SYsteMatical Trained learnIng aLgorithms for Oral carcInogenesiS Interpretation by Optical Coherence Tomography

Overview

This clinical trial aims to assess the efficacy of Optical Coherence Tomography (OCT) in the early diagnosis of oral cancer. It focuses on Oral Potentially Malignant Disorders (OPMDs) as precursors to Oral Squamous Cell Carcinoma (OSCC). Despite the availability of oral screening, diagnostic delays persist, underscoring the importance of exploring non-invasive methodologies. The OCT technology provides cross-sectional analysis of biological tissues, enabling a detailed evaluation of ultrastructural oral mucosal features. The trial aims to compare OCT preliminary evaluation with traditional histology, considered the gold standard in oral lesion diagnosing. It seeks to create a database of pathological OCT data, facilitating the non invasive identification of carcinogenic processes. The goal is to develop a diagnostic algorithm based on OCT, enhancing its ability to detect characteristic patterns such as the keratinized layer, squamous epithelium, basement membrane, and lamina propria in oral tissues affected by OPMDs and OSCC. Furthermore, the trial aims to implement Artificial Intelligence (AI) in OCT image analysis. The use of machine learning algorithms could contribute to a faster and more accurate assessment of images, aiding in early diagnosis. The trial aims to standardize the comparison between in vivo OCT images and histological analysis, adopting a site-specific approach in biopsies to improve correspondence between data collected by both methods. In summary, the trial not only evaluates OCT as a diagnostic tool but also aims to integrate AI to develop a standardized approach that enhances the accuracy of oral cancer diagnosis, providing a significant contribution to clinical practice.

Background and needs: Despite advancements in oral screening techniques, diagnostic delays persist, necessitating the exploration of non-invasive methodologies for early detection of oral cancer. The current standard diagnostic method, histological analysis, often requires invasive biopsies and can be time-consuming, leading to delays in treatment initiation. Moreover, traditional screening methods may not always detect early-stage oral lesions accurately. Therefore, there is a critical need to enhance diagnostic capabilities through the adoption of innovative technologies. In this context, Optical Coherence Tomography (OCT) emerges as a promising technology warranting investigation. OCT offers several advantages over conventional diagnostic approaches. Its non-invasive nature allows for real-time and non-invasive imaging of tissue morphology with high resolution, enabling clinicians to visualize structural changes in oral tissues. By providing cross-sectional images of tissue layers, OCT has the potential to identify subtle alterations indicative of early-stage oral lesions, including potentially malignant disorders (OPMDs) and squamous cell carcinoma (OSCC). Additionally, OCT can facilitate early detection by enabling repeated examinations over time, thereby monitoring lesion progression or regression without the need for repeated biopsies. The exploration of OCT as a diagnostic tool aligns with the urgent need to improve the efficiency and accuracy of oral cancer diagnosis. By leveraging the capabilities of OCT, clinicians can potentially expedite the identification of suspicious lesions, leading to timely intervention and improved patient outcomes. Moreover, the integration of OCT into routine clinical practice has the potential to reduce the burden associated with invasive procedures and diagnostic delays, ultimately enhancing the quality of care for individuals at risk of oral cancer. However, despite these potential benefits, several challenges remain. Currently, there is a lack of precise definition of OCT patterns specific to various oral lesions. This hinders the consistent interpretation of OCT images and limits its diagnostic utility. Additionally, the accurate alignment of OCT findings with histological analysis is essential for validation and clinical applicability. Yet, there is still a need for standardized protocols to ensure proper overlay of OCT images with corresponding histopathological features. Furthermore, while computerized OCT analysis holds promise for enhancing diagnostic accuracy, existing methodologies may be prone to biases. These biases must be addressed to develop robust algorithms capable of reliably detecting early signs of oral cancer, trained on standardized techniques of comparison between OCT and histology. Therefore, addressing these challenges through the standardization of OCT imaging protocols, the establishment of consistent OCT patterns, and the development of unbiased computerized analysis methods is imperative. Doing so will not only advance the clinical utility of OCT in oral cancer diagnosis but also improve patient outcomes by enabling earlier detection and intervention. Aims and approach: 1. Standardization of technique for OCT scans and biopsy of oral lesions: * Objective: To standardize the biopsy acquisition technique for both OCT and histological analysis, ensuring a reliable correlation between imaging modalities. * Approach: We will develop and implement a standardized biopsy acquisition protocol, optimizing tissue preservation and alignment with OCT imaging parameters. This may involve specialized instrumentation and procedural guidelines tailored to maximize diagnostic yield, focusing on standardization of site and dimension of optical and surgical sampling. Detailed protocols will be established for OCT imaging, ensuring consistent acquisition parameters across all sites. Similarly, histological processing of biopsy specimens will adhere to standardized protocols to maintain integrity and facilitate accurate correlation with OCT findings. A novel optical and histological procedure of Target biopsy will be performed and assessed. 2. Standardization of OCT patterns of oral carcinogenesis: * Objective: To establish standardized patterns for OCT imaging of OPMDs and OSCCs, enhancing diagnostic accuracy. * Approach: The evaluation of OCT images will entail meticulous analysis to identify consistent patterns reflective of various oral lesions. By correlating these patterns with histological findings, we aim to develop a comprehensive reference guide for interpreting OCT images with precision and consistency. 3. Creation of Image Dataset for the Development of Diagnostic Software: * Objective: To collect a comprehensive repository of OCT images, facilitating the development of digital diagnostic tools. * Approach: A robust dataset comprising OCT images and corresponding histological data will be meticulously curated. This dataset will serve as the foundation for training and validating machine learning algorithms aimed at developing sophisticated diagnostic software capable of detecting early signs of oral cancer with high sensitivity and specificity. By pursuing these objectives, we aim to not only evaluate the efficacy of OCT in early oral cancer diagnosis but also contribute to the standardization of diagnostic methodologies and pave the way for the integration of advanced technologies into clinical practice.