Methods of estimating bone mineral density in digital radiography
word thesis.zip (177.8Mb)
Robertson, Gregory Aidan James
MetadataShow full item record
Assessment of bone mineral density (BMD) through Dual X-ray Absorptiometry (DXA) is a well established clinical technique. However, use of DXA is limited to the elective outpatient setting as a modality for detection of osteoporosis, and not available in the acute trauma setting prior to fracture fixation. This limits its use as a modality for estimating BMD prior to fracture fixation, and so limits its ability to influence choice of fixation materials. Given the limited resources of the current health system, such a technique of pre-operative BMD estimation would have to be performed from pre-operative plain radiographs or, more recently, pre-operative digital radiographs. Various measures have been suggested as indicators of BMD in plain radiographs, including: use of cortical measures, cortical ratios and summation of cortical measures; use of textural measures; and use of aluminium grading systems. Promising results have been reported with these measures in plain radiographs, however significant limitations exist with these techniques including variations in film quality and magnification, failure to account for the effects of soft tissue attenuation and scatter phenomenon, and inconsistent film processing techniques. With the introduction of digital imaging to clinical practice, it has been suggested that many of these limitations can been corrected for by the digital processing technique. As such, digital radiography provides clinicians with a potential tool to provide pre-operative BMD measures, allowing the potential to modify choice of fracture fixation materials accordingly. However limited research has been performed in this field to validate this technique. In this thesis the possibility of estimating BMD from digital radiographs by comparing various methods against results obtained from DXA scanning was investigated. When considering radiographs of the hip, cortical measures and cortical indices showed good correlations with hip DXA results, with the correlation being strengthened by summations of cortical measures. Textural measure analysis showed poor correlation with hip DXA results. Use of aluminium equivalent grading showed poor correlation with hip DXA results. When considering radiographs of the wrist, cortical measures and cortical ratios showed varying correlations with forearm and hip DXA results, ranging from poor to good. Summation of cortical measures failed to provide improved correlation values. Use of aluminium equivalent grading showed good correlation with forearm and hip DXA results. In conclusion, this thesis shows the potential for estimation of BMD from digital radiographs in the pre-operative setting. For the proximal femur, the summation of cortical measures provided the best estimation of bone density, whereas for the distal radius aluminium equivalent grading provided the best estimate. Further analysis is however required to establish if these techniques provide an adequate indicator of fixation strength in bone and so effectively guide pre-operative selection of fracture fixations materials.