Butterfly Network (NYSE: BFLY) announced this week that its AI-enabled auto B-line counter had received FDA 510(k) clearance. Butterfly Network, based in Burlington, Massachusetts, develops handheld ultrasound technology. B-lines appear as bright, vertical lines on an ultrasound scan, indicating lung moisture associated with pulmonary air-space disease. This includes congestive heart failure, chronic pulmonary obstructive disease (COPD), pneumonia, and COVID-19.
As per Mass Device, using deep learning technology, the automatic B-line counter generates a B-line count from a six-second ultrasound clip. This process for interpreting B-lines differs significantly from the traditional manual counting methods. According to Butterfly Network, this could facilitate the evaluation of adults with suspected diminished lung function by medical professionals. This may enhance their capacity to make informed treatment decisions at the point of care.
According to Dr. Jonathan Rothberg, founder and interim CEO of Butterfly Network, “Our mission at Butterfly is to provide healthcare professionals and, eventually, consumers with a real-time, annotated, full-color window into the human body. Applying AI to make ultrasound easier to use is fundamental to Butterfly and will enable more clinicians across specialties to monitor, assess, and prescribe treatments in a more informed manner.
Our AI-enabled Auto B-line Counter enables providers to assess lung conditions faster and with greater confidence, which in turn will aid in the earlier detection, diagnosis, and treatment of cardiovascular diseases, the world’s leading cause of death, responsible for nearly 18 million deaths annually.”
The algorithm for the auto B-line counter makes use of cutting-edge instant percent counting. In addition to counting discrete B-lines, this assigns numbers to confluent B-lines based on the fraction of rib space occupied. According to Butterfly Network, this technique is more reliable than current individual line counting methods. Using this algorithm, providers with the proper training can simply place the probe and receive an accurate number count on their screen.
Butterfly Network’s AI algorithms were developed and trained using the secure Butterfly Cloud. This provides access to over 3.5 million anonymized ultrasound cines. There is a wide range of age, gender, body mass, ethnicity, and race within these data inputs, which originate from hundreds of sites across the United States. The company anticipates an “early summer” release for the AI-enabled auto B-line counter in the United States.
