AI-Powered Radiology Reporting Delivers Results for Patients and Staff at Lewisham and Greenwich NHS Trust

AI-Powered Radiology Reporting Delivers Results for Patients and Staff at Lewisham and Greenwich NHS Trust  

Lewisham and Greenwich NHS Trust is leading the way in patient care and innovation with the successful pilot of an AI-powered radiology reporting app in the Emergency Department. 

Tackling Real Challenges in the ED 

Radiology departments across the NHS face mounting pressures: rising numbers of imaging studies, workforce shortages, and increasing demand for rapid, accurate diagnoses. These challenges can lead to delays and, in some cases, missed fractures, impacting patient outcomes and increasing costs. 

The Solution: AI App for Fracture Detection 

To address these issues, the Trust piloted the Radiobotics RBfracture AI app – on a technology platform deepc, which is integrated with the hospital’s imaging systems. This tool analyses X-rays in real time, providing preliminary reports that support radiologists and ED staff in making faster, more accurate decisions. 

Impressive Results 

Improved Diagnostic Accuracy: The AI tool achieved a 96.5% fracture detection rate, helping to reduce missed fractures and ensuring patients receive timely treatment. 

Faster Care, Better Outcomes: By speeding up image interpretation, the app has reduced ED breaches and length of stay, enabling quicker decisions and improved patient flow. 

Cost Saving: The pilot demonstrated potential annual savings of over £114,000, including fewer ED recalls, fewer unnecessary follow-up appointments, and significant cost avoidance from preventing missed diagnoses and litigation. 

Positive Staff and Patient Feedback: Over 80% of clinical users reported satisfaction with the AI tool, and patient outcomes improved thanks to fewer missed fractures and reduced repeat visits. 

Operational Productivity: Radiologists, ED clinical practitioners, and reporting radiographers have demonstrated increased confidence in reporting and in the early turnaround of AI output. Internal validation of improving high accuracy may pave the way for workflow modifications, with AI handling straightforward assessments modifying urgent care/ED workflows, freeing up time for ED and Radiology to focus on more complex studies and potentially saving thousands of hours each year.The project was aligned with the AI Centre – a consortium of organisations exploring how AI can impact healthcare outcomes. This project was in partnership with deepc AI radiology platform.  

A Model for the Future 

The pilot’s success has paved the way for a full business case and funding request across Southeast London, with plans to expand the use of AI across many Trusts. The project has been supported by robust governance, clinical safety assessments, and alignment with national guidance from NICE maintaining human oversight at all times! 

Lewisham and Greenwich NHS Trust Launches Robust AI Policy to Safeguard Clinical Safety 

Lewisham and Greenwich NHS Trust (LGT) had unveiled a comprehensive Artificial Intelligence Policy, setting a new benchmark for clinical safety and responsible AI adoption in healthcare. 

As AI technologies rapidly transform the medical landscape, LGT’s policy ensures that innovation is matched by rigorous governance, patient safety, and ethical standards. The policy, effective from March 2025, applies Trust-wide and covers all AI systems—whether developed internally or procured from external vendors and aligns with SEL’s AI framework and the Kings AI Centre. 

Putting Clinical Safety First 

At the heart of the policy is a clear mandate: AI must enhance patient care without compromising safety. The Trust’s AI Governance Board oversees all AI solutions, ensuring that every application is evaluated for clinical and operational impact before deployment. Only those approved by this board can progress to further review and funding stages.  

The policy requires that all AI systems used for clinical purposes comply with strict regulatory standards. Any AI classified as a Software as a Medical Device (SaMD) must be used only for its approved purpose. Using general-purpose AI tools, such as Microsoft Copilot, for clinical tasks like summarizing patient notes is strictly prohibited unless specifically approved. This approach mitigates medicolegal risks and protects both patients and staff.  

Rigorous Approval and Monitoring 

Before any AI solution is implemented, it must pass through a standardized process: 

• Initial Assessment: Identifying the clinical need and defining the problem. 
• Stakeholder Engagement: Consulting with clinicians, staff, and patients. 
• Compliance Checks: Ensuring alignment with GDPR, NHS standards, and regulatory approvals. 
• Pilot Testing: Evaluating performance and safety in real-world settings. 
• Training: Ensuring staff are fully prepared to use AI safely. 
• Continuous Monitoring: Ongoing evaluation of AI’s impact on patient care, with regular reporting to governance groups.  

A Culture of Safety and Transparency 

The Trust’s policy also emphasizes transparency and staff empowerment. All staff are encouraged to report concerns about AI safety, and incidents are reviewed by the Information Governance Team. Lessons learned are shared across the organisation to drive continuous improvement.  

Leading the Way in Ethical AI 

By embedding clinical safety at every stage—from procurement to deployment and monitoring—Lewisham and Greenwich NHS Trust is leading the way in ethical, patient-centered AI adoption. The policy not only protects patients but also builds trust among staff and the wider community, ensuring that AI is a force for good in healthcare.