Fluorescent biomarkers are important indicators of disease, but imaging them can require specialized and often-expensive devices. Periodontal and dental diseases resulting from microbial plaque biofilms, if diagnosed early with biomarker images and expert knowledge, can be treated to prevent occurrences of serious systemic illnesses. We report two convolutional neural network classifiers trained with dentist annotations of disease signatures and fluorescent porphyrin biomarker images to identify dental plaque in white light images as a per-pixel binary classification task. The classifiers were trained and tested with millions of image patches from two datasets collected from 27 consenting adults using handheld intraoral cameras. The areas under the receiver operating characteristic curves for the test sets were calculated to be 0.7694 and 0.8720. Once trained, the classifiers predict the location of plaque in white light images without requiring specialized biomarker imaging devices or expert intervention. This generalized approach can be useful in other domains where diagnostic biomarker predicting can augment expert knowledge using standard white light images.
Why is this work important?
Biomarker imaging provides non-invasive indicators of disease and is used by human experts to augment disease diagnosis. Capturing biomarker images requires specialized and often expensive hardware, annotations, and analyses by experts, resulting in substantial diagnosis delays.
What has been done before?
Even when biomarker imaging is available, experts are often needed to interpret the resulting images. There is a rich literature on medical image segmentation, but many approaches—especially deep learning—require large amounts of images and operate on information from only a single given imaging modality.
What are our contributions?
We successfully learn assocations between images and union signatures of biomarker presence and expert disease annotations. By transforming the image-level segmentation problem into a region-based problem, we are able to learn from far fewer images than other approaches. We specifically test our approach on detecting the biomarker porphyrin and associated conditions in millions of image patches. Once trained, the classifiers predict the location of porphyrin in images without requiring specialized biomarker imaging devices or expert intervention.
What are the next steps?
We are developing processes incorporating numerous other biomarkers, conditions, and imaging modalities.