){kind=small}
Applying Data Science to Medical Data
In this era of unprecedented data availability, our researchers are at the forefront of transforming healthcare through data-driven insights.
Leveraging the power of medical data and health informatics, our projects aim to enhance diagnostic precision, treatment efficacy, and patient outcomes. From developing advanced machine learning algorithms for predictive analytics to creating secure and interoperable health information systems, our researchers are dedicated to revolutionising the way healthcare is delivered and managed.
-
Three-dimensional hybrid guidance system for cardiac interventional procedures
AI-enabled Multi-Tier Trust Management in the Internet of Medical Things
Improving early diagnoses for upper gastrointestinal (UGI) tract cancers
We also have a selection of projects in other themes with relevance to Data in Health:
PREDICTion of GastroIntestinal malignancy in patients with IRON Deficiency (PREDICT-GI-IRON)
CAVA - Continuous Ambulatory Vestibular Assessment
Participants
Summary
Dizziness affects up to 20% of the population, mainly the older population, placing time and cost pressures on health services. Diagnosis can be challenging, as there are no objective tests and self- reporting is subjective. Vertigo, a type of dizziness, results in an unusual eye-movement which can aid diagnosis, but it is often not present by the time patients visit their clinician. With previous funding from the NIHR and MRC, the CAVA device was developed to provide continuous, month-long recordings of eye-movements from patients. It records horizontal and vertical eye-movements by way of electrical signals produced by the eyes, captured using electrode pads adhered to the face, and an electronic logging over the ear. Artificially Intelligent Deep Neural Network Algorithms detect abnormal eye movements, enabling clinicians to provide a retrospective diagnosis. The vision for the CAVA system is to create a diagnostic pathway which improves the speed and accuracy of both diagnosis and access to treatment.
This work involves Mr John Phillips (NNUH), Professor Stephen Cox (CMP), Dr. Jacob Newman, Dr Emmanuel Jammeh.
Funding (previous)
Three-dimensional hybrid guidance system for cardiac interventional procedures
Participants
Summary
X-ray fluoroscopy is a common type of medical imaging that shows a continuous X-ray image on a monitor. Surgeons use it to guide devices (e.g. wires, catheters and stents) into the target area (e.g. heart) during minimally invasive cardiac surgeries. However, Xray images only produce 2D information, and the extended duration of X-ray radiation exposure is harmful to patients. The aim of this project is to build a new 3D hybrid guidance system which will: (a) provide 3D information allowing surgeons to act more efficiently reducing treatment time and thus overall X-ray exposure; and (b) make use of guidance information from the cardiac mapping system other than the X-ray system to significantly reduce the frequency of X-ray exposure.
To develop this system, we will use advanced computer vision techniques to detect devices and extract 3D blood vessel models from X-ray images, and then fuse these with existing 3D models inside the cardiac mapping system to provide the completed information to guide the procedure. Our proposed system will significantly reduce X-ray radiation exposure. This will benefit patients as X-ray radiation might cause the cancer in their later life. Furthermore, our proposed system will shorten the procedure time and reduce the cost of treatment for cardiovascular diseases.
Funding
AI-enabled Multi-Tier Trust Management in the Internet of Medical Things
Participants
Summary
This project is a collaborative effort between UEA and Edge Hill University, UK. The current funded amount for Phase 1 is £31,878, which can be extended to Phase 2 (up to £100k), subject to grant conditions. The project aims to improve the trustworthiness of IoMT devices by enhancing their privacy, reliability, safety, and resilience against cyber threats in the healthcare sector. The project is intended to support stakeholders such as the NHS, private healthcare providers, manufacturers, end users, and patients.
Partners
Improving early diagnoses for upper gastrointestinal (UGI) tract cancers
Participants
Summary
According to various studies, upper gastrointestinal (UGI) cancers are frequently missed during endoscopy and are often diagnosed later. The percentage of early diagnoses for these cancers remains low, around 12%, and has not changed significantly for several years.
The study's hypothesis is that demographic and clinical variables, extracted from routinely collected data in written endoscopy and histology reports at the time of a "cancer-negative" endoscopy, may serve as risk factors for Post Endoscopy Upper Gastrointestinal Cancers (PEUGIC). This study aims to assess the feasibility of using advanced text mining techniques on routinely collected endoscopy reports to predict missed UGI tract cancers.
Text mining methods have advanced rapidly in recent years, and deep learning in particular shows great promise. These methods may identify new features in medical reports associated with PEUGIC. We will compare various feature extraction techniques (e.g. n-grams, word embeddings) and different algorithms (such as classical text mining approaches and deep learning) to determine the most efficient approach.
This project is in collaboration with Norwich Medical School and the Department of Gastroenterology, Norfolk and Norwich University Hospital, so it will be interdisciplinary.
)
Partners
Publications
[1] Alexandre, L., Tsilegeridis-Legeris, T., & Lam, S. (2022). Clinical and endoscopic characteristics associated with post-endoscopy upper gastrointestinal cancers: A systematic review and meta-analysis. Gastroenterology, 162(4), 1123-1135. https://doi.org/10.1053/j.gastro.2021.12.270
[2] Alexandre, L., Tsilegeridis-Legeris, T., & Lam, S. (2021). P4 Clinical and endoscopic characteristics associated with interval upper gastrointestinal cancers: a systematic review and meta-analysis. Gut, 70, [A43]. https://doi.org/10.1136/gutjnl-2020-bsgcampus.79
[3] Edo-Osagie, O., De La Iglesia, B., Lake, I., & Edeghere, O. (2019). Deep Learning for Relevance Filtering in Syndromic Surveillance: A Case Study in Asthma/Difficulty Breathing. 491-500. Paper presented at International Conference on Pattern Recognition Applications and Methods 2019, Prague, Czech Republic. https://doi.org/10.5220/0007366904910500