New Methods for In-situ Quantitative measurement of Dentine Tubule Occlusion

Methods: Dentine samples (n=469), employed in an ethically approved in-situ clinical study involving 30 subjects (mean age 34.7 years) over four days, were used to investigate the level of tubule occlusion produced by two occlusion-based dentifrices containing 5% Novamin versus a control dentifrice containing NaF (1450ppm) or water. The samples were imaged pre and post treatment by Tandem Scanning Microscopy (TSM) and post treatment using Scanning Electron Microscopy (SEM). Tubule occlusion was determined using (i) a conventional approach of visually grading the amount of un-occluded dentine tubules on SEM images by three examiners using a visual ordinal scale (‘standard’) or (ii) computational software assessment of tubule patency with computer software specifically written using Image J (1.45s) to count the number of un-occluded dentine tubules.

Results: Statistically significant positive Spearman correlations existed between the ‘standard’ and the software designed for SEM (0.58) and TSM (0.42) (p<0.001) (n=469).  SEM visual grading and TSM computational analysis were both able to differentiate between the dentifrice’s formulated to produce occlusion of the dentine tubules and the control dentifrice or water (p<0.001). Intra-class correlation of the number of patent and un-occluded tubules counted visually and then by the software on thirty randomly chosen SEM or TSM images was 0.77 and 0.90 respectively.

Conclusions: This study has identified a novel computational-based image analysis routine capable of analysing dentine tubule occlusion quantitatively that may be used pre and post sample treatment. This routine allows for reproducible and objective assessment of dentine specimens.