A. Stassinakis et al., ACCURACY IN DETECTING BONE-LESIONS IN-VITRO WITH CONVENTIONAL AND SUBTRACTED DIRECT DIGITAL IMAGING, Dento-maxillo-facial radiology, 24(4), 1995, pp. 232-237
Objective: To implement direct digital imaging (DDI) in subtraction ra
diography and compare the accuracy of conventional and subtracted DD i
mages in detecting small bone lesions in vitro. Method: Alveolar bone
defects were produced in a section of a pig mandible, with slow-speed
burs 0.6, 0.7, 0.8, 0.9, 1.0, 1.2, 1.4 mm in diameter. Standardized DD
images were subtracted and displayed in black and white, contrast-enh
anced and pseudo-colour transformed formats. 370 pairs of slides taken
directly from the computer monitor were evaluated by eight observers.
Results: The area P(A) under the ROC curve with DDI was 0.67+/-0.1. T
his was significantly lower (p<0.001) than any of the three modes of s
ubtraction radiography (mean P(A) = 0.88+/-0.09). The detection of sma
ll lesions (bur diameter 0.6 mm) was significantly better (p<0.001) wi
th contrast enhancement. Observer agreement was smaller for DDI (kappa
= 0.22+/-0.09) compared with the subtraction images (mean kappa 0.64/-0.13) (p<0.001). Conclusion: The diagnostic characteristics of the D
DI system were significantly improved by digital subtraction with imag
e processing.