The number of people with chronic lung diseases is rapidly growing. Abnormal lung sounds are among symptoms of these diseases, which are detected during a quick physical examination with a stethoscope. However, detection and interpretation of auscultation findings is subjective and manual, research has shown that clinicians interpret lung sounds differently. This makes the auscultation results unreliable and communication between health professionals more challenging. There have been developed digital stethoscopes, that allow sound recording, but currently there exist no solutions for automatic and objective detection of lung sounds. Medsensio is developing a cloud- and subscription-based platform for automatic detection and classification of normal/abnormal lung sounds, using deep learning, with an exclusive license to access worlds biggest dataset of annotated lung sounds. The first product - training tool for learning lung sounds- built on this platform, is now market-ready. During conducted beta tests it received positive feedback and obtained affirmative results from analyzing its intended use. Medsensio has also received interest for the lung sound analysis platform from clinical perspective and believes that the platform has greater user value, commercial potential and societal impact as a clinical tool. We want to assess how feasible it is to build such application with all necessary regulatory approvals, using our platforms algorithms, and evaluate its market potential for developing a sustainable business model. Such feasibility study fully corresponds to the application description for EU project under SME instrument phase 1. In order to develop a corresponding project plan along with an application for SME instrument call, Medsensio is applying for a project establishment support from the Norwegian Research Council. Medsensio declares that the company's competence and resources to implement the EU project are reasonably sufficient to our best knowledge.
Project leader: Johan Fredrik Eggen Ravn
Institution: MEDSENSIO AS