The Laeneco team has developed a smart stethoscope that detects lung disease with greater accuracy than doctors. Next - about the components of the device and its capabilities.
Photo ©
Difficulties associated with the treatment of lung diseases
According to the World Health Organization, respiratory diseases account for 10% of the . And this is one of the most common reasons people go to clinics (after cardiovascular diseases).
The most common method for detecting lung disease is auscultation. It consists in listening to sounds caused by the activity of internal organs. Auscultation has been known since 1816. The first person to put it into practice was a French physician and anatomist. . He is also the inventor of the stethoscope and the author of a scientific work describing the main auscultatory phenomena - noises, wheezing, .
In the XNUMXst century, doctors have at their disposal ultrasound machines that allow not only to hear, but to see the internal organs. Despite this, the auscultation method still remains one of the main medical tools. For example, the importance of auscultation in medical practice is emphasized by Valentin Fuster, MD. In his he cited six cases (all occurring within 48 hours) where stethoscope diagnosis helped to make an accurate diagnosis that was not obvious on the pictures.
However, the method has its drawbacks. In particular, doctors do not have the means to objectively control the results of auscultatory examination. The sounds that the doctor hears are not recorded anywhere, and the quality of the assessment depends solely on his experience. According to various estimates, the accuracy with which a doctor is able to detect pathology is approximately 67%.
Engineers from — a startup that passed through the acceleration program of ITMO University. They have developed a smart stethoscope that uses machine learning algorithms to detect lung diseases from audio recordings.
Possibilities and prospects of the solution
An electronic stethoscope has a sensitive microphone that picks up a wider range of frequencies than the human ear. At the same time, doctors get the opportunity to increase the volume of audible noises. This is important when working with overweight patients, since sound penetrates through thick human tissues worse. Plus, the function is relevant for health workers at an age whose hearing acuity is no longer the same as in youth.
Deep neural networks help to identify sounds that indicate the presence of a disease. Now the accuracy of their work is 83%, but in theory this figure can be increased to 98%. The startup team is already collecting new data to expand the training sample.
Photo: /PD
A smart stethoscope works in tandem with a smartphone. The application gives users recommendations regarding diagnostics, saves and processes records, displays measurement results. Thanks to this device, people without medical education can use it.
The Laeneco team is convinced that the smart stethoscope will help reduce the risk of chronic lung disease and plans to expand the tool's capabilities. One of the main tasks is to develop functionality for detecting cardiac pathologies.
About Laeneco
Team consists of three people: Evgeny Putin, Sergey Chukhontsev and Ilya Skorobogatov.
Evgeny works as a programmer-engineer at ITMO University’s Computer Technologies Lab and leads the Kaggle Club for solving practical machine learning problems. He is also the author of the resource that can predict a patient's age from a blood test.
The second member of the team, Sergey, graduated from the Institute of Law at the Udmurt State University and is one of the authors of the network plant concept. It is designed to manage many independent industries.
As for Ilya, he is a graduate of ITMO University with a degree in Information Technology and Programming, who has been working on issues of production automation and workflow for a long time. The idea to create a smart stethoscope came to him when he was developing a sensor for analyzing the sounds emitted by machine tools.
In 2017, the Laeneco team passed the acceleration program . The participants formed the business model and developed the smart stethoscope MVP. The system was presented at the startup festival *SHIP-2017 in Finland and the St. Petersburg forum SPIEF'18. Also in 2018, the project became the winner of the pitch session "”, organized by the Technopark of ITMO University together with experts from Asia. At the same time, Laeneco received an offer to bring their product to the Japanese market.
Other habraposts of ITMO University:
PS If you are related to ITMO University and would like to talk about your project or scientific work in our blog on Habré, send potential topics pm.
Source: habr.com