OCAX (Oral CAncer eXplained): An Advanced AI-Based Approach for the Diagnosis and Interpretation of Oral Cancer

The OCAX project (Oral CAncer eXplained), funded under the PRIN PNRR 2022 program, aims to develop a cutting-edge system that uses Artificial Intelligence (AI) to support the early diagnosis and explanation of oral squamous cell carcinoma (OSCC). This disease, known for its high mortality and morbidity, can be managed more effectively through timely detection of precursor oral lesions.

Objectives

The OCAX project aims to significantly improve the early diagnosis and understanding of OSCC by integrating advanced technologies such as Artificial Intelligence (AI), Deep Learning (DL), and Case-Based Reasoning (CBR). The planned phases include:

• Data Collection and Annotation (DIAS Annotator): Creation of an extensive digital library containing clinical images and medical imaging data, such as magnetic resonance imaging (MR) and computed tomography (CT), systematically annotated by experts.
• Development of an Explainable AI System: Implementation of advanced methodologies for image classification and segmentation, enhanced by explainability features using tools like saliency maps and bounding boxes.
• Advanced Annotation Platform: Development of an intuitive web system to facilitate efficient data collection and annotation, supported by semi-automatic annotation algorithms.

These initiatives aim to make the diagnostic system not only accessible and reliable but also capable of improving the quality of diagnosis while reducing healthcare costs. OCAX represents an advancement in the application of AI in the medical field, promoting early diagnosis and informed clinical decisions.

The primary goal of OCAX is to simplify the identification of oral lesions at risk of becoming cancerous through the analysis of thousands of clinical images obtained from various devices. The platform, designed to be accessible to physicians in various clinical settings, aims to reduce diagnostic times and improve the quality of care.

A novel feature of OCAX is the ability to explain AI results through Case-Based Reasoning, enhancing the transparency and acceptance of AI tools among healthcare professionals. This aspect is crucial for improving the interaction between AI and clinicians, especially in complex contexts or less equipped centers.

Partnerships

The OCAX consortium includes prominent academic and clinical institutions:

• University of Salerno (UNISA): Responsible for the development of the Artificial Intelligence system and the annotation platform.
• University of Palermo (UNIPA) and University of Foggia (UNIFG): Involved in the collection and annotation of clinical images and diagnostic imaging data.
• External Collaborations: National medical centers and clinics that will contribute to the system validation and the collection of additional clinical cases.