Oesophageal adenocarcinoma is the fourth commonest cause of cancer mortality in men in the UK. Most cases arise within Barrett's oesophagus (BE). Our research initially focused on development of optical diagnostic and therapeutic tools for managing premalignant changes in Barrett's oesophagus. This expanded to include discovery of biomarkers of progression and prognosis. We then switched to antibody directed imaging & photodynamic therapy for established oesophageal cancer. Over the last few years as data science has taken off, we have moved into machine learning and artificial intelligence and have expanded our use of these tools into colon cancer also. 

We have, however, developed the largest tertiary patient referral base in the UK & one of the largest patient cohorts in Europe. Most of our patients are keen to support the reserach efforts and more than 500 enrol in research trials every year. 

1. Optical spectroscopy and imaging. In our NIH funded work in collaboration with Boston University, we showed that elastic scattering spectroscopy could diagnose pre-cancer (dysplasia) in BE in vivo. Not only is it possible to detect high grade dysplasia in vivo in real time with an accuracy of up to 90%, but it is also possible to detect DNA content (aneuploidy). This led to a CRUK project grant at UCL followed by an MRC collaborative grant to demonstrate this prospectively in N. Ireland Barrett's registry. It also led to CRUK funded evaluation of cytosponge biomarkers of risk in BE in collaboration with Professor Rebecca Fitzgerald in Cambridge. We then developed collaborations with Prof Paul Beard in UCL Medical Physics & Bioengineering to develop photoacoustic imaging and with Prof Peter Rich in the Department of Structural & Molecular Biology at UCL to develop mid-infra red spectroscopy for ex vivo analysis of Barrett's oesophagus biopsies. This has resulted in a spin-out company DynamX. We are now exploring AI to identify dysplastic lesions in vivo. Our ultimate vision is to obtain instant optical diagnosis with no need for histology at all. 

2. Ablative therapy to treat dysplasia in BE. We treated more than 500 patients in phase 1 and 2 trials of photodynamic therapy (PDT). We then set up a national registry for radiofrequency ablation treatment (RFA) of BE to which we have recruited over 2,000 patients nationally from 26 centres. The Registry is recognized by NICE and all patients undergoing this treatment should be enrolled in it. This success led to the NIHR BRIDE randomised controlled trial of argon plasma coagulation therapy compared to RFA. 

3. Antibody Directed Photodynamic Therapy (ADP). This Innovate-UK funded work developed both for oesophago-gastric cancer and for bacterial disease. The laboratory work started in vitro. We then developed new in vivo models for testing these novel approaches to oesophageal cancer.

4. Saliva to Predict disease rIsk using Transcriptomics and epigenetic (SPIT). The use of invasive endoscopic testing to exclude important disease has, perhaps, spiralled out of control. We need simpler, less invasive, approaches to triage patients at risk. We have been funded by CORE, the Rosetrees Trust and the Charles Wolfson Charitable Trust to explore different diagnostic approaches using saliva to identify both the presence and the risk of developing oesophageal and colon cancer. We are also exploring whether this approach can be used to diagnose Crohn's disease. 

5. AI to detect Colonic Lesions We have built on our previous collaborations and data science expertise to collaborate further both within and outside UCL to develop artificial intelligence approaches to diagnosing colonic polyps.