Radiotherapy plays a crucial role in cancer treatment, with nearly half of all cancer patients undergoing this life-saving intervention. Behind every successful treatment plan lies a meticulous and time-consuming process: clinicians manually contour tumors and organs on CT images to ensure treatment precision while protecting healthy tissue. This task demands deep expertise and often delays treatment, affecting both clinician efficiency and patient outcomes.
For clinicians, the repetitive nature of these tasks contributes to mounting workloads and, over time, burnout. For patients, every delay in treatment could mean waiting longer for potentially life-saving care.
To tackle these challenges, Guy’s and St Thomas’ Hospital, in partnership with King’s College London and deepc, is in the process of integrating AI into their radiotherapy workflows—the result: faster, more efficient treatment planning that enhances both clinician focus and patient care.
If you're interested in the broader journey of implementing AI in healthcare, you might also enjoy our blog, Navigating the AI Development Journey in Healthcare.
The Challenge: Streamlining Radiotherapy Without Compromising Safety
Manual segmentation of CT images is one of the most resource-intensive aspects of radiotherapy. Clinicians spend hours outlining tumors and protecting healthy tissue, a process demanding absolute precision. While essential for patient care, the repetitiveness of this task can lead to fatigue, and the time invested often means that patients must wait longer for treatment.
Integrating AI into this process is not just about saving time, it is about improving patient care outcomes while maintaining trust among clinicians. With these priorities at the forefront, it is equally critical to meet strict regulatory standards, ensure data security, and scale AI solutions across different hospital systems without disrupting existing workflows.
The Solution: Bridging Research and Clinical Practice with deepc
The Clinical Scientific Computing (CSC) Team at Guy’s and St Thomas’, led by Anil Mistry, has partnered with Dr. Teresa Guerrero Urbano and deepc to convert AI research into a clinically viable tool.
Through this collaboration, advanced AI segmentation models developed in academic research are being adapted to meet clinical safety and regulatory requirements. Using deepc’s platform, designed to meet stringent medical device quality and safety standards, enables the AI models to be securely deployed and integrated seamlessly into the hospital’s radiotherapy planning system, allowing clinicians to maintain control over the treatment planning process while benefiting from AI-driven support.
A crucial component of the solution is data governance. With strict protocols in place for anonymizing and securely storing patient data, the team must ensure compliance while allowing for continuous model refinement and improvement.
Real-World Results: Saving Time, Improving Outcomes
The deployment of AI-driven segmentation showed promising potential to improve efficiency. Early estimates suggest clinician workload could decrease by 20–30%, with complex cases like head and neck organ segmentation potentially dropping from 45 minutes to just 10 minutes.
AI-driven tools help clinicians focus on the complexities of treatment rather than getting bogged down in repetitive tasks, ultimately improving the quality of care provided. In addition to saving time, peer review of AI-generated contours helped maintain clinical safety standards and reduce automation bias. By adhering to stringent regulatory requirements, the workflows aim to build clinician confidence, paving the way for seamless integration into daily practice.
deepcOS® offers scalability to this project, making it easy for the hospital to expand AI deployment across other departments and preparing the groundwork for broader adoption across additional NHS Trusts.
The Future of AI in Radiotherapy: Expanding the Impact
Following the successful implementation of their radiotherapy model, the hospital plans to extend the use of AI to other areas of cancer care. Future developments include adaptive radiotherapy, which allows real-time adjustments to treatment based on patient response and integrating data from multi-modality imaging such as MRI and PET-CT for even more precise treatment planning.
With deepc’s scalable infrastructure, there’s also potential for expanding these innovations across the NHS network, allowing other hospitals to benefit from the same time-saving and accuracy-boosting solutions.
If you're curious about other healthcare institutions adopting AI, our recent webinar on How Sycai Medical Brought AI into Clinical Practice explores the journey of developing AI for detecting pancreatic, kidney, and focal liver lesions.
Key Takeaways for Healthcare Professionals
The success of AI integration at Guy’s and St Thomas’ Hospital highlights three essential lessons for healthcare professionals:
- Clinician engagement from the beginning ensures that AI tools are practical and easy to implement in daily workflows.
- Strict adherence to data governance and regulatory compliance helps to guarantee patient safety while enhancing trust in AI-driven solutions.
- Automating repetitive tasks frees clinicians to focus on delivering better, more personalized patient care, leading to improved outcomes.
Conclusion: Redefining Radiotherapy with AI-Driven Innovation
The collaboration between Guy’s and St Thomas’ Hospital, and deepc shows how AI can improve efficiency and streamline radiotherapy workflows. By automating time-intensive tasks and improving the efficiency of treatment planning, AI allows healthcare providers to focus on delivering the highest standard of care.
As this partnership expands to other NHS Trusts and clinical applications, it sets the foundation for a future where AI becomes a central tool in the fight against cancer; improving patient outcomes, reducing clinician workload, and streamlining healthcare delivery.
Watch the Full Webinar to discover how AI is already transforming radiotherapy at Guy’s and St Thomas’ Hospital.
