Flagging urgent suspicion of lung cancer with AI CXR algorithms: optimising feature selection and performance
Abstract
Objectives: To investigate the impact of feature selection and thresholding on the performance of a commercial chest X-ray (CXR) algorithm for flagging urgent suspicion of lung cancer (USC) in a West of Scotland population.
Materials and methods: The datasets comprised 2887 prospectively collected CXRs from general practitioners and outpatient routine care in the West of Scotland. An additional validation dataset included 983 retrospectively collected CXRs. The index test was an algorithm probability score; the reference standard was USC, derived from a blinded expert radiologist panel review of CXR. Three feature selection approaches were compared: radiologist-selected features with commercial thresholding of individual features, radiologist-selected features with regression weighting for the combined score, and a model using Akaike Information Criterion (AIC) and LASSO regression approaches for feature selection. Performance was evaluated at two thresholds: the Youden index and 15% positive flag rate.
Results: Regression-based feature selection using the AIC and LASSO approaches resulted in a four-feature model (consolidation, nodule, tuberculosis and cavity) compared with the five features (cavity, nodule, mass, hilar enlargement and mediastinal widening) selected by clinical experts. Sensitivity was highest in the AIC/LASSO model, but was significantly impacted by a 15% flag rate threshold. Regression weighting for the combined score demonstrated superior performance compared to individual feature thresholds.
Conclusion: In this study, regression-based feature refinement against an expert-derived reference standard outperformed clinically intuitive selection. Commercial CXR algorithms require local calibration and task-specific optimisation for USC detection. Workflow and post-market deployment considerations can be addressed by constraining positive flag rates through a unified model threshold.
Key Points: Question How should users select and combine AI chest X-ray algorithm features to achieve optimal deployment performance for detecting urgent suspicion of cancer in routine practice?
Findings Regression-based feature selection outperformed expert-selected features, and combining features under a single regression-weighted threshold outperformed individual feature thresholds for urgent cancer suspicion detection.
Critical relevance statement Optimising chest X-ray AI for urgent cancer suspicion requires data-driven, regression-based feature selection rather than clinical intuition; health systems therefore need local capability to validate and calibrate algorithms for their population and task.
Related
- Performance of Chronic Conditions Data Warehouse algorithms in Medicare Advantage and fee-for-service populations
- Performance of Claims-based Alzheimer’s Disease and Related Dementias Algorithms in Medicare Advantage and Fee-for-Service Populations
- Comprehensive comparison of six optimization algorithms for automated population pharmacokinetic model selection via two simulation studies and eleven real case studies
- Comparative Evaluation of Feature Selection Algorithms for Predicting Radiation-Induced Complications in Head and Neck Cancer Using Multimodal CT and MRI Radiomics
- Preoperative CT Suspicion of Main Pancreatic Duct Injury in Isolated Blunt Pancreatic Trauma: A Case Report With Operative Correlation
BibTeX
@article{duncan2026flagging,
title = {{Flagging urgent suspicion of lung cancer with AI CXR algorithms: optimising feature selection and performance}},
author = {Sean F. Duncan and Katie Dudman and Shalom Odigbo and Jesus Perdomo Lampignano and Shamie Kumar and Neil Hawkins and Alex McConnachie and John D. Maclay and David B. Stobo and Mark Hall and David J. Lowe},
journal = {Insights into Imaging},
year = {2026},
month = oct,
doi = {10.1186/s13244-026-02438-7},
url = {https://link.springer.com/article/10.1186/s13244-026-02438-7},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Flagging urgent suspicion of lung cancer with AI CXR algorithms: optimising feature selection and performance
AU - Sean F. Duncan
AU - Katie Dudman
AU - Shalom Odigbo
AU - Jesus Perdomo Lampignano
AU - Shamie Kumar
AU - Neil Hawkins
AU - Alex McConnachie
AU - John D. Maclay
AU - David B. Stobo
AU - Mark Hall
AU - David J. Lowe
PY - 2026
DA - 2026/10/31
JO - Insights into Imaging
DO - 10.1186/s13244-026-02438-7
UR - https://link.springer.com/article/10.1186/s13244-026-02438-7
AB - Objectives: To investigate the impact of feature selection and thresholding on the performance of a commercial chest X-ray (CXR) algorithm for flagging urgent suspicion of lung cancer (USC) in a West of Scotland population. Materials and methods: The datasets comprised 2887 prospectively collected CXRs from general practitioners and outpatient routine care in the West of Scotland. An additional validation dataset included 983 retrospectively collected CXRs. The index test was an algorithm probability score; the reference standard was USC, derived from a blinded expert radiologist panel review of CXR. Three feature selection approaches were compared: radiologist-selected features with commercial thresholding of individual features, radiologist-selected features with regression weighting for the combined score, and a model using Akaike Information Criterion (AIC) and LASSO regression approaches for feature selection. Performance was evaluated at two thresholds: the Youden index and 15% positive flag rate. Results: Regression-based feature selection using the AIC and LASSO approaches resulted in a four-feature model (consolidation, nodule, tuberculosis and cavity) compared with the five features (cavity, nodule, mass, hilar enlargement and mediastinal widening) selected by clinical experts. Sensitivity was highest in the AIC/LASSO model, but was significantly impacted by a 15% flag rate threshold. Regression weighting for the combined score demonstrated superior performance compared to individual feature thresholds. Conclusion: In this study, regression-based feature refinement against an expert-derived reference standard outperformed clinically intuitive selection. Commercial CXR algorithms require local calibration and task-specific optimisation for USC detection. Workflow and post-market deployment considerations can be addressed by constraining positive flag rates through a unified model threshold. Key Points: Question How should users select and combine AI chest X-ray algorithm features to achieve optimal deployment performance for detecting urgent suspicion of cancer in routine practice? Findings Regression-based feature selection outperformed expert-selected features, and combining features under a single regression-weighted threshold outperformed individual feature thresholds for urgent cancer suspicion detection. Critical relevance statement Optimising chest X-ray AI for urgent cancer suspicion requires data-driven, regression-based feature selection rather than clinical intuition; health systems therefore need local capability to validate and calibrate algorithms for their population and task.
SN - 1869-4101
ER -