A novel clinical evaluation programme aims to examine the impact of artificial intelligence (AI) on radiology decision making in the lung cancer pathway and the patient outcomes from these decisions. The programme, called INPACT*, is a collaboration between Aidence, clinical consultancy Hardian Health, and the University of Edinburgh, with funding from the NHSX through the NHS AI in Health and Care Award.
The radiology workflow
The study will focus on radiologists reporting on chest CT scans in routine clinical practice. In this setting, patients undergo scans for varied health conditions. If the radiologist finds an abnormal growth in the lungs – which may be an indicator of lung cancer – they make an assessment of the nodule’s characteristics and a recommendation for managing its evolution.
Most of the lung nodules detected in routine practice are benign, but some are cancerous. Their identification and follow-up are crucial to an early lung cancer diagnosis, when a cure may still be possible. However, these are challenging tasks for UK physicians. Miguel O. Bernabeu, Senior Lecturer in Medical Informatics and Deputy Director of the Bayes Centre at The University of Edinburgh and Chief investigator in the programme, explains:
“Spotting lung nodules on a chest CT scan which might potentially develop into malignant disease is painstaking and time-consuming work for already-stretched NHS radiologists.”
An innovative approach
Research into AI medical imaging solutions is often centred around their performance, for example, the percentage of anomalies correctly flagged on a scan. These studies take place in controlled research conditions that do not match real clinical workflows. INPACT aims to gain insight into an understudied area: the impact of AI technology on human decisions.
As part of INPACT, radiologists will perform their lung nodule analysis with the support of Aidence’s Veye Lung Nodules, an intelligent software that automatically finds, segments, measures, and tracks the growth of pulmonary nodules on chest CT scans. The programme will investigate Veye Lung Nodules’ influence on how radiologists manage pulmonary nodules.
Miguel O. Bernabeu, on the approach and value of this programme:
“We hope to prove, through one of the first studies of its kind, how artificial intelligence software can improve the detection and management of actionable nodules in the lung cancer pathway, leading to better and more appropriate follow-up for at-risk patients. This has the potential to free up valuable radiology time and contribute to earlier diagnosis, thereby improving the life chances of those who face the prospect of this deadly disease.
“We are immensely grateful to the NHS teams who are working with us to make this study possible.”
Hugh Harvey, Managing Director at Hardian Health, added:
“Digital health technology has enormous potential to improve detection of disease, leading to better health outcomes and longer patient survival. We will be examining the value of such benefits, and associated costs, through our health economic modelling and will be evaluating the case for national adoption at scale.”
Timeline and expected results
Aidence’s Veye Lung Nodules will be implemented in 20 new hospitals in the UK. Six of these will participate in the evaluation and will be confirmed by the end of the year. The results of the study will appear in reports and scientific publications in 2022.
David King, Project & Delivery Manager at Aidence, on the expected results:
“The NHS AI Award is a unique opportunity for us to generate real-world evidence of the impact new and emerging technologies can make to health and care outcomes. This will only improve trust and encourage adoption of proven tools that can make a lasting difference.”
Dr Indra Joshi, Director of AI, NHSX, said:
“Using AI technology like Veye to help with early diagnosis and treatment of lung cancer will have a huge impact on people’s chances of recovery. The AI Award funding is going where it is needed most – testing AI innovations that are tackling some of the biggest killers.”
Dan Bamford, Deputy Director AI Award, Accelerated Access Collaborative, added:
“I am excited to see 20 hospitals signed up to trials for the Veye Lung Nodule technology. Getting real-world testing is an essential part of our evaluation through the AI Award and will help to ensure that AI is used safely and ethically for health and care.”
Early lung cancer detection
Earlier detection can have life-changing consequences for those affected. It increases the likelihood of successful treatment and long-term survival while avoiding the massive costs of late-stage treatments.
David King, on behalf of Aidence:
“This year, over 52,000 people in the UK will receive a lung cancer diagnosis. Until we can improve the detection of the disease at its earliest, still treatable stages, lung cancer will remain the UK’s biggest cancer killer.
We are proud to partner with the NHS to explore how AI technology can improve the life chances of everyone affected by this deadly disease.”
*INPACT – Investigating Nodule Protocol Adherence using CADe/x Technology: A real-world evaluation of the impact on, and outcomes from, radiology decision making using AI software for pulmonary nodule management.
About Hardian Health
Hardian Health is a clinical digital consultancy helping startups and tech developers to tailor their strategic approach by focusing on scientific validity, regulations, health economics and procurement.
About the University of Edinburgh
The University of Edinburgh is a public research university. It is one of Scotland’s four ancient universities and the sixth-oldest university in continuous operation in the English-speaking world.
About the AI award
The Artificial Intelligence (AI) in Health and Care Award aims to benefit patients by combining the power of artificial intelligence with the expertise of the NHS to improve health and care outcomes. The Award is making £140 million available over four years to accelerate the testing and evaluation of technologies most likely to meet the aims set out in the NHS Long Term Plan.