Non-Invasive diagnostics of post-COVID patients with persistent dyspnea supported by KI - NICK
Project leader: Dr. Isabell Pink
Research field
- Basic and translational research on post COVID
Who is involved?
- Dr. Isabell Pink (Project leader, MHH)
- Dr. Jannik Ruwisch (MHH)
- Dr. Dominik Berliner (MHH)
- Prof. Dr.-Ing. Bodo Rosenhahn (LUH)
- Priv.-Doz. Dr. Jens Spiesshoefer (RWTH Aachen)
What is the aim?
Dyspnea is often complained by post COVID patients (pCp), regardless of initial disease severity. As in other symptoms, the pathomechanism is still unclear and diagnostic procedures of pulmonary function test are most often in normal range. Previous works showed impaired respiratory muscle strength using invasive methods in former hospitalized patients. Non-invasive techniques showed pathological values in non-hospitalized pCp, although this method is not independent of patient cooperation. The study aims to quantify dyspnea as sequel after a mild COVID-19 by using a combination of two non-invasive techniques for diagnostic of the diaphragm and impulse oscillometry (IOS). IOS provides insights into the mechanical properties of the distinct compartments of respiratory system, as a component in the pathophysiology of dyspnea. Patients will be recruited via the outpatient clinic (n=100) established in 2021. Data analysis will be carried out in cooperation with the Institute for Information Processing. Using machine learning, correlations between the measurements of standard diagnostic procedure and the study-diagnostics are evaluated. The work focuses on interpretable models, sparse methods and feature importance ranking to identify relevant procedures for diagnostics. To avoid spurious correlations, potential findings will be evaluated in interdisciplinary discussion between clinicians and data scientists. This offers the possibility of a comprehensive examination of pCp with dyspnea by local pulmonologists and further therapeutic approaches by using an efficient AI-human expert cycle.