Morphological segmentation and fractal analysis for the classification of colon polyps from en face optical coherence tomography (OCT) images

Colorectal cancer (CRC) is one of the top causes of malignancy in both men and women. Although screening has significantly reduced CRC mortality, colonoscopy suffers from inadequate inspection and sampling of the tissue, a limitation that can be addressed by Optical Coherence Tomography (OCT). The use of fractal analysis, which has been shown to estimate the scatterer size from OCT data, could help improve the classification of colon polyps as compared to using morphological information alone. For this study, thirty polyps were imaged immediately post excision, histologically processed, and both the OCT and H&E images annotated by an expert. Multi-step segmentation was used to segment the crypt regions. Fractal analysis of those areas was employed to estimate the nuclear size and classify the polyps as normal or tubular adenoma. This process resulted in an accuracy of 81% (92% sensitivity, 40% specificity) as confirmed by histology. The poor specificity can be partially attributed to the small number (only 5) of normal polyps and, more importantly, to the confirmed presence of other histological features which might influence the fractal analysis. The proposed approach must be expanded to include more polyps and further enhanced to improve its specificity. However, these preliminary results provide evidence that this method has the potential to perform scatterer size estimation and tissue classification from en face images, thus, providing a robust approach for the improvement of the accuracy of endoscopic OCT and, in the future, the effectiveness of colonoscopy.