News: 1651746610

  ARM Give a man a fire and he's warm for a day, but set fire to him and he's warm for the rest of his life (Terry Pratchett, Jingo)

Watchdog rubber-stamps cavity-detecting neural network for dentists

(2022/05/05)


VideaHealth, a dental AI startup spun out of MIT and Harvard, has received FDA approval in the US to commercialize its cavity-detection algorithm for clinical use.

The software is designed to help dentists look for signs of tooth decay in X-ray scans. Officials granted Videa 510(k) clearance, meaning it has successfully demonstrated its product can be marketed as safe and effective for human use and functions similarly to an approved device.

The algorithm demonstrated it could reduce the number of caries missed by dentists by up to 43 percent, and slash dental lesion detection errors by 15 percent on average in the FDA trial, the biz said.

[1]

"Our technology supports the dentist's ability to detect diseases more reliably and more accurately. It streamlines workflows and also increases case acceptance rates since our AI detection solution offers an impartial secondary analysis," a Videa spokesperson told The Register .

[2]

[3]

The cavity detection algorithm is based on a convolutional neural network, and was trained and tested on a giant dataset with over 100 million samples of X-ray images collected from the dental industry, insurance companies, and universities.

"Our biggest priority as a team is ensuring that our solution is effective across diverse patient populations and helps dentists deliver the most accurate diagnoses," the upstart's CEO Florian Hillen [4]explained in a statement. "This paves the way for more appropriate dental treatment recommendations and the opportunity for dentists to foster deeper patient engagement."

[5]DARPA to build life-saving AI models that think like medics

[6]Google helps develop AI-driven lab machine to diagnose Parkinson's

[7]AI-designed drug to treat deadly disease now tested on humans

[8]AI models still racist, even with more balanced training

Videa argued that dental technology had been pretty stagnant before machine learning came along. "With the exception of the introduction of digital sensors and intraoral cameras, dental technology advancements have been scarce in the last 30 years," the spokesperson said.

The company believes AI will better standardize care for patients. "There [is] a lack of standard quality in dentistry. If you go to three different dentists you might get three different opinions."

[9]

"AI is the next big thing that will help move the industry forward. It will radically accelerate and improve decision-making for dentists and dental insurers and ensure patients receive more accurate diagnoses and better preventative care," the spokesperson added.

Videa was founded in 2018, and was spun out of Harvard and MIT research. It has raised a total of $26.4 million so far with funding rounds led by Zetta Venture Partners and Pillar VC. ®

Get our [10]Tech Resources



[1] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2YnP0qQwaJnXqHTyZb0T33wAAABQ&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0

[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44YnP0qQwaJnXqHTyZb0T33wAAABQ&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0

[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33YnP0qQwaJnXqHTyZb0T33wAAABQ&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0

[4] https://www.videa.ai/news/videahealth-dental-ai-solution-receives-fda-510-k-clearance-establishing-industry-wide-benchmark-for-clinical-accuracy

[5] https://www.theregister.com/2022/03/30/darpa_ai_medicine/

[6] https://www.theregister.com/2022/03/25/google_ai_parkinsons/

[7] https://www.theregister.com/2022/02/24/ai_drug_humans/

[8] https://www.theregister.com/2022/05/01/ai_models_racist/

[9] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44YnP0qQwaJnXqHTyZb0T33wAAABQ&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0

[10] https://whitepapers.theregister.com/



Should not the headline ...

alain williams

have talked about the AI drilling down through the data ?

Re: Should not the headline ...

sniperpaddy

Yep. it will tell you the tooth, the whole tooth, and nothing but the tooth.

'Rubber-stamp' poor choice for headline

KLane

The usual meaning of 'Rubber-stamp' is to approve something with little or no checking/validation. Really not a proper usage.

In these matters the only certainty is that there is nothing certain.
-- Pliny the Elder