In the event you’ve ever had a PET scan, you understand it’s an ordeal. The scans assist medical doctors detect most cancers and observe its unfold, however the course of itself is a logistical nightmare for sufferers.
It begins with fasting for 4 to 6 hours earlier than coming into the hospital — and good luck to you in the event you dwell rurally and your native hospital doesn’t have a PET scanner. If you get to the hospital, you’re injected with radioactive materials, after which you will need to wait an hour whereas it washes by your physique. Subsequent, you enter the PET scanner and have to aim to lie nonetheless for half-hour whereas radiologists purchase the picture. After that, it’s a must to maintain bodily away from the aged, younger individuals, and pregnant ladies for as much as 12 hours since you’re actually semi-radioactive.
One other bottleneck? PET scanners are concentrated in main cities as a result of their radioactive tracers should be produced in close by cyclotrons — compact nuclear machines — and used inside hours, limiting entry in rural and regional hospitals.
However what in the event you may use AI to transform CT scans, that are way more accessible and inexpensive, into PET scans? That’s the pitch of RADiCAIT, an Oxford spinout that got here out of stealth this month with $1.7 million in pre-seed financing. The Boston-based startup, which is a Prime 20 finalist in Startup Battlefield at TechCrunch Disrupt 2025, has simply opened a $5 million elevate to advance its medical trials.
“What we really do is we took the most constrained, complex, and costly medical imaging solution in radiology, and we supplanted it with what is the most accessible, simple and affordable, which is CT,” Sean Walsh, RADiCAIT’s CEO, informed TechCrunch.
RADiCAIT’s secret sauce is its foundational mannequin — a generative deep neural community invented in 2021 on the College of Oxford by a group led by the startup’s co-founder and chief medical info officer, Regent Lee.
The mannequin learns by evaluating CT and PET scans, mapping them, and choosing out patterns in how they relate to one another. Sina Shahandeh, RADiCAIT’s chief technologist, describes it as connecting “distinct physical phenomena” by translating anatomical construction into physiological perform. Then the mannequin is directed to pay further consideration to particular options or features of the scans, like sure forms of tissue or abnormalities. This centered studying is repeated many occasions with many alternative examples, so the mannequin can establish which patterns are clinically vital.
Techcrunch occasion
San Francisco
|
October 27-29, 2025
The ultimate picture that goes to medical doctors for evaluation is created by combining a number of fashions working collectively. Shahandeh compares the strategy to Google DeepMind’s AlphaFold, the AI that revolutionized protein construction prediction: Each methods be taught to translate one sort of organic info into one other.
Walsh claims the group at RADiCAIT can mathematically show that their artificial or generated PET photographs are statistically just like actual chemical PET scans.
“That’s what our trials show,” he mentioned, “that the same quality of decision has been made when the doctor, radiologist, or oncologist is given a chemical PET or [our AI-generated PET].”
RADiCAIT doesn’t promise to interchange the necessity for PET scans in particular therapeutic settings, like radioligand remedy, which kills most cancers cells. However for diagnostic, staging, and monitoring functions, RADiCAIT’s know-how may make PET scans out of date.
“Because it’s such a constrained system, there’s not enough supply to meet demand for diagnostics and theragnostics,” Walsh mentioned, referring to a medical strategy that mixes diagnostic imaging (i.e., PET scans) with focused remedy to deal with ailments (i.e., most cancers). “So what we’re looking to do is soak up that demand on the diagnostic side. PET scanners themselves should pick up the slack on the theragnostic side.”
RADiCAIT has already begun medical pilots particularly for lung most cancers testing with main well being methods like Mass Common Brigham and UCSF Health. The startup is now pursuing an FDA medical trial — a costlier and concerned course of that’s driving RADiCAIT’s $5 million seed spherical. As soon as that’s permitted, the subsequent step will likely be to do business pilots and exhibit the business viability of the product. RADiCAIT additionally desires to run the identical course of — medical pilots, medical trials, business pilots — for colorectal and lymphoma use circumstances.
Shahandeh mentioned RADiCAIT’s strategy to utilizing AI to yield legitimate insights with out the burdens of inauspicious and costly checks is “broadly applicable.”
“We’re exploring extensions across radiology,” Shahandeh added. “Expect to see similar innovations linking domains from materials science to biology, chemistry, and physics wherever nature’s hidden relationships can be learned.”
If you wish to hear extra about RADiCAIT be a part of us at Disrupt, October 27 to 29 in San Francisco. Study extra right here.
