Oatmeal Health posted its first LungAI study on medRxiv, a retrospective validation on 2,747 held out NLST low dose CT exams with 89 confirmed cancers. LungAI reached an AUROC of 0.92, and combined with Lung-RADS v2022 it cut cancers sent to a 12 month rescan by 46 percent and benign biopsy referrals by 20 percent at roughly unchanged biopsy volume. The paper is a preprint, and its AUROC gain over Lung-RADS alone was not significant.
Our readRadiology and screening program buyers now have a public, citable performance record for the nodule risk score, including stated limits (single cohort, solid nodules only, vendor funded with author equity). LungAI remains investigational and not FDA cleared, so this is evidence for evaluation, not a purchasable US product.