Chicago, IL
info
The Need
Precise auscultation, or listening to heart sounds, provides doctors with important information related to the health of the heart. Unfortunately, due to the manual nature of auscultation, this type of detection can be highly subjective and labor intensive. As a result, doctors have become more reliant on expensive, but useful automated metrics of determining cardiovascular health. There are also a variety of other external factors reshaping the medical landscape including; burgeoning patient expectations, the desire of administrators for a modern, boutique clinical setting, the emergence of point of care clinics, and a broad push for cost containment.
To help improve the diagnostic accuracy of listening to the heart, HeartSounds offers a computer-aided auscultation platform known as PATCH™ (Passive Auscultation Technology to determine Cardiovascular Health). The PATCH™ will provide physicians with cost-effective, non-invasive, objective devices for use in detecting specific cardiac dysfunctions and abnormalities. Utilizing its patented technology, HeartSounds’ PATCH™ will meet the needs of patients, physicians, and hospitals. Because of its adaptive, dynamic algorithms, the PATCH™ allows patients to relax and move freely as monitoring occurs, while at the same time offering an improved analysis of cardiovascular health. By providing the physician with a more consistent and objective measure of the heart, he/she can better assess the status of the patient, and, thus, more confidently recommend treatment. Hospitals can continue to modernize their offerings while also enjoying the cost savings associated with improved referrals and early detection.
The technology that underlies HeartSounds’ devices is a digital signal processing technique known as Independent Component Analysis (ICA), a method of Blind Source Separation (BSS). BSS is a relatively new idea, with the seminal work performed by Jutten, Herault and Guerin in 1988 where they proposed an algorithm that can isolate specific desired sounds from the unnecessary concomitant sound data that occurs. Historically, simple filtering techniques have been used in an attempt to isolate concurrent sounds and, in many cases, these techniques are effective. However, when the desired sounds and the unwanted sounds are similar, overlapping, or of low amplitude, these simplistic techniques break down. The ability of ICA to ignore both the sameness and temporal overlap of the source signals is a key element of HeartSounds’ devices. The ICA algorithm is also adaptive, which means it is less likely to lose data due to physical movement. ICA can be applied to “any physically measured set of signals, and to any number of signals (e.g., images, biomedical data, or stock prices).” The company’s products leverage the distinctive abilities of ICA in conjunction with other classic and novel mathematical methods to extract uniquely independent source sound data.
C. Jutten, J. Herault A.Guerin, Artiÿcial Intelligence and Cognitive Sciences, Manchester University Press, Manchester, 1988 pp. 231–248.
Stone, J V: Independent Component Analysis, Encyclopedia of Statistics in Behavioral Science, Vol 2 p 908, 2005
HeartSounds currently has two published patents, both pending payment of issuance fees. One patent encompasses the general method of extracting discrete source sounds from composite heart sound information utilizing ICA. The second patent protects the extraction of the heart sounds of a fetus from composite fetal heart sounds and background noise. These patents also grant the company exclusive rights to a sensor head containing multiple microphones to capture heart sounds. This device design is utilized across all current and anticipated product lines. Following the advice of IP attorneys from Bell, Boyd and Lloyd, HeartSounds is also in the process of filing an improvement patent that further clarifies the composition of the noise components during fetal heart sound extraction.
Chicago, IL
info