CAI-X is the framework surrounding the many projects with clinical AI at OUH and SDU.
The project aims to develop AI algorithms based on Natural Language Processing to identify previous incidents of venous thromboembolism in patients’ electronic health records for thrombosis risk assessment.
The purpose of the project is to improve diagnostics of cervical cancer with artificial intelligence. The project will develop, test and implement an AI solution to guide the colposcopist in real time to find the best spots for biopsies.
This experimental phantom study aims to assess the effect of DLIR in combination with MAR and Dual energy CT.
The DETECT Acute project is a clinical two-centre study aiming to assess the diagnostic performance of low-dose computed tomography (CT) with a novel Deep Learning Image Reconstruction algorithm (DLIR) in patients with acute abdominal conditions.
The aim is to provide efficient, fast and cheap diagnostics of patients with suspected skin cancer through the use of a telemedical solution called teledermoscopy
The goal of this project is to develop a prototype of FibroBot, a robot that will be able to measure the level of scar tissue in the liver with a Fibroscan probe.
The MERIS algorithm, which is based on EHR data, is a validated and evidence-based tool developed to identify patients in high risk of receiving unsuitable medication through medication reviews.
This project will examine the potential of AI-based chatbots in the healthcare sector, starting with patients with endometriosis who are in contact with the Department of Gynaecology and Obstetrics at Odense University Hospital.
The aim is to aid clinical decision making using artificial intelligence (AI) to find critical information in electronic health records.
The aim is to develop a data-driven prediction model capable of identifying individuals with Ischaemic Heart Disease at risk of developing anxiety or depression.
The goal of the project is to create a complete and validated AI-assisted pathway that improves CCE diagnostics making the technology clinically viable for the good of patients, health care systems and society.
This project aims to develop new ways of detecting epileptic seizures by analysing involuntary movements of eyes and face.
The aim is to test and validate the effect of a sentinel algorithm applied on health care data for early recognition of elderly community-dwelling adults at risk of acute hospital admission.
The RELIP project will develop an algorithm to use existing information in the patient recordto ensure early identification of a problematic alcohol consumption among hospitalised patients.
This PhD project will develop and validate a decision aid for use in primary health care to assess the risk of advanced fibrosis in patients with alcoholic- and non-alcoholic fatty liver disease.
The project developed the model MAS-AI which is an HTA-based framework to support the introduction of novel AI technologies into healthcare.
The objectives of this project are to validate and test an AI system in an existing radiology workflow and analyse the impact on cancer detection and recall rates as well as workload.
In this project, we have developed a machine-learning algorithm with the ability to improve the diagnostic accuracy of SWE pictures in patients with chronic fatty liver disease.
The goal is to explore whether cutaneous variables that are not apparent to the medical professional naked eye (i.e. invisible facial redness and the variation of redness due to heart pulsation) could be markers of diabetes and in particular diabetic neuropathy.
The project aims to reduce the number of unforeseen deteriorations of patients in medical emergency departments.
The FAST-MRI project aims to investigate the accuracy, workload changes and time-to-diagnosis when implementing an AI system for stroke MRI diagnostics in a Danish setting.
The study aims to develop a software system that utilises deep-learning by fully convolutional networks in order to perform automatic grading of retinal images to detect diabetic retinopathy.
The project aims to use AI to recognise audiogram patterns. The aim is to recognise at least 8 to 10 of the most common audiogram subtypes.
The project uses machine learning to analyse big data from previous medical history.
Automatic selection and imputation of vital signs in emergency departments.
The project is a branch of the automated ultrasound scanning robot RoPCA which will ensure systematic and uniform method to perform ultrasound scans, thereby ensuring physicians the best possible starting point for the determination of treatment. Such systematic scanning is not possible today due to inaccuracy in human performance.
Full title: Development of machine learning algorithms for prediction of cancer and other serious conditions
The Department of Pathology analyses a great deal of tissue samples for metastases; each sample is not time-consuming, but automisation of the analyses would save quite som time due to the number of samples.