In vitro assessment of incipient approximal carious lesions using computer-assisted densitometric image analysis.

Objectives: To investigate the use of computer-assisted densitometric image analysis (CADIA) to detect incipient approximal carious lesions in vitro and to compare areas of demineralization and mechanically induced lesions by radiographic visual inspection.

Methods: Thirty-six extracted non-carious human molars were demineralized in a defined approximal area of interest (AOI) using a gel system (pH 4.8) applied over 15 weeks. Standardized digital radiographs were taken before and after creation of caries-like lesions and grayscale values were measured within the AOI. For another group of eight teeth, artificial defects were mechanically induced by a diamond bur and densitometric analysis was performed. Statistical significance was tested using the Wilcoxon test and set at p < or = 0.05. In addition, radiographs were interpreted by visual examination by six experienced dentists. The sensitivity and specificity for both study groups were determined in comparison to a baseline group of untreated teeth. Polarized light microscopy and SEM analysis were used to evaluate changes in the subsurface mineralization pattern and surface micromorphology.

Results: Densitometric comparison of images from identical regions in control teeth revealed no significant grayscale value differences within the AOI, whereas a significant decrease (p < or = 0.05) was observed in the demineralized group. Mechanically induced lesions also showed a significant densitometric decrease (p < or = 0.05). The sensitivity of visual inspection of radiographs of demineralized areas was 0%. In contrast, the sensitivity for the detection of a mechanical defect of 0.1mm, which was the same subsurface extension as the histologically determined demineralization, increased to 52%. Specificity was 100% for all groups.

Conclusions: CADIA is able to detect very small changes in mineralization of incipient carious approximal lesions in vitro. The findings presented in this investigation may improve the subtraction method used for diagnosis and monitoring of approximal caries, as well as prophylactic measures.