A little over a month ago, 3D bioprinting startup Aether, based in San Francisco with a new University of California location at Johnson & Johnson Innovation, released its latest announcement. The startup, already well-known for its Aether 1 3D bioprinter, has entered the world of medical imaging software, and developed the Automatic Segmentation and Reconstruction (ASAR) process – one of the more exciting medical AI technologies – to power it.
This news was all the more exciting because Aether had no prior AI background. ASAR rapidly combines an array of AI and image processing techniques with adaptable deep learning models, so users can quickly transform raw medical images into segmented tissues and organs, then easily reconstruct them as digital 3D models without having to use calibration or editing tools.
This AI-powered software will increase productivity with this automatic segmentation, and advance the development of organ 3D printing; it’s also, as Aether CEO Ryan Franks told us, the “perfect complement” to a new method just published regarding how to turn already segmented medical images into 3D files.
“Aether’s process also segments organs like the brain pixel by pixel and converts them to 3D files within seconds, which will also include slicing. So it’s lightning quick, ” Franks told 3DPrint.com.
“Ours will produce stl files, allowing you to use a different bio-material for each stl, a requirement for printing organs and tissues with living cells.
“These stl files can be used for any tool type, allowing users to utilize Aether 1’s microvalves for liquids, syringes for viscous biomaterials like our new SynthBone UV curable hydroxyapatite, and complex fabrication techniques which also incorporate lasers and other tools.”
According to Aether, many large companies and brands in the medical field jumped on the bandwagon after the April announcement with their own new medical imaging software, but really, if it isn’t powered by AI, not much has changed – it’s still the same slow method of manual segmentation.
In the meantime, Aether has stayed on course with ASAR, and just announced several major advances and developments for its innovative new AI, including excellent Jaccard scores.
Franks explained to us, “Jaccard is the preferred metric for measuring accuracy of multi-class segmentation, especially with uneven amounts of each class, for example images with many more pixels of cortical gray matter, than pixels of brain stem.”
Aether’s models have already received extremely high scores on the Jaccard scale for each of these eight classes. All of the classes were able to produce Jaccard scores of 0.984 – 0.999, which is the equivalent of scoring 98.4% – 99.9%.
“While Aether AI is still in early training, our Jaccard accuracy scores have been higher than anything I’ve seen in the field,” Franks told 3DPrint.com. “We would love to use this groundbreaking new ASAR technique to help make MIT’s process faster, easier, and more scalable.”
Aether AI also has a new trick up its sleeve with multi-class segmentation, and was able to segment a human brain MRI into eight separate classes: background, cortical gray matter, basal ganglia, white matter, cerebrospinal fluid, ventricles, cerebellum, and the brainstem.
“Anyone, with zero training, can now convert raw unsegmented medical images into extremely complex multi-tool prints with a few simple clicks,” Franks told us.
Aether believes that ASAR’s capabilities are extremely far ahead of other software currently on the market, so it would make good business sense to charge a premium price for the technology. But the startup is more focused on developing a broad customer base, and has made the decision to “prioritize affordable access.” So Aether AI will be available for a much lower cost than the software it’s competing against, and when it’s combined with the advanced Aether 1 3D bioprinter and Aether’s advanced line of bioinks, users will have the world’s first integrated organ 3D printing system.
The ultimate goal of 3D bioprinting is to create replacement organs, and Aether AI is a big step in the right direction. It provides a major breakthrough in medical imaging, as the process makes it possible to analyze and edit organ and tissue types selected for segmentation in 2D and 3D environments, without being distracted by the surrounding anatomy.
By quickly and easily turning raw medical images into segmented, 3D printable files, soon doctors will be able to use patient CT and MRI images with the Aether 1 to create 3D printable organ replicas for the purposes of personalized medical devices, surgical training, and hopefully one day, implantable organs.
“It’s so absolutely unbelievable to see massive medical breakthroughs like these happening so fast,” Franks told us. “If you believe in the exponential growth of technology like Aether does, you can imagine the miraculous upcoming technologies that are right around the corner.”
Discuss this story and other 3D printing topics at 3DPrintBoard.com or share your thoughts in the Facebook comments below.[Source/Images: Aether]
You May Also Like
3D Printing News Briefs: February 26, 2020
In our 3D Printing News Briefs today, we’ve got a workshop and a partnership, along with student discounts for 3D modeling and students 3D printing parts for an electric vehicle....
EnvisionTEC-exocad Partnership Signals Industry Integration
For LMT Lab Day 2020, a number of 3D printing companies are rolling out their dental 3D printing product lines. This year, there’s the sense that there is a change...
New Partnership: BEGO’s Dental Materials Allow Formlabs Customers to 3D Print Crowns & Bridges
BEGO, headquartered in Germany, has been a leader in the dental field for 130 years—and as pioneers in 3D printing for the last two decades, they now specialize in all...
3D Printing News Briefs: February 12, 2020
In today’s 3D Printing News Briefs, we’ve got materials news, then some collaborations and an education story to share. First, KRAIBURG TPE is examining if its thermoplastic elastomers are suitable...
View our broad assortment of in house and third party products.