VERTEBRAL FRACTURE ASSESSMENT ON DUAL-ENERGY X-RAY ABSORPTIOMETRY

Background:Vertebral Fracture Assessment (VFA) is a new feature available on modern densitometers. Yet, the assessment of vertebral fracture (VF) status has not become standard practice.Objectives:Our study aimed to evaluate the reliability of VFA as assessed by a rheumatologist and a radiology technician.Methods:We conducted a cross-sectional study assessing the performance of low-radiation single energy x-ray absorptiometry VFA for the detection of VF. We selected patients who were assessed for osteoporosis according to screening protocols. Bone mineral densitometry was measured using standard methods over the lumbar spine L1-L4, the total proximal femur, and results were expressed as T-scores. All VFA were independently evaluated by 2 experienced readers: a rheumatologist and a radiology technician for the identification of VF (T4-L4). VF was classified according to the Genant grading system: grade 1 for an anterior, mid or posterior reduction of 20–25% in vertebral height; grade 2 for a reduction of 25– 40% and grade 3 for a reduction of more than 40% in vertebral height. A score for the inter-rater reliability between the readers was expressed using the kappa statistic.Results:One hundred patients were included with a mean age of 66.9 ± 9.5 years [46.7-83] years. There was a female predominance (91%). Nearly half of patients had osteopenia (48.9%), 27.7% had osteoporosis and 23.4% had a normal bone mineral density. On VFA scans, the non-visible vertebra was mostly located in the upper thoracic spine (60%). The mean number of VF was 1.2 [0-3] for both readers. According to the doctor’s evaluation, 25% of patients had at least one VF, of which 75.9% had a Genant grade 1, 17.2% had a Genant 2, and 6.9% had a VF grade 3. According to the technician evaluation, at least one VF was found in 36% of patients. A grade 1 was assessed in 91.7% of cases, a grade 2 in 8.3% of patients but no VF grade 3 was assessed. A kappa score for the inter-rater reliability between the readers for VFA was 0.545 (p=0.000). The overall agreement by grade between the readers was 0.785 (p=0,000). The exclusion of non-visible vertebra resulted in a better agreement (k=0.853). Further analysis excluding vertebra T4 to D10, revealed a very good agreement (k=0.9).Conclusion:Our study showed a low agreement between the readers on VFA and a better agreement when non-visible vertebrae were excluded. Thus, caution should be advocated when relying exclusively on this device.Disclosure of Interests:None declared.