ECR 2026: Poster Presentation
Edgar Wakelin, Jeroen Mollink, Marios Koulakis, Maurits Engbersen, Bryan Haslam
Abstract
Purpose
Despite lung cancer being a leading cause of mortality in Europe, eligibility requirements for screening are complex and region specific, limiting adoption rates. Artificial intelligence (AI) that identifies lung nodules in routine CT scans could help detect lung nodules that would otherwise be missed, but it is unclear if the same AI can be used for both screening and routine CTs given the different patient populations scanned. Here we characterize the performance of an AI algorithm to detect lung nodules within routine chest CT scans and compare to screening low dose CT (LDCT).
Methods and Materials
A total of 310 CT scans (143 LDCT (73 with nodules, 70 without) and 167 routine (92 with nodules, 75 without)) were included. Five thoracic radiologists segmented all nodules measuring 4–30mm, with majority voting (≥3/5 rads) used to define the ground-truth. AI performance was evaluated using: FROC analysis, reporting the average sensitivity across false-positive rates (from 0.125 to 8.0 false positives per scan (FPPS)); nodule sensitivity; and FPPS at the selected operating point.
Results
Across the 310 CT scans, 330 nodules were identified. The median nodule diameter was 6.0±4.71mm. All scans were successfully processed by the AI and produced similar detection performance between algorithms (FROC sensitivity: LDCTs: 0.732(95%CI:0.672-0.795), routine: 0.791(95%CI: 0.743-0.837)). At the device operating point the nodule sensitivity was also similar (routine: 0.811(95%CI:0.754-0.867), LDCT: 0.793(95%CI:0.713-0.866)) and the FPPS was similar (routine: 0.894(95%CI:0.700-1.114), LDCTs: 1.013(95%CI:0.703-1.388)).
Conclusion
The AI model successfully identified lung nodules within routine chest CT scans with high sensitivity and low FPPS showing similar performance to screening LDCT exams. This performance highlights the potential for using the same AI to support early lung cancer detection for both screening LDCT and routine CT scans.
Disclaimers:
Disclaimers: Chest Suite comprises multiple applications including Veye Lung Nodules, Veye Reporting, DeepHealth Lung AI, DeepHealth Viewer and HealthCCSng. Veye Lung Nodules and Veye Reporting are manufactured by Aidence B.V. and distributed by DeepHealth, Inc. Neither Veye Lung Nodules nor Veye Reporting are FDA cleared for distribution in the US. DeepHealth Lung AI is FDA 510(k) pending. DeepHealth Viewer is manufactured by eRAD, Inc. and distributed by DeepHealth, Inc. in the US. HealthCCSng is manufactured by Nanox AI, Ltd. and distributed by DeepHealth, Inc. DeepHealth Lung AI is 510(k) pending. Any claims made about Chest Suite may reference claims associated with its individual components. Not all products and functionalities are commercially available in all countries
Any claims made about Chest Suite may reference claims associated with its individual components.
Not all products and functionalities are commercially available in all countries. For clearance and commercial availability in your geography of functionalities listed and compatibility with other systems, please contact your account manager.