Related Experiment Video
Updated: Jun 30, 2026
05:33
Introduction of an Integrated Pathology Image Management, Artificial Intelligence, and Reporting System
Published on: July 11, 2025
CITOBOT AI for real-world cervical cancer screening using colposcopy imaging
Marcela Arrivillaga1, Daniela Neira2, David Steven Rivero3
1Department of Public Health and Epidemiology, Pontificia Universidad Javeriana, Cali, Colombia.
Frontiers in Public Health
|June 29, 2026
Summary
Artificial intelligence (AI) shows promise for cervical cancer screening. CITOBOT AI achieved high accuracy in identifying women needing further examination, potentially improving early detection in resource-limited settings.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Oncology
Background:
- Cervical cancer is a significant global health issue, especially in low- and middle-income countries.
- Delayed screening and diagnosis hinder early detection and treatment.
- AI in medical imaging offers potential to enhance screening accuracy, consistency, and timeliness.
Purpose of the Study:
- To evaluate the real-world screening performance of CITOBOT AI, an artificial intelligence system.
- To assess CITOBOT AI's accuracy using colposcopy imaging for cervical cancer screening.
- To validate the AI system in a public hospital setting in Colombia.
Main Methods:
- A cross-sectional study involving 650 women screened at Siloé Hospital, Cali, Colombia.
- Colposcopy-guided biopsy served as the reference standard.
- 2,648 cervical images were analyzed using AI, with performance evaluated by accuracy, sensitivity, specificity, and AUC.
Main Results:
- CITOBOT AI demonstrated high performance: 94.3% accuracy, 93.4% sensitivity, 94.9% specificity, and 0.98 AUC.
- Positive and negative predictive values were 92.9% and 96.2%, respectively.
- Performance metrics were consistent across validation subsets.
Conclusions:
- CITOBOT AI shows strong potential as a screening support tool for cervical cancer using colposcopy images.
- The AI system can help identify women requiring further colposcopy and biopsy.
- External validation in diverse datasets is necessary before widespread implementation.
