About project DELTA
Project DELTA: integrateD diagnostic solution for EarLy deTection of oesophageal cAncer
Background
Incidence of the cancer type oesophageal adenocarcinoma (OAC) has increased 6-fold since the 1990s and carries a dismal prognosis. The UK has some of the worst outcomes from this disease in Europe. Clinical guidelines have focussed on minimising endoscopy referrals unless patients have "alarm symptoms" suggestive of cancer. Nevertheless, General Practice referral rates vary widely, and low endoscopy referral rates have been linked with poor outcomes from oesophageal cancer.
A major risk factor for this cancer is chronic heartburn caused by reflux. Three to six percent of individuals with reflux predominant symptoms may have the precursor lesion called Barrett's oesophagus, but only around 20% of patients with Barrett's are diagnosed. It is estimated that the burden of OAC could be reduced by up to 50% as a result of increasing the proportion of individuals with reflux symptoms who are investigated. This is a formidable task since heartburn symptoms affect between 5%-20% of the population and account for up to 10% of GP consultations. GPs therefore focus on controlling reflux symptoms with acid-suppressant medication, particularly proton pump inhibitor therapy (PPI). PPIs are highly effective but patients often continue taking them life-long and there are increasing concerns about long-term side effects including osteoporosis, pneumonia and recently allergy.
In 2008, the Chief Medical Officer, Sir Liam Donaldson, raised oesophageal cancer as a public health concern and identified an urgent need to develop a need for a safe, minimally invasive, affordable test applicable to the office setting to diagnose Barrett's oesophagus. We have developed a new minimally invasive test for patients with reflux that can be performed in the GP surgery. This test is called Cytosponge-TFF3 which has been tested in over 4,000 individuals across 3 continents with very promising results.
Project DELTA aim
Our vision is to redesign the clinical pathway for reflux. A new algorithm applied to NHS prescribing databases will flag symptomatic individuals at risk for oesophageal cancer to their GP. Individuals at risk will be offered a Cytosponge test in a nurse led clinic. The TFF3 stained cells collected by the Cytosponge will be assessed using Artificial Intelligence to increase the throughput and reduce the cost. Individuals at high risk for cancer will be referred for endoscopy and PPI use will be rationalised. These changes will result in an economic benefit to the NHS, a social benefit for early detection of a lethal cancer and a reduction in over-use of PPI medication.
Project DELTA began on 1 February 2020 and will run until 31 January 2023. The project will be split into four work packages:
Data Mining/algorithm development for identification of individuals at risk
To improve on the state of the art we need to incorporate Cytosponge-TFF3 into the clinical pathway by identifying suitable patients. To achieve this goal Professor Hippisley-Cox’s team will mine a variety of databases to identify at-risk individuals (including electronic health records and endoscopy and public databases). These linked data-sets are available in an anonymised form on a national scale. This includes nationwide drug prescribing data from 2015 which has already been used to conduct successful research studies (PMID: 29991628).
Dr Hall’s team will explore the regulatory and ethical challenges associated with data flows into and through the proposed patient pathway. It will also include consideration of the nature, veracity and volume of the training data that might be available to machine learning developers. This work will include consideration of the policy landscape for utilising data mining using AI/ML for personalised prevention; issues relating to the nature and quality of the data (to the extent that they impact on regulatory factors such as bias and discrimination); the reasonable expectations of key stakeholders to data processing by AI/ML (including patients and health providers); and assessing the potential legal and regulatory challenges associated with compliance with the GDPR, particularly if data mining is solely automated.
Building a transferable operating model and clinical implementation for a nurse/pharmacist-led clinic
This work package seeks to recruit patients and collect data to re-design and evaluate the clinical pathway such that we systematically identify those at risk, perform a simple test to inform who needs endoscopy and in so doing rationalise the use of long-term PPI medication. Due to the cost-effectiveness of Cytosponge™ -TFF3 compared with endoscopy these changes will likely result in an economic benefit to the NHS, a social benefit for early detection of a lethal cancer and a reduction in over-use of PPI medication.
Professor Fitzgerald’s team will explore a number of options for Cytosponge clinics in both primary and secondary care settings. The team will build on the evidence achieved to date and move towards Cytosponge rollout, assessing the feasibility and practical implementation steps required to introduce the Cytosponge - TFF3 as a triage test for endoscopy to identify Barrett’s oesophagus, early cancer and other oesophageal conditions.
A patient app will be developed by Dr Walter’s team for use in clinical care to enhance how patients experience the Cytosponge test for improved patient outcomes. The work led by Dr Waller aims to ensure that the app meets patient and provider needs and is acceptable to patients. It will also evaluate the impact of the app on patient reported experience. This will be done by interviewing patients and providers about the content and format of the proposed App, and gather views on the feasibility and acceptability of providing this App.
WP2 will utilise the results of Ai/ML data mining to invite at-risk patients for screening via a nurse-led clinic. Dr Hall’s team will explore specific regulatory challenges including the potential legal and regulatory challenges relating to information provision, obtaining consent and associated with compliance with relevant data processing legislation (GDPR and the Data Protection Act 2018).
Digital pathology and AI
The Cytosponge processing differs from endoscopic biopsy handling and requires quality control measures and expertise in Cytology for reporting. Deep learning methods have been shown to achieve excellent performance on diagnostic tasks, but it is still an open challenge how to optimally combine them with expert knowledge and existing clinical decision pathways. This question is particularly important for the early detection of cancer, where high volume workflows might potentially benefit substantially from automated analysis.
In this work package Marcel Gehrung’s team will develop and validate a deep learning framework to analyse samples of the Cytosponge -TFF3 test, a minimally invasive alternative to endoscopy, for detecting Barrett's Esophagus, the main precursor of esophageal cancer. Our approach based on a research prototype exploits screening patterns of expert gastrointestinal pathologists and established decision pathways to define eight triage classes of varying priority for manual expert review.
By generating hotspot identification and simultaneous smart report generation our software will be able to substantially reduce the time taken by a pathologist to review Cytosponge samples. This approach will further lay the foundation for tailored, semi-automated decision support systems embedded in clinical workflows with a perspective of using the technology in other diagnostic pathways.
Digital pathology developed in WP3 will be used at two points in the implementation pathway – (1) to evaluate the status of cellular samples from the Cytosponge test and (2) to analyse endoscopy samples from patients who have a positive screening result, for subsequent diagnosis, management and treatment. Dr Hall’s team will assess the legal, regulatory and ethical challenges associated with compliance with the GDPR, particularly if the screening invitation is solely automated. Where relevant, this element may also include consideration of relevant aspects of compliance with the EU Medical Devices Regulation, to be implemented in May 2021.
Dr O’Donovan will be responsible for leading the pathology systems. She will be training pathologists, developing cytosponge training material for Pathologists as well as recruiting and training a team of pathologists to report cytosponge samples. In addition she will lead the implementation of lab processing systems, sample reporting and standardized reports for Clinical users. She will also work with the IT team of developers to set up the AI systems
Health economic and implementation research
The aim of WP4 as a whole is to investigate factors affecting the implementation of the novel pathway involving Cytosponge®-TFF3.
Prof Morris’s team will undertake a discrete choice experiment (DCE) to analyse preferences for different pathways from patient and health professional perspectives. The objective of the DCE is to analyse the preferences of members of the public and health professionals for the Cytosponge-TFF3 test within the clinical care pathway for reflux disease. This will provide new data on what these stakeholders think about this innovation, which issues matter to them the most when considering this, and which factors they prioritise most strongly.
Prof Morris’s team will also calculate the cost-effectiveness of the novel pathway from an NHS perspective versus current practice. The objective of the economic evaluation is to undertake a cost-utility analysis and budget impact analysis of the novel Cytosponge®-TFF3 test clinical pathway versus current practice. This analysis will allow us to make preliminary judgements as to whether the novel pathway will represent good value for money compared with current practice to the NHS. In addition, it will permit us to identify the main cost drivers and important cost-effectiveness parameters of the pathway.
Dr Walter’s team will use qualitative and quantitative mixed methods approaches with patients and providers to understand the barriers and facilitators to implementation of the novel pathway.
The relationship between the work packages is illustrated in the below flow chart:
