Oatmeal Health
Oatmeal Health is an AI-driven platform specialized in cancer screening, focusing primarily on underserved patient communities. The platform provides critical s...
Last verified:
What is Oatmeal Health?
Oatmeal Health is an AI-powered lung cancer screening and diagnostics platform that covers the entire lung cancer screening journey from patient identification to reimbursable diagnosis. The platform combines three core components: AI Patient Identification that analyzes EHR data to predict patients with 20+ pack-year smoking history even when undocumented, Care Navigation plus Shared Decision Making visits via telehealth with AI navigators helping FQHC nurse practitioners, and LungAI CADx which delivers automated malignancy risk assessments through native integration with PACS/CADe vendors.
The key feature is LungAI, a 100-million-parameter Vision Transformer that delivers a continuous 0-100 malignancy risk score per nodule, backed by a model trained on 100,000 CT scans. LungAI achieves 0.96 AUROC on NLST data compared to Lung-RADS 0.88, with 94% sensitivity and approximately 88% specificity. It integrates directly into existing CADe workflows with zero new clicks, no new workstation, no new login, and no IT project. The platform includes prior-scan comparison built in and longitudinal tracking.
Oatmeal Health is primarily for FQHCs (Federally Qualified Health Centers) and community health centers serving underserved Medicare/Medicaid populations, imaging centers and health systems looking to add reimbursable AI services, CADe technology partners wanting to embed their CADx layer, and life sciences/pharma companies accessing the FQHC network for clinical trial enrollment. The solution is reimbursable under CPT 0721T for hospital outpatient settings following FDA clearance, covering AI-assisted quantitative image analysis.
Oatmeal Health pricing
Pricing model: Free
Oatmeal Health operates on a zero-cost, zero-hassle model for providers with risk-based revenue sharing. The platform is potentially free of charge for FQHCs and patients, with expenses covered by CMS. Health centers and imaging can bill through the new AI CPT code 0721T at $650 per scan as an overread. Partnership economics are structured around volume and integration scope. No upfront costs or subscription fees are mentioned - the model focuses on reimbursement sharing from the billable CPT code.
Oatmeal Health pros
- 0.96 AUROC on NLST data vs Lung-RADS 0.88
- 94% sensitivity with approximately 88% specificity
- 100-million-parameter Vision Transformer model
- Trained on 100,000 CT scans
- Continuous 0-100 malignancy risk score per nodule
- Integrates directly into existing CADe workflow
- Zero new clicks required for radiologists
- No new workstation or login needed
- No IT project required for implementation
- Prior-scan comparison built in automatically
- Longitudinal tracking of nodules over time
- Reimbursable under CPT 0721T
- 9x faster patient identification than manual chart review
- 0.97 AUROC for EHR-based patient identification
- 91% scan agreement rate from care navigation
- Reduces false positives from up to 40% significantly
- Reduces false negatives from approximately 10%
- Addresses 30% inter-reader variability issue
- 1.5-3 minute read time per scan
- Brings tumor board-level insight to every radiologist
- Increases survival from 9% to 77-92% with early detection
- Avoids $600K+ in late-stage treatment costs
- 4-6x ROI from prevention for health plans
- Zero-cost zero-hassle model for providers
Oatmeal Health cons
- Still awaiting FDA clearance for full deployment
- Retrospective study results may differ in prospective use
- Only covers lung cancer screening not other cancers
- Requires existing CADe workflow integration
- Reimbursement only for hospital outpatient settings
- CPT 0721T reimbursement post-FDA clearance only
- Limited to patients with 20+ pack-year smoking history
- Primarily targets underserved FQHC populations
- One integration with CADe partner needed initially
Frequently asked questions about Oatmeal Health
What is LungAI and how does it work?
LungAI is a 100-million-parameter Vision Transformer CADx diagnostic tool that delivers a continuous 0-100 malignancy risk score per nodule. It integrates a 3D Vision Transformer nodule encoder, a Sybil-derived whole-scan encoder, and a dense decoder that fuses both with patient demographics and prior scan data to output per-nodule malignancy probability. The model is trained on 100,000 CT scans and achieves 0.96 AUROC on NLST data.
How does LungAI compare to Lung-RADS?
LungAI achieves 0.96 AUROC on NLST held-out set compared to Lung-RADS v2022 at 0.88. At the high-sensitivity operating point, LungAI achieved 94% sensitivity and approximately 88% specificity compared to 91% sensitivity and 82% specificity for Lung-RADS alone. Lung-RADS depends on size-based rules only without shape or texture AI, while LungAI analyzes shape, texture, density, and patient history.
Is LungAI reimbursable and what is the CPT code?
Yes, LungAI is reimbursable under CPT 0721T for hospital outpatient settings following FDA clearance. CPT 0721T covers AI-assisted quantitative image analysis. Health Centers and Imaging can bill $650 per scan as an overread. The reimbursement is available post-FDA clearance.
How does AI Patient Identification work?
The AI analyzes EHR data including notes, labs, imaging, and billing codes to predict patients with 20+ pack-year smoking history even when undocumented. It achieves 9x faster patient identification than manual chart review with 0.97 AUROC. This automatically surfaces eligible patients from the EHR without manual chart review by radiologists.
What is the care navigation process?
AI navigators help FQHC nurse practitioners conduct shared-decision making visits via telehealth. The system helps educate patients and drives a 91% scan agreement rate, feeding the diagnostic pipeline. This occurs between patient identification and the actual CT scan diagnosis step.
How long does it take to read a scan with LungAI?
LungAI reduces read time to 1.5-3 minutes per scan compared to 5-10 minutes for manual Lung-RADS reads. The automated malignancy risk assessment appears directly in the existing CADe viewer without workflow disruption.
Who is the primary target customer for Oatmeal Health?
The primary targets are FQHCs (Federally Qualified Health Centers) and community health centers serving underserved Medicare/Medicaid populations, imaging centers and health systems looking to add reimbursable AI services, CADe technology partners wanting to embed the CADx layer, and life sciences/pharma companies needing access to FQHC networks for clinical trials.
What false positive and false negative rates does LungAI address?
Current Lung-RADS screening has up to 40% false positive rate causing unnecessary biopsies and patient anxiety, and approximately 10% false negative rate where real cancers are missed until too late. LungAI achieves 94% sensitivity (reducing false negatives) and approximately 88% specificity (reducing false positives) compared to Lung-RADS's 91% sensitivity and 82% specificity.
How does LungAI integrate with existing workflows?
LungAI scores appear directly in the existing CADe viewer with zero new clicks. There is no new workstation, no new login, and no IT project required. One integration with a CADe partner unlocks their entire hospital network. The integration is native to existing CADe viewers.
What is the value proposition for patients?
Earlier detection increases survival from 9% to 77-92%. Fewer unnecessary biopsies mean less anxiety and fewer complications for patients. The platform ensures everyone, especially underserved populations, has access to life-saving early detection services with no burden of costs for essential cancer screening.