Follow us on X
Follow us on Linkedin
Real-world-data enabled assessment
for health regulatory decision-making

Use cases

The REALM platform is a medical device software that is being validated and improved through five case studies. These use cases involve the application of AI-based algorithms to address various unmet clinical needs. Here are the details of each use case:

DuneAI (UM) is a use case that involves evaluating an automated segmentation software for detecting and segmenting non-small-cell lung cancer tumours in CT scans. This deep learning technique-based software has been validated on a large dataset and has shown high sensitivity and specificity in addition to fast segmentation times when compared to human experts. The purpose of this evaluation is to assess the software's usability and usefulness in a clinical setting and to improve its integration into clinicians' workflow.

COPowereD (COM) aims to predict hospitalisation or acute exacerbations in patients with chronic obstructive pulmonary disease by using medical AI algorithms that include patient-reported outcomes. The solution is a part of the COMUNICARE application, which is a generic solution for patient empowerment and self-efficacy in chronic care and post-hospitalisation contexts. The goal is to reduce the number of hospital admissions for chronic obstructive pulmonary disease exacerbations.

Pharmacogenomics Passports to Practice (PGx2P) is a use case that aims to implement preventive pharmacogenetics testing of a panel of genetic variants approved for clinical use by the Dutch Pharmacogenetics Working Group guidelines. This project is led by VITO and will recruit five general practitioners in the region of Antwerp, Belgium, to evaluate PGx2P with patients who could benefit from the testing. The project aims to facilitate stratified prescription of commonly used medications, resulting in a >25% reduction of overall adverse drug responses and >20% better adherence to prescriptions in PGx2P tested individuals.

The STAR project (UoL) aims to develop a model-based decision-support system to enable precise regulation of blood glucose levels in intensive care unit (ICU) patients. The project involves selecting 2-5 ICU clinicians and 30 ICU nurses from Hungary, New Zealand, and Belgium to evaluate the system's usability in a clinical setting. The goal is to assess the usability of STAR in a clinical setting, comparing results from experienced users versus naïve clinical users.

AI models for COPD and ASTHMA inpatient risk stratification (ASCOPD) (TRAQBEAT) is a use case that proposes a solution for the inpatient risk stratification of patients suffering from asthma and chronic obstructive pulmonary disease. The solution includes a Traqbeat sensor and biomarkers measuring methods for predicting inpatient mortality, acute respiratory failure, and ventilator dependence. The proposed solution is aimed at reducing hospitalisation duration and cost and increasing the quality of life of patients suffering from these diseases.

All five demonstrators will be thoroughly evaluated using both qualitative and quantitative methods to ensure their perceived usefulness, usability, and accuracy in a clinical setting. The expected benefits of implementing these AI-based solutions in the clinic include improved detection accuracy, faster and more reliable segmentation, less variability in clinical outcomes, reduced hospitalisation duration and cost, and increased quality of life for patients suffering from various diseases.