Abstract 2105: Characterization of oral carcinoma risk in patients with oral epithelial lesions employing an AI-assisted cytopathology-on-a-chip platform

Abstract Oral squamous cell carcinoma (OSCC) accounts for >350, 000 new cases and >177, 000 deaths worldwide. Oral potentially malignant disorders (OPMD), harboring oral epithelial dysplasia (OED) are 15 times more common than OSCC - a significant proportion transforming to malignancy. Furthermore, field cancerization may contribute to multifocal disease-risk comprising other oral sites. Oral epithelia lesions (OEL) are often challenging to diagnose for front-line clinicians and process of monitoring patients for new/evolving local disease lacks consensus - challenging opportunities for triage, surveillance and early detection. To address these gaps there is a compelling need to complete accurate surveillance testing of OEL for disease evolution in at-risk patients chairside/point-of-care, aiding clinical decision making. This study leverages an NIH-funded, multisite time-course investigation of patients with a history of OSCC or OED, utilizing AI-linked cytopathology capabilities derived from a microfluidics engine analyzed with a three-color fluorescence analyzer. Minimally invasive oral brush cytology specimens were collected from suspicious lesions and contralateral sites, with clinical and histopathological diagnoses blinded until patient exit. The study employs a chairside-ready, AI-assisted cytopathology-on-a-chip (COC) platform. This platform integrates microfluidics-facilitated immunoassays, cytomic measurements, automated analysis of cellular and molecular signatures, and result generation. The output includes a numerical index that profiles disease severity and progression, linked to the patient’s clinical and histopathological data, all delivered in approximately 40 minutes. Initial results from this study highlight the strong potential of a disease severity numerical index for assessing risk and distinguishing between oral cavity cancer patients who experience recurrence and those who do not. Similarly, OED patients at risk of malignant transformation exhibit higher numerical indexes. Moreover, contralateral lesions provide an advanced perspective on patients with a stronger predisposition to field cancerization. This noninvasive approach aligns well with findings obtained through more invasive biopsy and histological analysis. The AI-assisted chairside-ready COC minimally-invasive platform has strong potential to aid oral lesion characterization, for early-identification of disease risk and evolution. These capabilities include use of static numerical indexes and associated time-course changes. Measurement of contralateral-site specimen also offers new insights for referral or routine-surveillance, towards achieving better outcomes, disease prevention and management. Citation Format: Kritika Srinivasan Rajsri, Michael McRae, Nancy Ruel, A Ross Kerr, Nadarajah Vigneswaran, Adam Jacobson, Jonathan Shum, Rachelle Wolk, Nicolaos Christodoulides, John T. McDevitt. Characterization of oral carcinoma risk in patients with oral epithelial lesions employing an AI-assisted cytopathology-on-a-chip platform [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 2105.