Logo image
Back
C108-10 Development of a Standard Field of View for Isotropic Voxel Quantitative Lung Imaging Utilizing Ultra-High-Resolution Photon Counting Computed Tomography
Abstract   Peer reviewed

C108-10 Development of a Standard Field of View for Isotropic Voxel Quantitative Lung Imaging Utilizing Ultra-High-Resolution Photon Counting Computed Tomography

J J Kitzmann, J Atha, K Schroeder, K Knoernschild, E A Hoffman and J C Sieren
American journal of respiratory and critical care medicine, Vol.212(Supplement_1), aamag1625465
05/01/2026
DOI: 10.1093/ajrccm/aamag162.5465

View Online

Abstract

Rationale Metrics associated with quantitative Computed Tomographic (qCT) lung imaging have historically been biased by patient body size due to variability in voxel spacing associated with in-plane field of view (FOV). This resolution variability impacts airway, vascular, and parenchymal density and dimension measurements. A significant limitation has been the 512x512 matrix size of energy integrating CT systems. With the introduction of photon counting detector CT (PCD-CT), improvements in spatial resolution (0.2mm) along with 1024x1024 matrix images create potential for consistent, high-resolution, isotropic voxel spacing across all patients. We hypothesize that, with the expanded matrix size, a universal FOV for qCT is now possible, minimizing inter-subject bias. Methods Chest CT scans from 904 COPDGene participants imaged at University of Iowa Health Care, ranging from never-smokers to GOLD stage 4 COPD, were analyzed [1]. Full thoracic anatomy was segmented automatically using TotalSegmentator [2] which created a lung mask containing all lung lobes and a bone mask containing the T2-C7 vertebrae, costal cartilage, and ribs. A minimum bounding-box was fit to the masked image targeting the bone mask for left-right (lateral) and anterior-posterior (AP) bounding and the lung mask for transverse bounding. Bounding-box dimensions were used as a proxy for thoracic cavity size. Dimensions were divided by 1024 to calculate the minimum voxel spacing required for full coverage. Data were analyzed and plotted using GraphPad Prism. Results The maximum lateral dimension, which represented the limiting factor, was 385mm. For a 512x512 matrix this would correspond to 0.76mm voxel spacing, while a 1024x1024 matrix allows 0.39mm voxel spacing. Lung dimension scaled linearly with height and weight but showed no correlation with COPD severity. Males had larger thoracic dimensions than females, with a lateral average of 325.8mm (max 385.1), compared to 281.7mm (max 341.0). Conclusion We propose a 0.4mm isotropic voxel size for ultra-high resolution qCT lung imaging with PCD-CT to ensure consistent measurements across diverse body sizes. While finer resolutions (e.g. 0.32mm) could accommodate most patients, stratifying resolution by sex or BMI may introduce new biases. A universal voxel size enabled by the improved spatial resolution of PCD-CT and a 1024x1024 matrix size may standardize qCT metrics and improve cross-subject comparability in lung imaging studies. References: [1]-Regan EA et al. COPDGene® 2019: Redefining the diagnosis of chronic obstructive pulmonary disease. Chronic Obstr Pulm Dis (Miami). 2019;6(5):384-399. doi:10.15326/jcopdf.6.5.2019.0149 [2]-Wasserthal J et al. TotalSegmentator: Robust Segmentation of 104 Anatomic Structures in CT Images. Radiology: Artificial Intelligence. 2023;5(5):e230024. doi:10.1148/ryai.230024 This abstract is funded by: NIH 1S10OD034285-01A1
 Chronic obstructive pulmonary disease

Details

Metrics

1 Record Views
Logo image