For any company building medical AI, independent peer review is one of the most important milestones there is. We're proud to share that Dinmar Labs' first validation study — conducted with our U.S. university partner — has been accepted for publication in a peer-reviewed journal.
What we studied
The study evaluated our CT lung nodule detection model on a large, independent, retrospective dataset from our partner institution — data our model had never seen during development. Radiologists read each study twice, separated by a washout period: once unaided and once with Dinmar AI assistance.
Key findings
- Higher sensitivity: AI-assisted reads detected more confirmed nodules than unaided reads, with the largest gains for small nodules.
- Stable specificity: The improvement in detection came without a meaningful increase in false positives.
- Faster reading: Average reading time decreased with AI assistance, as automated measurement replaced manual steps.
- Less variability: Agreement between readers increased when AI assistance was available.
"Peer review is how science separates promise from proof. We are proud to have our work examined — and validated — by independent experts."
— Roger Dinmar, Founder & CEOLimitations
As the authors note, this was a retrospective study at a single institution. Prospective, multi-site evidence is essential — which is exactly what our ongoing clinical trial with both university partners is designed to provide.
Request the paper
Researchers, clinicians and investors can request a copy of the paper and supplementary materials through our contact page.
Dinmar Labs Closes Its First Funding Round