Abstract: Squamous cell cancer of the anus (SCCA) is increasing in the aging cohort of people living with HIV (PWH), who are at 40-80x higher risk for SCCA than the general population. The increase in invasive anal cancer, and its precursor, high-grade squamous intraepithelial lesions (HSIL), has led to a focus on prevention in PWH. The Anal Cancer HSIL Outcomes Research study demonstrated immediate anal HSIL treatment decreases invasive cancer incidence rates, providing the impetus for increasing screening capacity with anal colposcopy or high resolution anoscopy (HRA). Unfortunately, HRA-guided histologic diagnosis is resource intensive and inefficient, requiring extensive clinician training, separate patient visits for diagnosis and treatment, and biopsy- associated patient discomfort, an issue exacerbated by the poor specificity of HRA-directed biopsies. There is a great need to develop novel, low cost, point-of-care approaches that are responsive to global health care contexts and which offer an accurate, real-time “optical biopsy” thus transforming the current standard of practice by decreasing unnecessary biopsies, reducing loss of follow up, and facilitating a “see and treat” approach in low- and high-resource settings. In our current R01, we successfully optimized and validated a novel, low cost (<$2,500) Mobile High- Resolution Micro-Endoscope (HRME) which provides “optical biopsies” of HRA-identified abnormal areas by delineating the cellular and morphologic features of neoplasia. Our team also (a) produced a faster-frame capture (70 frames/sec) HRME, (b) developed automated, machine learning-based algorithms for image interpretation, and (c) showed that this fast and artificial intelligence (AI)-based (FAST) HRME had comparable sensitivity (91% vs 91%) and significantly improved specificity (87% vs 52%, p=0.0001) for HSIL detection, using gold-standard histology (AUC of 0.96). In this competitive renewal, we will develop computer-assisted AI-based identification of HRA-suspicious areas to enhance FAST-HRME to create a fully automated AI-based HRA-HRME platform and evaluate this in diverse clinical environments in USA and Brazil to facilitate implementation. Our hypothesis is this HRA-HRME technology will improve the efficiency, accuracy, and clinical impact particularly for novice HRA providers in the community by facilitating the identification of HRA-abnormal areas, providing real-time “optical biopsy” with HRME, and enabling real-time decision making (treat vs. biopsy vs. no biopsy) regardless of provider skill level. Finally, we will assess the barriers and facilitators for adoption and utilization of the AI HRA-HRME in diverse global health settings to facilitate utilization. Successful results can be applied to other cancers (oral, cervix, skin, stomach, etc.) requiring community-based screening and surveillance.