Quantitative CT reconstruction kernel harmonization for multi-site lung cancer screening

用于多部位肺癌筛查的定量CT重建核协调

阅读:1

Abstract

BACKGROUND: Published reference CT protocols for lung cancer screening are not optimized to produce uniform image appearance across different scanner manufacturers and models or to conform to quantitative imaging profiles for robust small lung nodule size and volume measurements, which are important in clinical management of screen-detected nodules. PURPOSE: This study used widely available phantoms and software to identify lung cancer screening CT reconstructions that enable accurate and reproducible nodule size measurements and to match reconstructions across scanner manufacturers and models to provide a consistent image appearance to interpreting physicians. METHODS: ACR CT accreditation phantom scans were used to measure the modulation transfer function (MTF) and noise power spectrum (NPS) for various reconstruction kernels for six CT scanner models from three manufacturers. A reference kernel was chosen, and other kernels were matched based on agreement between MTF and NPS for each candidate kernel. Kernels were validated for conformance to the Quantitative Imaging Biomarkers Alliance (QIBA) Small Lung Nodule profile using a commercial phantom and conformance testing software. RESULTS: Medium sharp kernels with similar MTF and NPS curves and QIBA-compliant imaging performance were identified for the six scanner models. One medium smooth kernel with iterative reconstruction and one common lung kernel did not conform to the QIBA profile. CONCLUSIONS: MTF and NPS comparisons to a validated reference CT protocol can be used to identify candidate scanner-specific reconstructions appropriate for lung cancer screening that conform to the QIBA protocol, thus producing accurate and reproducible nodule size measurements, and that should provide uniform image appearance to the interpreting radiologist.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。