The Commoditization of Ultrasound: The Hand-held Avalanche
G. Wayne Moore
05/09/2021
Diagnostic ultrasound systems have been through multiple technologically important inflection points over the last 40 years – including the first one I was directly involved in – the transition from analog to digital scan conversion. However, hand-held battery-operated ultrasound devices are not new. Indeed, a company called Organon Teknika introduced the pictured Minivoid linear array imager in 1977, they have in 2021 become quite technologically sophisticated and are proliferating in the market (see photos of just a sample of these devices on the following page). The MiniVisor was a commercial failure in the ultrasound market for multiple reasons, but the main reason was that prospective end-users, i.e., ultrasound savvy physicians, found the image quality produced by the device basically non-diagnostic. That might have been because the 1-inch viewing screen was unreadable. But the creation of the MiniVisor was visionary in that it pointed to a future where clinically-useful, hand-held ultrasound devices would ultimately find their way into the hands of anyone who might want one. The future is now. However, as with any potentially disruptive new category device, one must also consider the potential unintended effects of the diffusion of that technology running too far ahead of the diffusion of the knowledge required in how to properly use it. Providing a hammer to a person does not make them a carpenter.
Enter artificial intelligence (AI). AI, some say, will truncate or even obviate the learning cycle of non-ultrasound users in how to both acquire and diagnose an ultrasound image. This reasoning is roughly analogous to saying that autonomous vehicles obviate the need for any passenger in the car knowing how to drive. While it may seem to make sense when you look at an impressive diagram of how the car will operate, the data to date suggests otherwise. Would you let your wife and newborn baby travel home from the hospital in an autonomous vehicle today? As we understand AI today, simply marrying AI with a hand-held ultrasound unit in the hands of a non-ultrasound trained user is no guarantee of an accurate diagnosis.
What we are witnessing in the market is the commoditization of diagnostic ultrasound. These small, hand-held, cabled, and wireless units sell from $1,000 to $5,000 depending on what “cloud support” services you also order with the device and can be purchased in the United States by any licensed physician – ultrasound trained, or not (or you can simply go on the web and buy one yourself!). The image and Doppler quality of these various units’ range, depending on the claimed intended use, from “questionable” to “fairly-good” – but those quality level calls are based on input I have received from trained ultrasound users. Someone who does not know what a good image or a good Doppler is supposed to look like has no way of knowing if what they are looking at is good or not. Having said that, these units in the hands of trained ultrasound users are a blessing to many around the world who otherwise would not have access to diagnostic imaging of any kind. That is a good thing.
It may be that someday we will have totally autonomous ultrasound devices and cars but, in my judgement, it is not today. Call me old school, but I prefer to drive myself.
Until Next Month,
Wayne
If there is any subject you would like me to weigh-in on in future articles please email me – [email protected], I would enjoy hearing from you.