Automatic enumeration of adherent streptococci or actinomyces on dental alloy by fluorescence image analysis

The aim of the present study was to develop an automated image analysis method to quantify adherence of Streptococcus sanguinis or Actinomyces viscosus on surfaces of a currently used dental alloy. Counting such bacterial strains was difficult because of their arrangement, thus S. sanguinis being a coccus arranged in chains or pairs, and A. viscosus a long complexly arranged polymorph rod. Direct counting of fluorescently stained adherent bacteria was done visually and with image analysis methods. To differentiate these two morphotypes, two programs were developed: (i) for streptococci, thresholding and selection of the object maxima, and (ii) for actinomyces, two step thresholding and processing of the characteristic points of the object skeletons. The triplicate enumerations for each bacterial strain were not significantly different (p > 0.005) and correlations between visual counting and automated counting were significant (r = 0.91 for S. sanguinis and r = 0.99 for A. viscosus, p <00.0001). These rapid and reproducible methods, allowed us to count either cocci or rods, adherent on an inert substratum, in high density conditions.