In order for the WELMO automatic detection algorithms to be properly trained, two rounds of data collection from real patients were organized and executed in May and November 2019. During these data collection activities, lung sounds and EIT signals were gathered from patients with remarkable clinical signs as well as from healthy volunteers. The process took place at the Respiratory Department of “G. Papanikolaou” General Hospital in Thessaloniki, Greece and the outpatient clinic of the same department. The task was carried out by researchers form AUTH, UCOMP and CAU in two weekly sessions on the aforementioned dates.
Patients hospitalized or examined due to any respiratory disease who presented remarkable adventitious lung sounds and/or abnormal radiology images were recruited. They were submitted to spirometry, measurement of pulse oximetry and auscultation with stethoscopes capable of digital recording of lung sounds. At the same time, the patients were monitored for their Electric Impedance Tomography (EIT) signals with the EIT cables attached to patches around their thorax. The auscultation recordings and the EIT signals were monitored during certain maneuvers, including shallow and deep breathing, moving the arms, coughing or speaking. The same procedure was followed in the cases of the healthy participants. In total, 32 patients and healthy subjects were included in the study.
The data collection phases provided the consortium with a valuable and large database unlike any other existing biosignals dataset, since it combines lung sounds together with simultaneous EIT recordings for patients with a variety of respiratory pathologies. It will serve as a baseline for further signal analysis and development of dedicated classification algorithms for lung sounds and the interpretation of EIT signal features. The overall process also helped the researchers identify common pitfalls and practical issues during patients’ recruitment in the specific hospital (where the actual pilot trial with patients for the WELMO system will take place), paving the way for a successful clinical trial towards the final stages of the project.
Dr. Evangelos Kaimakamis
Aristotle University of Thessaloniki