The AI quietly reading the scans nobody ordered a test for - turning a hospital's routine CTs into early warnings.
Somewhere right now, a chest CT is being read for a cracked rib, and on that same image - already captured, already paid for, already archived - sits a quiet white smear of calcium around a heart valve. Nobody ordered a test for it. Most of the time, nobody looks. That overlooked smear is the entire reason Bunkerhill Health exists.
The company calls it opportunistic screening, which is a polite way of saying: the data was already there, we just bothered to read it. Bunkerhill builds FDA-cleared AI that scours routine imaging for the diseases hiding in plain sight - coronary calcium, aortic valve calcium, thinning bone, a widening aorta. The scan was for something else. The finding could save a life. That is the whole, deceptively simple pitch.
Bunkerhill Health is building the system of action for health systems.
It is a useful phrase, because most healthcare AI stops at the verb identify. It flags a problem and hands you a notification. Bunkerhill's argument is that identifying isn't the hard part anymore - acting is. Closing the loop from a finding on a screen to a follow-up appointment, a prior authorization, a referral that actually happens. That is the bunker the company is trying to take.
Bunkerhill grew out of Stanford's Center for AI in Medicine and Imaging - AIMI, the kind of place where brilliant models are trained and then, too often, parked in a paper. Co-founders Nishith Khandwala and David Eng were researchers there. Khandwala had built an algorithm to detect arterial blockages.
Then in 2020, his father had a heart attack. The irony was not lost on him: the technology to flag exactly that kind of risk existed, sat on his own hard drive, and was nowhere near a real clinic. The gap between what AI could do and what hospitals actually used stopped being an academic problem. It became the company.
Most solutions either identify problems without solving them, or address only narrow use cases. We deliver end-to-end AI workflows that drive real clinical impact.
Carebricks is the platform; the FDA-cleared algorithms are the proof. Each one reads a scan ordered for some other reason and returns a finding a clinician can act on.
Lets health systems create, share, and deploy AI agents for clinical and operational tasks - from actionable findings to prior authorization to referral prioritization, wired into the EHR.
First and only AI cleared to detect and quantify coronary artery calcium and aortic valve calcium on contrast-enhanced, non-gated routine chest CT.
First FDA-authorized AI to identify mitral annular calcification, a marker of cardiovascular risk, on routine CT scans.
First AI cleared to automate abdominal aortic diameter quantification on both contrast and non-contrast CT.
Assesses bone mineral density from spinal structures on non-contrast abdominal CT for adults 30+ - opportunistic osteoporosis detection.
Sold to hospitals as cleared algorithms plus the Carebricks agent layer - increasingly with a reimbursement path as CMS adds billing codes for AI analysis.
Roughly $30 million raised and about 29 people - a small machine that ships regulatory firsts. The names behind it read like a who's-who of the venture and healthcare worlds.
Bars indicate participation, not relative investment amounts. Figures approximate, drawn from public reporting.
Secured a national CMS reimbursement pathway and FDA clearance for the first AI algorithms to evaluate coronary and aortic valve calcium on contrast-enhanced chest CTs.
FDA clearance for Bunkerhill MAC - the first-ever AI to detect mitral annular calcification on routine CT.
FDA clearances for Bunkerhill AAQ (abdominal aortic diameter) and Bunkerhill BMD (bone mineral density).
Partnership with UTMB Health to advance clinician-led AI at enterprise scale.
Bunkerhill's customers are U.S. hospitals and health systems - the organizations sitting on archives of unread diagnostic data. Featured names include UTMB Health, HCA Florida Healthcare, MedStar Health, Endeavor Health, WVU Medicine, and Stamford Health.
For a clinician, the appeal is leverage: catch the osteoporosis, the calcium, the widening aorta now - on a scan the patient already had - instead of finding out at the next emergency. For a health system, it is unread value turned into measured outcomes, and with the new CMS code, into reimbursable ones.
Profile compiled from public sources including the company website, Y Combinator, and industry press. Funding and team figures are approximate and may have changed since publication.
Return to the chest CT for the cracked rib, the one with the quiet smear of calcium nobody ordered a test for. In the old world, that image gets read for the rib, signed off, and filed. The smear waits, patiently, for the day it stops being a smear and becomes an emergency.
In Bunkerhill's world, an algorithm reads it on the way past, quantifies it, and hands a clinician something to do about it - before the emergency. Same scan, same patient, same machine. The only thing that changed is that someone finally bothered to look. That, in the end, is what a company founded on a missed warning is built to fix.