Which is the optimal threshold for defining functional lung in single-photon emission computed tomography lung perfusion imaging of lung cancer patients?. Issue 2 (February 2018)

• Record Type: Journal Article
• Authors: Ding, Xiuping; Man, Xianfeng; Sun, Mingping; Xing, Jun; Fan, Tingyong; Gao, Min; Yan, Hongjiang; Li, Baosheng
• Is Part Of: Nuclear medicine communications. Volume 39:Issue 2(2018:Feb.)
• Publication Date: 2018-02
• Subjects: lung cancer -- radiation pneumonitis -- receiver–operator characteristic curves -- single-photon emission computed tomography; Nuclear medicine -- Periodicals; 616.07575
• DOI: 10.1097/MNM.0000000000000774
• Languages: English
• ISSNs: 0143-3636

Abstract

Abstract : Purpose: The aim of this study was to investigate the optimal threshold for the functional lung (FL) definition of single-photon emission computed tomography (SPECT) lung perfusion imaging. Patients and methods: Forty consecutive stage III non-small-cell lung cancer patients underwent SPECT lung perfusion scans and PET/CT scans for treatment planning, and the images were coregistered. Total lung and perfusion lung volumes corresponding to 10, 20, …, 60% of the maximum SPECT count were segmented automatically. The SPECT-weighted mean lung dose (SWMD x % ) and the percentage of FL volume receiving more than 20 Gy ( F x % V 20 ) of different thresholds were investigated using SPECT-weighted dose–volume histograms. Receiver–operator characteristic curves were used to identify SWMD and FV20 of different thresholds in predicting the incidence of radiation pneumonitis (RP). Results: Eleven (27.5%) patients developed RP (grades 1, 2, 3, and 4 were 10.0, 7.5, 7.5, and 2.5%, respectively) after treatment. The largest area under the receiver–operator characteristic curve was 0.881 for the ability of SWMD to predict RP with 20% as the threshold and 0.928 for the ability of FV20 with 20% as the threshold. Conclusion: The SWMD20% and FV20 of FL using 20% of the maximum SPECT count as the threshold may be better predictors for the risk of RP.