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

Eric C Yang,David R Brenes,Imran S Vohra

doi:10.1117/1.jmi.7.5.054502

Eric C Yang, David R Brenes + Show 1 more

Open Access

https://doi.org/10.1117/1.jmi.7.5.054502

Copy DOI

Abstract

.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 , 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.

Highlights

In recent years, in vivo optical microscopy devices that visualize subcellular tissue features have been utilized to distinguish cancerous and precancerous tissue from benign tissue at a diverseJournal of Medical ImagingDownloaded From: https://www.spiedigitallibrary.org/journals/Journal-of-Medical-Imaging on 08 Nov 2021 Terms of Use: https://www.spiedigitallibrary.org/terms-of-useSep∕Oct 2020 Vol 7(5)Yang et al.: Algorithm to quantify nuclear features and confidence intervals for classification. . .range of anatomic sites.[1,2,3,4] Compared to traditional tissue biopsies and histopathologic analysis, these devices are noninvasive and provide real-time results
Reflectance confocal microscopy, confocal laser microendoscopy, and high-resolution microendoscopy (HRME) are examples of optical microscopy technologies that have been shown to visualize the morphology of the nuclei in the oral epithelium.[5,6,7,8]
Because abnormal nuclear morphology is a key hallmark of oral neoplasia, numerous clinical studies have demonstrated their potential to diagnose oral neoplasia.[9,10,11,12,13,14,15,16]

Summary

Introduction

Range of anatomic sites.[1,2,3,4] Compared to traditional tissue biopsies and histopathologic analysis, these devices are noninvasive and provide real-time results. Their potential application spans the spectrum of cancer care, from the diagnosis and surveillance of premalignant lesions, to margin delineation during surgical resections, and monitoring patients following treatment for recurrence. Studies have shown that a feature called the number of abnormal nuclei per mm[2], calculated from HRME images, can distinguish oral cancer and highgrade dysplasia from benign tissue with high accuracy (the ND algorithm).[19,20]

Methods

Results

Discussion

Conclusion