Search

Abstract Oral epithelial dysplasia (OED) is a premalignant histopathological diagnosis given to lesions of the oral cavity. Its grading suffers from significant inter-/intra-observer variability, and does not reliably predict malignancy progression, potentially leading to suboptimal treatment decisions. To address this, we developed an artificial intelligence (AI) algorithm, that assigns an Oral Malignant Transformation (OMT) risk score based on the Haematoxylin and Eosin (H&E) stained whole slide images (WSIs). Our AI pipeline leverages an in-house segmentation model to detect and segment both nuclei and epithelium. Subsequently, a shallow neural network utilises interpretable morphological and spatial features, emulating histological markers, to predict progression. We conducted internal cross-validation on our development cohort (Sheffield; n = 193 cases) and independent validation on two external cohorts (Birmingham and Belfast; n = 89 cases). On external validation, the proposed OMTscore achieved an AUROC = 0.75 (Recall = 0.92) in predicting OED progression, outperforming other grading systems (Binary: AUROC = 0.72, Recall = 0.85). Survival analyses showed the prognostic value of our OMTscore (C-index = 0.60, p = 0.02), compared to WHO (C-index = 0.64, p = 0.003) and binary grades (C-index = 0.65, p

This systematic review analyses and describes the application and diagnostic accuracy of Artificial Intelligence (AI) methods used for detection and grading of potentially malignant (pre-cancerous) and cancerous head and neck lesions using whole slide images (WSI) of human tissue slides. Electronic databases MEDLINE via OVID, Scopus and Web of Science were searched between October 2009 - April 2020. Tailored search-strings were developed using database-specific terms. Studies were selected using a strict inclusion criterion following PRISMA Guidelines. Risk of bias assessment was conducted using a tailored QUADAS-2 tool. Out of 315 records, 11 fulfilled the inclusion criteria. AI-based methods were employed for analysis of specific histological features for oral epithelial dysplasia (n = 1), oral submucous fibrosis (n = 5), oral squamous cell carcinoma (n = 4) and oropharyngeal squamous cell carcinoma (n = 1). A combination of heuristics, supervised and unsupervised learning methods were employed, including more than 10 different classification and segmentation techniques. Most studies used uni-centric datasets (range 40-270 images) comprising small sub-images within WSI with accuracy between 79 and 100%. This review provides early evidence to support the potential application of supervised machine learning methods as a diagnostic aid for some oral potentially malignant and malignant lesions; however, there is a paucity of evidence using AI for diagnosis of other head and neck pathologies. Overall, the quality of evidence is low, with most studies showing a high risk of bias which is likely to have overestimated accuracy rates. This review highlights the need for development of state-of-the-art deep learning techniques in future head and neck research., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2020 Elsevier Ltd. All rights reserved.)