MON-803 Discrepancies in FRAX Scores calculated by Hologic Horizon DXA SystemTM

Abstract Disclosure: H.M. Babu: None. A.J. Spiro: None. N.O. Vietor: None. T.D. Hoang: None. M.K. Shakir: None. Background: Fracture risk assessment tool (FRAX) has been instrumental in calculating the 10-year probability of hip and major osteoporotic fractures using clinical risk factors and femoral neck bone density (BMD). This score is used to initiate antiresorptive treatment, typically in patients with risks of a major osteoporotic fracture ≥ 20% or hip fracture ≥ 3%. The FDA has cleared Hologic’s Horizon DXA SystemTM to incorporate the FRAX tool into its report. Trabecular Bone Score (TBS) assesses bone microarchitecture which can be used to calculate more accurate TBS adjusted FRAX scores. In our tertiary center, we noticed discrepancies between the FRAX scores reported by the HologicTM machines versus FRAX score calculated using the web-based tool. We also noticed differences in fracture risk prediction between unadjusted FRAX score and TBS adjusted FRAX score. We analyzed 150 DXA reports to find possible causes of these discrepancies. Methods: A retrospective analysis of 150 DXA reports was conducted using data from a previously approved research protocol including patients aged 40-90 years, weighing less than 275 lbs. Patients were classified into 3 categories, based on normal, osteopenic and osteoporotic BMD values. HologicTM only measures FRAX scores for patients that have BMD in the osteopenic range. FRAX scores were calculated by inputting data including patient risk factors from patient questionnaires and the calculated BMD into the FRAX web-based tool. We also collected data for TBS adjusted FRAX score. Results: A difference in FRAX scores was found in 42 out of 90 patients with osteopenia. In 12 patients, the difference was significant enough to alter management. All patients with discrepancy had one or more clinical risk factors present in the questionnaires. It was concluded that the differences in FRAX scores between HologicTM data and scores obtained from the web-based tool were due to the inaccurate entry of risk factors. Additionally, when the web-based tool calculated FRAX scores were compared to the TBS adjusted FRAX scores for osteopenic patients, in 1 patient the difference was significant enough to alter management. Conclusion: These findings highlight the importance of ensuring accurate and complete data entry into DXA machines. Training operators and implementing verification processes could mitigate these discrepancies and improve clinical decision making in osteoporosis management. This is critical as antiresorptive therapies can have potential adverse side effects. Furthermore, having a questionnaire with the same format as the HologicTM system can increase compliance and decrease inaccuracies with the entry of risk factors. We also recommend using the TBS adjusted FRAX scores, if available as it can change decision making for treatment of patients with borderline T-scores and unadjusted FRAX scores. Presentation: Monday, July 14, 2025