Atypical femoral shaft fractures (AFF) in NHS Grampian: incidence, underlying associations and assessment of DXA scanner software designed for early identification of AFF.
Year: 2021
Type: dissertation
Study design: Observational Study
Abstract: Objectives: The objectives of this study were to assess the clinical utility of extended femur DXA scan software in the identification of incomplete atypical femoral fractures in a routine clinical population; to assess the short term in-vivo precision of extended femur scans; and to investigate of the incidence of atypical femoral fractures between 2008 – 2018 within NHS Grampian, to provide a context for DXA scan findings. Method: A short term precision study was undertaken using the GE Lunar extended femur DXA scanning software, with 30 participants exposed to duplicate extended femur scans. From this, analysis was undertaken to assess and compare beaking index measurements and bone mineral density measurements from the software. Audits of scan acquisition, assessment and software stability were undertaken in order to assess the scan software and staff compliance with training, the service was routinely scanning the extended femur in all patients over 20 years of age. The images acquired were used to assess the utility of the software in identifying and measuring the endosteal femoral cortex, in contrast to the visual assessment of the same areas by experienced members of staff. A retrospective review of patients within NHS Grampian identified as having femoral fractures over a ten year period was carried out in order to identify the incidence of those suffering atypical femoral fractures within this healthcare service. Results: The visual assessment of the femoral cortex was found to be more effective and efficient than the scan software alone in identifying abnormalities, in line with the findings of previous studies comparing automated analysis with visual assessment. There was no identified difference in bone mineral density precision errors at the hip precision using the extended femur scanning software compared to standard hip measurements in the 30 patients included in the study, but there were some discrepancies in duplicated beaking index measurements in part thought to be caused by slight differences in positioning for scans. Least significant change was measured as 5.68% at femoral neck and 3.96% for total hip across the study, well within the parameters of ISCD accepted figures of 6.9% for femoral neck measurements and 5% for total hip. A negative predictive value of 100% was found when using the software in a clinical setting for six months, with a positive predictive value of 0.01% and an accuracy rate of 82.07%. Around 20% of patients scanned were found to have peaks ≥1mm on extended femur DXA scan automated analysis, however these were found to have an entirely normal appearance on visual inspection of images acquired. Audits of scan positioning, analysis and assessment of extended femur scans identified several positioning anomalies which were addressed when identified. Of the 7102 patients reviewed with femoral fractures over a period of ten years, 13 (0.18%) were identified as suffering atypical femoral fracture, with one of those patients having bilateral AFF. All the patients suffering AFF were also found to have varying lengths of bisphosphonate exposure. Discussion: Measurement of bone mineral density (BMD) by DXA is routinely used to diagnose osteoporosis and monitor treatment response. When comparing scans it is important to distinguish between real changes in BMD as opposed to changes related to the measurement process itself i.e. the precision of measurements. There is no published evidence of this type of beaking index precision study having been undertaken, indeed studies performed on extended femur scanning have all utilised slightly different methodology, making direct comparison impossible. Atypical femoral fracture is a rare but recognised complication of osteoporosis treatment and the use of extended femur scan software demonstrates a promising ability to identify and assess incomplete AFF in conjunction with routine bone mineral density measurement, identifying abnormal thickening or peaks in the lateral femoral cortex aided by automated software measurements of beaking index. However, in light of the study findings, visual assessment of the femoral cortex must also be undertaken by the operator to ensure false positives are eliminated from further investigation. It would be highly unlikely in the context of current literature that 20% of a general population would exhibit signs of iAFF on extended femur scanning. Conclusion: The extended femur scan acquisition and automated analysis was found to expose the patient population to a slightly higher radiation exposure. The automated analysis was not acceptable as a stand alone assessment of the femoral cortex, visual assessment was essential in tandem to ensure software anomalies were not reported as suspicious peaks, the investigation of which would place pressures on imaging services and also un-necessary anxiety to patients. Patient positioning as per the GE Lunar scan handbook is highlighted as important in terms of reproducibility of scans and also for accurately measuring the femoral cortex. For patients and clinicians there is reassurance that although such abnormalities are rare, affecting 0.18% of the local population, abnormalities will be identified by the software and highlighted by the reporting clinician, allowing monitoring and intervention as appropriate.
Source: Figshare
Author: D Smith
Institution: University of Exeter
Language: English
Subfield: Orthopedics and Sports Medicine
Field: Medicine
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W7055059492
