Digital Microradiography: In Vitro Validation of a Novel Imaging Technique

Abstract

Microradiography has been widely used in mineralized tissue research in determining the mineral content within the specimens being studied. There are considerable limitations of this ageing gold standard such as unavailability of the high-resolution spectroscopic plates and prolonged imaging and processing times. The present study aimed at developing and validating a novel digital microradiographic technique that is not restricted by the limitations of conventional microradiography. Reproducibility of digital microradiography was investigated by studying 4 repeated images of 10 randomly selected sectioned implant-bone specimens acquired by 2 examiners over 2 weeks. The acquired images were analyzed by both manual and automated digital subtractions. Correlation between digital and conventional microradiography was performed by digital subtraction of 23 matching images acquired by both microradiographic techniques. A comparison between manual and automated digital subtraction enabled evaluation of the influence of the digital subtraction protocols on the results of the subtraction. A direct digital microradiographic technique has been developed which does not require analogue recording medium and film processing. The robustness of the digital microradiography was evidenced by the narrow range of means and standard deviations for intra- and inter-examiner reproducibility. The intra-examiner means and standard deviations ranged from 126.54-128.42 and 4.11-5.34 respectively. The inter-examiner means and standard deviations ranged from 126.71-129.87 and 4.68-5.70 respectively. The detection threshold for the digital microradiography was 5 gray scales or 3.9 percent, which was comparable to digital radiography. The high concordance between conventional and digital microradiography was demonstrated by the mean and standard deviation of 8.69 and 1.75 gray scales respectively. There was a statistically significant difference between the results obtained by manual and automated digital subtraction, but the clinical significance is yet to be determined.