Harmonizing the National Lung Screening Trial: Kernel resolution and field of view compensation
9 affiliations
- 1Department of Electrical and Computer Engineering, Vanderbilt University, Nashville, Tennessee, United States of America
- 2Department of Computer Science, Vanderbilt University, Nashville, Tennessee, United States of America
- 3Department of Biomedical Engineering, Vanderbilt University, Nashville, Tennessee, United States of America
- 4Insitro, South San Francisco, California, United States of America
- 5Department of Radiology and Radiological Sciences, Vanderbilt University Medical Center, Nashville, Tennessee, United States of America
- 6Division of Allergy, Pulmonary and Critical Care Medicine, Department of Medicine, Vanderbilt University Medical Center, Nashville, Tennessee, United States of America
- 7Department of Medicine, Vanderbilt University Medical Center, Nashville, Tennessee, United States of America
- 8Department of Thoracic Surgery, Vanderbilt University Medical Center, Nashville, Tennessee, United States of America
- 9Vanderbilt University Institute of Imaging Science, Vanderbilt University Medical Center, Nashville, Tennessee, United States of America
Abstract
Low-dose chest computed tomography scans acquired in multi-centre studies exhibit heterogeneity in the form of varying spatial resolution, reconstruction kernel and semantic field-of-view that hinder the reproducibility of quantitative measures and generalizability of downstream analyses. The National Lung Screening Trial (NLST) provides a large, open collection of low-dose chest CT scans characterized by heterogeneous spatial resolution, reconstruction kernels and field-of-view configurations. This diversity presents an opportunity to standardize the data, enabling reproducible biomarker estimation and broader reuse of the NLST as a homogeneous quantitative imaging resource. In this work, we present a standardization pipeline that was applied to 105,191 scans from 23,669 participants in the NLST. The pipeline addresses the following elements: (1) Resampling through-axial-plane (z) spatial resolution to 1.0 mm (2) Harmonization of reconstruction kernels across intra-vendor (paired) and inter-vendor (unpaired) settings using an anatomy guided multipath cycleGAN where scans were mapped to a reference Siemens B30f (soft) and Siemens B50f (hard) kernel (3) Semantic field-of-view extension across all scans. Quantitative consistency across percent emphysema and body composition assessment (skeletal muscle and subcutaneous adipose tissue volumes) was evaluated in paired, cross-sectional and longitudinal settings. For paired data, harmonization improved agreement of quantitative measures relative to the soft kernel reference. Cross-sectional analyses demonstrated reduced inter-kernel variability, with small to medium effect sizes (Cohen’s d = 0.2 to 0.5) for percent emphysema and improved correlations with anthropometric measures post field-of-view extension. Longitudinal analysis using linear mixed effect models, showed reduced kernel bias and improved longitudinal consistency of biomarkers reflected by the increase in intraclass correlation coefficients. This work provides a harmonized NLST imaging resource that enables consistent and reliable quantitative biomarker estimation for lung cancer screening, with open source code released at https://github.com/MASILab/Harmonized_NLST_dataset and the harmonized dataset to be made publicly available once data return to the NLST is completed.
Related
- Harmonizing national drug lists: an eight-pattern ATC/DDD analysis
- High-Resolution Esophageal Manometry in Mongolia: The First National Experience and Diagnostic Spectrum of Esophageal Motility Disorders
- Twenty-Five Years of Colorectal Cancer Screening in the Czech Republic: National Registry Trends in Program Performance and Cancer Burden
- Interval colorectal cancer risk by haemoglobin concentration in Australia's National Bowel Cancer Screening Program
- Transporting Lung Cancer Screening Effects From National Lung Screening Trial to a US Screening-Eligible Population
BibTeX
@article{krishnan2026harmonizing,
title = {{Harmonizing the National Lung Screening Trial: Kernel resolution and field of view compensation}},
author = {Aravind R. Krishnan and Michael E. Kim and Chenyu Gao and Gaurav Rudravaram and Trent Schwartz and Andre T. S. Hucke and Lucas W. Remedios and Thomas Z. Li and Kaiwen Xu and Kim L. Sandler and Yency Forero Martinez and Sanja Antic and Fabien Maldonado and Lianrui Zuo and Bennett A. Landman},
journal = {PLOS One},
year = {2026},
month = oct,
doi = {10.1371/journal.pone.0352075},
url = {https://doi.org/10.1371/journal.pone.0352075},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Harmonizing the National Lung Screening Trial: Kernel resolution and field of view compensation
AU - Aravind R. Krishnan
AU - Michael E. Kim
AU - Chenyu Gao
AU - Gaurav Rudravaram
AU - Trent Schwartz
AU - Andre T. S. Hucke
AU - Lucas W. Remedios
AU - Thomas Z. Li
AU - Kaiwen Xu
AU - Kim L. Sandler
AU - Yency Forero Martinez
AU - Sanja Antic
AU - Fabien Maldonado
AU - Lianrui Zuo
AU - Bennett A. Landman
PY - 2026
DA - 2026/10/05
JO - PLOS One
DO - 10.1371/journal.pone.0352075
UR - https://doi.org/10.1371/journal.pone.0352075
AB - Low-dose chest computed tomography scans acquired in multi-centre studies exhibit heterogeneity in the form of varying spatial resolution, reconstruction kernel and semantic field-of-view that hinder the reproducibility of quantitative measures and generalizability of downstream analyses. The National Lung Screening Trial (NLST) provides a large, open collection of low-dose chest CT scans characterized by heterogeneous spatial resolution, reconstruction kernels and field-of-view configurations. This diversity presents an opportunity to standardize the data, enabling reproducible biomarker estimation and broader reuse of the NLST as a homogeneous quantitative imaging resource. In this work, we present a standardization pipeline that was applied to 105,191 scans from 23,669 participants in the NLST. The pipeline addresses the following elements: (1) Resampling through-axial-plane (z) spatial resolution to 1.0 mm (2) Harmonization of reconstruction kernels across intra-vendor (paired) and inter-vendor (unpaired) settings using an anatomy guided multipath cycleGAN where scans were mapped to a reference Siemens B30f (soft) and Siemens B50f (hard) kernel (3) Semantic field-of-view extension across all scans. Quantitative consistency across percent emphysema and body composition assessment (skeletal muscle and subcutaneous adipose tissue volumes) was evaluated in paired, cross-sectional and longitudinal settings. For paired data, harmonization improved agreement of quantitative measures relative to the soft kernel reference. Cross-sectional analyses demonstrated reduced inter-kernel variability, with small to medium effect sizes (Cohen’s d = 0.2 to 0.5) for percent emphysema and improved correlations with anthropometric measures post field-of-view extension. Longitudinal analysis using linear mixed effect models, showed reduced kernel bias and improved longitudinal consistency of biomarkers reflected by the increase in intraclass correlation coefficients. This work provides a harmonized NLST imaging resource that enables consistent and reliable quantitative biomarker estimation for lung cancer screening, with open source code released at https://github.com/MASILab/Harmonized_NLST_dataset and the harmonized dataset to be made publicly available once data return to the NLST is completed.
SN - 1932-6203
ER -