South Korean startup Lunit, developer of two FDA-cleared AI fashions for healthcare, went public this week on the nation’s Kosdaq inventory market.
The transfer marks the maturity of the Seoul-based firm — which was based in 2013 and has for years been a part of the NVIDIA Inception program that nurtures cutting-edge startups.
Lunit’s AI software program for chest X-rays and mammograms are utilized in 600 healthcare websites throughout 40 international locations. In its residence market alone, round 4 million chest X-rays a yr are analyzed by Lunit AI fashions.
Lunit has partnered with GE Healthcare, Fujifilm, Philips and Guardant Well being to deploy its AI merchandise. Final yr, it achieved FDA clearance for 2 AI instruments: one which analyzes mammograms for indicators of breast most cancers, and one other that triages essential findings in chest X-rays. It’s additionally obtained the CE mark in Europe for these, in addition to a 3rd mannequin that analyzes tumors in most cancers tissue samples.
“By going public, which is only one step in our lengthy journey, I strongly consider that we’ll succeed and achieve our mission to overcome most cancers by means of AI,” mentioned Brandon Suh, CEO of Lunit.
Lunit raised $60 million in enterprise capital funding late final yr, and its present market cap is a few $320 million, based mostly on its newest closing worth. Following its latest regulatory approvals, the startup is increasing its presence within the U.S. and the European Union. It’s additionally creating further AI fashions for 3D mammography.
Forging Partnerships to Deploy AI for Radiology, Oncology
Lunit has 4 AI merchandise to assist radiologists and pathologists detect most cancers and ship care:
- Lunit INSIGHT CXR: Educated on a dataset of three.5 million instances, this software detects 10 of the most typical findings in chest X-rays with 97-99% accuracy.
- Lunit INSIGHT MMG: This product reduces the prospect that physicians overlook breast most cancers within the screening mammography by 50%.
- Lunit SCOPE IO: Demonstrating 94% accuracy, this AI helps establish 50% extra sufferers eligible for immunotherapy by analyzing tissue slide pictures of greater than 15 kinds of most cancers, together with lung, breast and colorectal most cancers.
- Lunit SCOPE PD-L1: Educated on greater than 1 million annotated cell pictures, the software helps precisely quantify expression ranges of PD-L1, a protein that influences immune response.
GE Healthcare made eight AI algorithms from INSIGHT CXR out there by means of its Thoracic Care Suite to flag abnormalities in lung X-rays, together with pneumonia, tuberculosis and lung nodules.
Fujifilm included INSIGHT CXR into its AI-powered product to investigate chest X-rays. Lunit AI connects to Fujufilm’s X-ray gadgets and PACS imaging system, and is already utilized in greater than 130 websites throughout Japan to detect chest nodules, collapsed lung, and fluid or different international substances within the lungs.
Philips, too, is adopting INSIGHT CXR, making the software program accessible to customers of its diagnostic X-ray options. And Guardant Health, a liquid biopsy firm, made a $26 million strategic funding in Lunit to help the corporate’s innovation in precision oncology by means of the Lunit SCOPE tissue evaluation merchandise.
Accelerating Insights With NVIDIA AI
Lunit develops its AI fashions utilizing varied NVIDIA Tensor Core GPUs, together with NVIDIA A100 GPUs, within the cloud. Its clients can deploy Lunit’s AI with an NVIDIA GPU-powered server on premises or in the cloud — or inside a medical imaging machine utilizing the NVIDIA Jetson edge AI platform.
The corporate additionally makes use of NVIDIA TensorRT software program to optimize its educated AI fashions for real-world deployment.
“The objective right here is to optimize our AI in precise person settings — for the particular NVIDIA GPUs that function the AI,” mentioned Donggeun Yoo, chief of analysis at Lunit.
Over time, Lunit has introduced its work at NVIDIA GTC and as an NVIDIA Inception member on the prestigious RSNA convention for radiology.
“It was very useful for us to construct credibility as a startup,” mentioned Yoo. “I consider becoming a member of Inception helped set off the larger acknowledgements that adopted from the healthcare trade.”
Be part of the NVIDIA Inception neighborhood of over 10,000 expertise startups, and register for NVIDIA GTC, operating on-line Sept. 19-22, to attend the session “Accelerate Patient-Centric Innovation with Makers and Breakers in Healthcare Life Science.”
Subscribe to NVIDIA healthcare news.