Purpose
Osteoporosis is very common, affecting over two million people in the UK [1]. It is thelargestrisk factor for fragility fractures, which are associated withhigh morbidity and mortality as well assignificant care costs for health systems.
Osteoporosis is usually identified using a fracture risk assessment tool, in association with Dual Energy X-ray Absorptiometry (DXA) for equivocal cases. However osteoporosis remains vastly underdiagnosed [2].
Changes in bone density can also be identified on standard radiographs. Twenty million radiographs were taken in the UK in 2016/17 [3] and...
Methods and materials
Full Research Ethics Committee approval was not required for this study. Local approvals were granted by the hospital Research and Development department under Governance Arrangements for Research Ethics Committees (GafREC) provisions (ref. GF0235).
Aconvenience sample of 150 individuals who had undergone both pelvic radiography and femoral neck DXA within a maximum interval of 6 months during 2016 was identified retrospectively.Patients were classified by DXA according to the World Health Organisation (WHO) definitions (figure 2).The sample was intentionally balanced to include equal proportions of individuals with...
Results
Baseline accuracy (the expected prediction accuracy achieved through random guessing) was calculated to be 33.3%.
When the predictive models generated by each algorithm were evaluated using the training dataset and ten-fold cross validation the best performing model was J48, with a prediction accuracy of 43.33% (table 1).
When J48 was evaluated using the independent test dataset its accuracy was 41.33% (8% improvement over baseline).
Conclusion
Alimitation of this study is that the radiographs were acquired using the x-ray systems of a single manufacturer. The test dataset also intentionally comprised equal proportions of each DXA classification, making it unrepresentative of any specific population. Use of a wider variety of x-ray systems and a more representative patient population would provide greater external validity.
The J48 algorithm demonstrated a modest improvement over baseline accuracy. The textural analysis and machine learning approach described shows some potential for the prediction of osteoporosis using routine pelvic...
Personal information and conflict of interest
A. Creeden; Coventry/UK - nothing to disclose
References
1. Mitchell, P., Dolan, L., Sahota, O., Cooper, A., Elliot, M., McQuillian, C., Stone, M., Hosking, D., Sandhu, B., Shervington, P., Moger, S. and Mullan, K. (2010). Osteoporosis in the UK at breaking point. London: International Longevity Centre – UK.
2.Marsh, D., Currie, C., Brown, P., Cooper, A., Elliott, J., Griffiths, R., Hertz, K., Johansen, A., McLellan, A. R., Mitchell, P., Parker, M., Sahota, O., Severn, A., Sutcliffe, A. and Wakeman, R. (2007) The care of patients with fragility fracture. London: British Orthopaedic Association.
3.NHS...