Introduction/Background World Health Organization classified the introduction of artificial intelligence (AI) in colposcopy as one of the main events that could reduce the incidence of cervical cancer worldwide. Cervical cancer incidence and mortality disproportionally affect low-middle income countries (LMICs) due to suboptimal organization of screening programs (Pimple, 2019, Minerva Gynecol.). Telecolposcopy and mobile screening programs can potentially lower cervical cancer incidence and mortality in LMICs (Hitt, 2016, Telemed J E Health). Furthermore, AI-guided digital colposcopy devices could strengthen diagnostic ability and accelerate cervical cancer elimination worldwide (Hue, 2020, BMC Med; Kim, 2022, Healthcare (Basel)). This study evaluates the effectiveness of AI mobile colposcopy in diagnosing high-grade cervical intraepithelial neoplasia (CIN 2,3) and cervical cancer.

Methodology This study included 28 female patients who underwent traditional colposcopy (Olympus OCS-500, analog video system), mobile colposcopy (MobileODT Eva) using Visualcheck AI application for image analysis, and multifocal targeted biopsy with histopathological examination.

Results According to traditional colposcopy and histopathological analysis, CIN 2 was detected in 8 cases, CIN 3 in 15 cases, and cervical cancer in 5 cases. Mobile colposcopy with artificial intelligence analysis detected 23 pathological cases (CIN 2 – 6, CIN 3 – 12, Cervical cancer – 5). Visualcheck AI sensitivity for CIN 2 is 75%, 80% for CIN 3, and 100% for cervical cancer. The overall positive predictive value (PPV) of Visualcheck AI was 82%.

Conclusion AI mobile colposcopy system showed high sensitivity in detecting high-grade CIN and cervical cancer. High PPV indicates reasonable diagnostic reliability of artificial intelligence in diagnosing high-grade CIN and cervical cancer; however, more robust studies are needed to produce conclusive results.

Disclosures No potential conflict of interest to report for all authors.