Mildly dysplastic oral lesions with optically-detectable abnormalities share genetic similarities with severely dysplastic lesions.

OBJECTIVE: Optical imaging studies of oral premalignant lesions have shown that optical markers, including loss of autofluorescence and altered morphology of epithelial cell nuclei, are predictive of high-grade pathology. While these optical markers are consistently positive in lesions with moderate/severe dysplasia or cancer, they are positive only in a subset of lesions with mild dysplasia. This study compared the gene expression profiles of lesions with mild dysplasia (stratified by optical marker status) to lesions with severe dysplasia and without dysplasia. MATERIALS AND METHODS: Forty oral lesions imaged in patients undergoing oral surgery were analyzed: nine without dysplasia, nine with severe dysplasia, and 22 with mild dysplasia. Samples were submitted for high throughput gene expression analysis. RESULTS: The analysis revealed 116 genes differentially expressed among sites without dysplasia and sites with severe dysplasia; 50 were correlated with an optical marker quantifying altered nuclear morphology. Ten of 11 sites with mild dysplasia and positive optical markers (91%) had gene expression similar to sites with severe dysplasia. Nine of 11 sites with mild dysplasia and negative optical markers (82%) had similar gene expression as sites without dysplasia. CONCLUSION: This study suggests that optical imaging may help identify patients with mild dysplasia who require more intensive clinical follow-up. If validated, this would represent a significant advance in patient care for patients with oral premalignant lesions.

Algorithm to quantify nuclear features and confidence intervals for classification of oral neoplasia from high-resolution optical images.

Purpose: In vivo optical imaging technologies like high-resolution microendoscopy (HRME) can image nuclei of the oral epithelium. In principle, automated algorithms can then calculate nuclear features to distinguish neoplastic from benign tissue. However, images frequently contain regions without visible nuclei, due to biological and technical factors, decreasing the data available to and accuracy of image analysis algorithms. Approach: We developed the nuclear density-confidence interval (ND-CI) algorithm to determine if an HRME image contains sufficient nuclei for classification, or if a better image is required. The algorithm uses a convolutional neural network to exclude image regions without visible nuclei. Then the remaining regions are used to estimate a confidence interval (CI) for the number of abnormal nuclei per mm 2 , a feature used by a previously developed algorithm (called the ND algorithm), to classify images as benign or neoplastic. The range of the CI determines whether the ND-CI algorithm can classify an image with confidence, and if so, the predicted category. The ND and ND-CI algorithm were compared by calculating their positive predictive value (PPV) and negative predictive value (NPV) on 82 oral biopsies with histopathologically confirmed diagnoses. Results: After excluding the images that could not be classified with confidence, the ND-CI algorithm had higher PPV (65% versus 59%) and NPV (78% versus 75%) than the ND algorithm. Conclusions: The ND-CI algorithm could improve the real-time classification of HRME images of the oral epithelium by informing the user if an improved image is required for diagnosis.