AI-linked cytomics-on-a-chip to enable near-real time identification and management of oral lichenoid lesions in primary care

Oral cancer, which affects 54,000 Americans per year, has one of the worst outcomes and highest expense rates of all cancers, in large part due to late-stage diagnosis. Late-stage oral cancer has a 5-year survival rate of only 18%. Early detection leads to an improved 5-year survival rate (85%), less aggressive treatment regimens, and lower treatment cost burdens. Yet, only 28% of oral and oropharyngeal cancers are diagnosed at an early stage, mostly due to the lack of opportunistic screening. At any point in time, over 10% of the US population has an oral lesion, some of which will lead to cancer. For example, oral lichenoid conditions (OLC) have a 1.4-2.85% 5-year malignant transformation rate. But general practitioner dentists, periodontists otolaryngologists and oral and maxillofacial surgeons find it difficult to discriminate lesion risk based on visual and macroscopic appearance because benign, pre-cancerous, and malignant lesions often present with similar symptoms. While there is significant activity in the development of AI-linked diagnostic models for risk profiling oral lesions, the vast majority of these are not suitable for point-of-care (POC) use. There is a critical need for more accessible and accurate screening tools. To address this major gap in care, OraLiva proposes to develop a tool for AI-linked cytomics-on-a-chip to enable practitioners to identify and effectively manage OLC in primary care. In Phase I equivalent work, OraLiva validated C-AIIDE (Cytology Artificial Intelligence IDEntification), the first portable, programmable single cell cytology platform for POC use, for the ability to distinguish clinically significant oral lesions, including oral epithelial dysplasia and oral squamous cell carcinoma. A 4-site NIH-sponsored international prospective clinical study recorded over 200 cellular features related to biomarker expression, nuclear parameters, and cellular morphology per cell, encompassing ~2000 cells per patient, representing nearly 13 million indexed cells. This cytology platform has revealed 128 new features for better cellular discrimination, relative to established cytology predictors, among the diagnostic categories studied. For this Direct-to-Phase II project, OraLiva will advance a tool for front-line dentists/clinicians to assess lesions in primary care settings by undertaking five specific aims: 1) Design and produce alpha instrumentation units that combine all the major features and functionality integrated into stand-alone analyzers that are suitable for both research measurements and commercial launch of the initial oral lesion laboratory developed test (LDT); 2) Define optimal cytosignatures to differentiate between OLC+ and OLC- lesions; 3) Compile cytopathology database for OLC+ and OLC- lesions and develop new AI linked diagnostic models using the new cytosignatures and the new instrumentation; 4) Develop a clinically actionable intuitive OLC status report that can be accessed via cloud-based infrastructure; and 5) Define optimal beta product requirements for the future widely distributed regulated point of care platform. This program has potential to have major impact by identifying lesions that may be monitored in primary care when interventions are more impactful, less expensive, and less invasive. Project Number: 1R44DE033934-01 | Fiscal Year: 2025 | NIH Institute/Center: National Institute of Dental and Craniofacial Research (NIDCR) | Principal Investigator: Nicolaos Christodoulides (+1 co-PI) | Institution: ORALIVA, INC., NAPLES, FL | Award Amount: $1,178,476 | Activity Code: R44 | Study Section: Special Emphasis Panel[ZRG1 MSOS-D (10)] View on NIH RePORTER: https://reporter.nih.gov/project-details/10921587