Share this article and save a life!
Lung-RADS was never designed for AI. That’s the problem.
For years, we’ve used a 1-to-4 categorical scoring system to decide who needs follow-up after a chest CT. It works. But it was built for human radiologists reading films in isolation, not for AI systems processing thousands of scans with probabilistic outputs.
Here’s the tension nobody talks about:
Lung-RADS generates a false positive rate of 18% to 40+%. That sounds like a statistics footnote. It isn’t. In the National Lung Screening Trial, 1.7% of all screened patients underwent an invasive procedure for a lesion that turned out to be benign. Each unnecessary biopsy carries real clinical risk: pneumothorax, hemorrhage, infection. And real psychological cost: months of patient anxiety waiting on a result that was never cancer.
Scale that across 16 – 28 million eligible Americans and you’re not talking about edge cases. You’re talking about a structural flaw in how we triage.
And it cuts both ways. The same categorical system that over-refers on false positives misses roughly 10% of real cancers entirely.
AI changes the math.
Oatmeal Health’s LungAI outputs a continuous 0-100 malignancy probability score, not a category. At equivalent sensitivity to Lung-RADS, it reduces false positives by one-third: specificity improves from 82% to 88%. At equivalent specificity, it catches 3% more cancers Lung-RADS would miss. (Pre-FDA Cleareance) Launching soon.
That number travels with the patient into the EHR, into the care navigation workflow, and into the shared decision-making conversation.
Why does this matter for FQHCs and safety-net clinics?
Because the patients most likely to fall through the cracks are uninsured or Medicaid patients at community health centers. Their PCP is managing 2,000 patients. A nodule flagged for 3-month follow-up gets lost in the noise. A probability score of 7.2% routes differently than a Lung-RADS 3 label. High-probability findings get acted on. Lower-probability findings get appropriately scheduled. The nodule doesn’t disappear because a category label undersold the risk.
Stage I lung cancer has a 77% five-year survival rate. Stage IV has 9%. The difference is almost always a follow-up that happened or didn’t.
We keep building better detection tools. We’re still terrible at the follow-through.
The next frontier in lung cancer screening is not finding more nodules. It’s making sure the right people act on the ones we already find.
Are your workflows ready for a world where every chest CT comes back with a malignancy probability score, not just a Lung-RADS category?
♻️ Repost if AI should replace categorical nodule scoring with continuous probability scores before another Stage IV diagnosis happens at follow-up.
👉 Follow me for daily healthcare updates and get the deeper analysis in my free newsletter → https://lnkd.in/eJKFuB_p
Share this article and save a life!
Author:
CEO/Co-Founder @ Oatmeal Health | AI Lung Cancer Screening | Almost Became a Doctor | Engineer | Follow to Share What I’ve Learned Along the Way
I help patients get the care they need earlier, preventing late-stage cancer.
That’s been the throughline across three companies and almost 20 years in healthcare. At ReferralMD, we fixed broken referral networks so patients didn’t fall through the cracks. At Oatmeal Health, it’s lung cancer: building the diagnostic and screening infrastructure so the 85% of cases caught too late get caught early instead.
Today as CEO of Oatmeal Health, I lead a team embedding AI into radiology workflows to turn routine lung CT scans into reimbursable cancer risk assessments. We partner with FQHCs to reach underserved communities, and with health systems and payers to make early detection economically sustainable. Think HeartFlow or Cleerly, but for lungs.
Between companies, I advised at Techstars and Plug and Play, mentoring founders building in digital health. That experience shaped how I think about what separates companies that ship from companies that stall: distribution, reimbursement, and clinical trust, not just technology.
I’m a CancerX alumnus, a 3x healthcare founder, and someone who believes the biggest problems in cancer aren’t scientific. They’re operational.
We’re hiring mission-driven builders at Oatmeal Health. If you want to work on something that matters, reach out.
When I’m not working, I’m traveling, mentoring, and keeping up with one very energetic husky. 🐾
Substack – The Oatmeal Bite:
Millions of patients get less care because of who they are, where they live, or how they look. I’m fighting to change that. CEO @OatmealHealth, a startup built for the underserved. The Oatmeal Bite: intel for clinicians, investors, and advocates.
Jonathan Govette
CEO of Oatmeal Health
Substack:
https://oatmealhealthjonathangovette.substack.com/
