Cervical Screening Decision Support Platform
This project develops a decision support platform for cervical screening, focusing on how AI-assisted imaging and data integration can support clinical workflows across cytology, HPV testing, colposcopy, and patient communication. The goal is to provide tools that help clinical teams review information more efficiently and consistently, while keeping clinicians in control of decisions and maintaining strong expectations around governance, security, and privacy. The current implementation is a research prototype intended for demonstration and experimentation, not for clinical use.
The Problem
Cervical cancer screening programmes place a high workload on cytology and colposcopy services, which must maintain quality and timely communication with patients even as staffing, training, and demand pressures increase.
Our Solution
Our work investigates how AI-enabled tools can support, rather than replace, clinical teams in cervical screening. We focus on representing screening knowledge and workflows in a structured way, so that software agents can coordinate information from images, lab results, clinical notes, and patient communications into a single, consistent view for clinicians. The emphasis is on workflow support, explainability, and governance rather than fully automated diagnosis.
What We Did
Our comprehensive approach to delivering this solution
Developed deep learning models using PyTorch for cell detection and region-of-interest highlighting
Built a FastAPI backend to orchestrate analysis pipelines
Created a web interface for clinicians to review images and AI-generated suggestions
Explored patterns for secure, privacy-conscious data handling and configuration
Containerised the system to support reproducible deployment in test environments
Ran technical evaluations on limited datasets to guide further research
Technology Stack
Technologies and frameworks powering this solution
Key Features
Core capabilities that make this solution powerful and effective
Automated cell and region-of-interest detection
Configurable analysis pipeline for cytology image processing
Interactive review interface for AI-generated suggestions
Basic reporting views to summarise analysis runs
Support for batch processing in test environments
Achievements
Measurable impact and recognition
Winner - Health Innovation Hub Ireland AI Call
Early research prototype and demo implementation
Ongoing experimentation with AI-assisted cervical screening workflows
Interested in Similar Solutions?
Let's discuss how we can help bring your vision to life with cutting-edge AI solutions tailored to your needs.