Repeatability of real world, non-research chest CT scan-based lung density metrics

Quantitative CT (QCT) is carried out with attention to quality control (scanner make and model, calibration, lung volume and acquisition protocol), and bears financial and radiation cost. We wished to determine if non-research CT scans without such controls could yield reproducible data. 62 subjects (53 COPD, 6 asthma, 2 micronodules, 1 sarcoidosis) from a community respirology practice had had 2 non-contrast CT scans judged free of significant infiltrates, performed on 5 models of scanner in 9 different community hospitals for clinical indications within 14 months and had available spirometry and lung volumes performed respecting ATS criteria within 13 months of CT scans. Images were analyzed with AirwayInspector (airwayinspector.acil-bwh.org) for LAA<-950HU, lung density (LD) at 15th percentile + 1000HU and total lung volume (TLV). Means, Bland-Altman analysis and intraclass correlation coefficients were determined for TLV, LAA, LD and LD corrected for both predicted and measured TLC. Differences were determined with Student t-tests. Significance was set at p<05. Results are shown in Table 1.Real-world CT scans, if properly selected, can yield reproducible QCT data. Correcting LD with PFT measured TLC improves reproducibility more than correcting with predicted TLC. If validated at other centres, these findings suggest the pool of observational QCT data could be vastly expanded at little dollar and no radiation cost.
Source: European Respiratory Journal - Category: Respiratory Medicine Authors: Tags: Imaging Source Type: research