Agreement between radiographic and photographic trabecular patterns.

PURPOSE

It has been hypothesized that photographs can facilitate the interpretation of the radiographic characteristics of trabecular bone. The reliability of these photographic and radiographic approaches has been determined, as have various agreements between the two approaches and their correlations with biomechanical characteristics.

MATERIAL AND METHODS

Fourteen vertebral bodies were obtained at autopsy from 6 women and 8 men aged 22-76 years. Photographs (n = 28) and radiographs (n = 28) were taken of midsagittal slices from the third lumbar vertebra. The radiographs and photographs were digitized and the geometric properties of the trabecular architecture were then determined with a digital image analysis technique. Information on the compressive strength and ash density of the vertebral body was also available.

RESULTS

The geometric properties of both radiographs and photographs could be measured with a high degree of reliability (Cronbach's alpha > 0.85). Agreement between the radiographic and photographic approaches was mediocre as only the radiographic measurements showed significant correlations (p < 0.05) with the biomechanical characteristics. We suggest that optical phenomena may result in the insignificant correlations between the photographs and the biomechanical characteristics.

CONCLUSION

For digital image processing, radiography offers a superior description of the architecture of trabecular bone to that offered by photography.