Establishment of a computer-based system for initial approximal caries detection

Objectives: To establish a computer-based software system for better detection of initial approximal caries based on the performance of bite-wing radiography. Methods: With the basis of histologic caries diagnosis(microradiography)of some digitized bite-wing images, using MATLAB and Visual C++, we developed a computer-based software system that is capable of diagnosing approximal caries. Then a total of 190 approximal surfaces from 95 extracted posterior teeth were examined by bite-wing radiography .The proximal surfaces of these teeth were detected automaticly by this system ,and scored by 8 experienced dentists for the extent of lesions. The histological appearance of sectioned teeth was used as the validating criterion (gold standard) to assess sensitivity, specificity, accuracy , and reproducibility. Results: The sensitivity of this system was 0.728, better than the diagnostic ability of optic estimation at the level of enamel caries, and the reproducibility was good. The Pearson's correlation(0.722) is better than optic estimation(0.584)too. Conclusions: The diagnostic level of this system is better than optic estimation, it is a reasonably sensitive and accurate method for the detection of initial posterior approximal caries in present clinic practice, and a tool of providing objective data.