PINPOINT: Integrating Computer Assisted Detection and Virtual Nodule Clinics for Early Lung Cancer Detection in Routine Practice, a Pilot Study

Background Promoting early detection of lung cancer is essential to improve survival and reduce treatment intensity. Incidental pulmonary nodules (IPN) detected on routine chest CT scans represent an opportunity for early diagnosis but are frequently missed or not managed appropriately. Computer-assisted detection (CAD) software and patient tracking through virtual nodule clinics (VNCs) may improve IPN detection and promote appropriate clinical management. Research question Is implementation of CAD combined with VNCs feasible and effective for improving IPN detection and management in routine clinical practice? Study design and methods A dedicated IPN program integrating CAD for analysis of all chest CT scans and a navigator-supervised VNC was implemented across 6 European hospitals. Aggregated clinical and radiological data were collected from the VNC application. Feasibility outcomes including CAD failure rate, nodule detection rate, patient flagging in the VNC, lung cancer diagnosis, and site-level adoption and continued software use, were evaluated using descriptive statistics. Results Data were collected from program initiation until January 2025, with a mean duration of 14 months (range 8-25). 65,344 patients were included. CAD software successfully analyzed 99.5% of scans (range 99.1–99.6%) and detected abnormalities in 62.7% (40,954) of scanned patients. Radiologists flagged 619 patients with IPN for management in the VNC, ten of which (1.6%) were diagnosed with lung cancer. All lung cancers were early-stage. CAD use was maintained in all sites, with radiologists reporting increased diagnostic confidence. The VNC was perceived as a valuable safety-net, although insufficient resources and workflow disruptions limited consistent use. Interpretation This pilot demonstrates feasibility of CAD and VNCs in routine workflows and informs future efficacy studies. CAD showed promising technical reliability and detection performance, and its sustained use suggests clinical acceptability. VNC uptake could be further optimized by supporting the navigator role with dedicated resources and improving ease of use.