Oral cancer is a growing health issue in a number of low- and middle-income countries (LMIC), particularly in South and Southeast Asia. The described dual-modality, dual-view, point-of-care oral cancer screening device, developed for high-risk populations in remote regions with limited infrastructure, implements autofluorescence imaging (AFI) and white light imaging (WLI) on a smartphone platform, enabling early detection of pre-cancerous and cancerous lesions in the oral cavity with the potential to reduce morbidity, mortality, and overall healthcare costs. Using a custom Android application, this device synchronizes external light-emitting diode (LED) illumination and image capture for AFI and WLI. Data is uploaded to a cloud server for diagnosis by a remote specialist through a web app, with the ability to transmit triage instructions back to the device and patient. Finally, with the on-site specialist’s diagnosis as the gold-standard, the remote specialist and a convolutional neural network (CNN) were able to classify 170 image pairs into ‘suspicious’ and ‘not suspicious’ with sensitivities, specificities, positive predictive values, and negative predictive values ranging from 81.25% to 94.94%.

Deep learning has become a powerful tool in solving image classification problems¹ and has been widely used in medical image analysis.² Convolutional neural networks (CNNs) show the potential for automated classification of different cancer lesions, such as lung cancer,³ skin cancer,⁴ and breast cancer.⁵ But the lack of interpretability and explainability makes CNNs less understandable.⁶ To push the envelope of CNN-based medical diagnostic methods, more rigorous neural network models and training methods must be developed to provide both visually interpretable and explainable analysis.^{7 – 9}

Non-invasive strategies that can identify oral malignant and dysplastic oral potentially-malignant lesions (OPML) are necessary in cancer screening and long-term surveillance. Optical coherence tomography (OCT) can be a rapid, real time and non-invasive imaging method for frequent patient surveillance. Here, we report the validation of a portable, robust OCT device in 232 patients (lesions: 347) in different clinical settings. The device deployed with algorithm-based automated diagnosis, showed efficacy in delineation of oral benign and normal (n = 151), OPML (n = 121), and malignant lesions (n = 75) in community and tertiary care settings. This study showed that OCT images analyzed by automated image processing algorithm could distinguish the dysplastic-OPML and malignant lesions with a sensitivity of 95% and 93%, respectively. Furthermore, we explored the ability of multiple (n = 14) artificial neural network (ANN) based feature extraction techniques for delineation high grade-OPML (moderate/severe dysplasia). The support vector machine (SVM) model built over ANN, delineated high-grade dysplasia with sensitivity of 83%, which in turn, can be employed to triage patients for tertiary care. The study provides evidence towards the utility of the robust and low-cost OCT instrument as a point-of-care device in resource-constrained settings and the potential clinical application of device in screening and surveillance of oral cancer.

Head and neck squamous cell carcinomas (HNSCC) includes multiple subsites that exhibit differential treatment outcome, which is in turn reflective of tumor stage/histopathology and molecular profile. This study hypothesized that the molecular profile is an accurate prognostic adjunct in patients triaged based on clinico-pathological characteristics. Towards this effect, publically available micro-array datasets (n = 8), were downloaded, classified based on HPV association (n = 83) and site (tongue n = 88; laryngopharynx n = 53; oropharynx n = 51) and re-analyzed (Genespring; v13.1). The significant genes were validated in respective cohorts in The Cancer Genome Atlas (TCGA) for correlation with clinico-pathological parameters/survival. The gene entities (n = 3258) identified from HPV based analysis, when validated in TCGA identified the subset specifically altered in HPV+ HNSCC (n = 63), with three genes showing survival impact (RPP25, NUDCD2, NOVA1). Site-specific meta-analysis identified respective differentials (tongue: 3508, laryngopharynx: 4893, oropharynx: 2386); validation in TCGA revealed markers with high incidence (altered in >10% of patients) in tongue (n = 331), laryngopharynx (n = 701) and oropharynx (n = 404). Assessment of these genes in clinical sub-cohorts of TCGA indicated that early stage tongue (MTFR1, C8ORF33, OTUD6B) and laryngeal cancers (TWISTNB, KLHL13 and UBE2Q1) were defined by distinct prognosticators. Similarly, correlation with perineural/angiolymophatic invasion, identified discrete marker panels with survival impact (tongue: NUDCD1, PRKC1; laryngopharynx: SLC4A1AP, PIK3CA, AP2M1). Alterations in ANO1, NUDCD1, PIK3CA defined survival in tongue cancer patients with nodal metastasis (node+ECS-), while EPS8 is a significant differential in node+ECS- laryngopharyngeal cancers. In oropharynx, wherein HPV is a major etiological factor, distinct prognosticators were identified in HPV+ (ECHDC2, HERC5, GGT6) and HPV- (GRB10, EMILIN1, FNDC1). Meta-analysis in combination with TCGA validation carried out in this study emphasized on the molecular heterogeneity inherent within HNSCC; the feasibility of leveraging this information for improving prognostic efficacy is also established. Subject to large scale clinical validation, the marker panel identified in this study can prove to be valuable prognostic adjuncts.