AVIEW Lung Nodule CAD
Computer-Aided Detection (CAD) software designed to assist radiologists in the detection of pulmonary nodules (diameter 3–20 mm) during the review of CT examinations of the chest for asymptomatic populations. Adjunctive second-reader after the radiologist's initial read. Validated using non-contrast chest CT.
Regulatory
| Regulatory status | 510(k) cleared |
| FDA submission pathway | 510(k) |
| FDA decision date | 2025-12-03 |
| FDA 510(k) number | K251203 |
| CE marked | Yes |
Validation summary
| Study design | Retrospective |
| Cohort size | 151 patients; 282 samples — MRMC reader study (K221592): 151 chest CTs (103 negative controls + 48 cases with one or more lung nodules); 11 board-certified radiologists, blinded, 4-week washout between unaided and AI-aided reads. Standalone study: 282 chest CTs (140 with nodule data + 142 without) from three geographically distinct US clinical sites; gender split 132M/150F. K251203 (2025) is a UI-only minor change with no new clinical study. |
| Number of validation sites | 3 |
| Site geography | Multi-center (US) |
| Comparator | Pathologist consensus |
| Primary endpoint | AUC |
| Primary result | MRMC: AUC 0.73 (unaided, 95% CI 0.66–0.79) vs 0.92 (aided, 95% CI 0.89–0.95), Δ=0.19; Sensitivity 0.68 (unaided, 95% CI 0.62–0.73) vs 0.91 (aided, 95% CI 0.89–0.94), Δ=0.23; FP/scan 0.48 (unaided, 95% CI 0.28–0.69) vs 0.28 (aided, 95% CI 0.15–0.42), Δ improvement 0.24. Standalone: AUC 0.961 (95% CI 0.939–0.983); Sensitivity 0.907 (95% CI 0.846–0.95); Specificity 0.704 (95% CI 0.622–0.778); Sensitivity at FP/scan<2: 0.889 (95% CI 0.849–0.93) at FP/scan=0.504. |
| External validation | Yes — Standalone test set sourced from three geographically distinct US clinical sites; anonymized medical data purchased separately from any training or internal validation data per K221592 Decision Summary. Both screening and incidental populations included. · Subgroup analyses performed for challenging/confounding cases per K221592; consistent performance. |
| Peer-reviewed | Yes |
| FDA decision summary | View on accessdata.fda.gov |
Key publications
- Outstanding negative prediction performance of solid pulmonary nodule volume AI for ultra-LDCT baseline lung cancer screening risk stratificationLung Cancer · 2022
Limitations noted
[Source-detail preserved: Study design: Multi-center MRMC (multi-reader, multi-case); Primary endpoint (original): AUC_aided_vs_unaided] (i) Nodule diameter range 3–20 mm — does not cover sub-3mm or >20mm nodules; (ii) validated using non-contrast CT only; (iii) MRMC cohort enriched (48 positives / 103 negatives) is not population-prevalence representative; (iv) K251203 is a UI-only change — no new clinical data submitted; (v) Lung Cancer 2022 paper is an independent real-world Moscow Lung Cancer Screening evaluation, not the pivotal trial.
Deployment
| Available | Yes |
Sources
- https://www.accessdata.fda.gov/cdrh_docs/pdf25/K251203.pdf · fda_db · 2026-05-12
- https://www.accessdata.fda.gov/cdrh_docs/pdf22/K221592.pdf · fda_db · 2026-05-12
Elbakri A. OncologyAI Registry: AVIEW Lung Nodule CAD. Available at: https://oncologyairegistry.org/tools/coreline-aview-lung-nodule.html. Accessed [date].