Medsensio is developing a cloud-based platform for automatic detection of lung sounds, using deep learning. This will provide decision support to any health practitioner, that performs lung examination with a digital stethoscope, in order to accurately, objectively and consistently detect abnormal lung sounds in patients with lung issues. Lung diseases are a growing problem, characterized by: ? increasing prevalence; ? shortage of experienced health personnel and workload pressure; ? economic burden due to direct/indirect cost. Auscultation is a low-cost, non-invasive procedure to examine lungs. However, its interpretation is subjective and manual, research has shown that clinicians interpret lung sounds differently. There is therefore an unmet need for tools that can eliminate subjectivity, inconsistency and enable wider circles of health staff to provide such examination. Our solution is easy-to-use: take a digital stethoscope – record lung sounds – sounds are sent to the cloud – get automatic analysis in our platform – use the platform’s answer to create a record, monitor and communicate the findings further. Our innovativeness is in the following: ? state-of-the art deep learning methods; ? trained on world’s biggest annotated dataset; ? compatible with existing digital stethoscopes. The platform is already used in a training tool (https://lungsounds.medsens.io/) and we have a documented industry interest for our tool used in a clinical setting (clinical tool – product 2). Medsensio has the right team in place and an exclusive partnership with UiT and the Tromsø Study. The main objective for FORNY 2020 is to ensure that our clinical tool can get recurring revenues via stepped go-to-market approach with sales to pilot validations and later via distributors, in support of getting funding to pursue certification. For the success of the project we will establish strategic relationships with potential partners, stethoscope manufacturers, and customer groups.
Project leader: Anna Dranovska
Institution: MEDSENSIO AS