Differentiating malignant and benign necrotic lung lesions using kVp-switching dual-energy spectral computed tomography

BMC Medical Imaging

Xiao Fan¹  Fa-Jin Lv²  Tian-You Luo²  Qi Li²  Xing-Tao Huang³ 

[1] Department of Radiology, Children’s Hospital of Chongqing Medical University, No. 136 Zhongshan Road Two, Yuzhong District, Chongqing, China;Department of Radiology, The First Affiliated Hospital of Chongqing Medical University, No. 1 Youyi Road, Yuzhong District, Chongqing, China;Department of Radiology, University of Chinese Academy of Sciences Chongqing Renji Hospital (Fifth People’s Hospital of Chongqing), No. 24 Renji Road, Nan’an District, Chongqing, China;

关键词: Lung neoplasms;    Inflammation;    Necrosis;    Dual-energy computed tomography;    Quantitative imaging;   

DOI  :  10.1186/s12880-021-00611-6

来源: Springer

【 摘 要 】

BackgroundNecrotic pulmonary lesions manifest as relatively low-density internally on contrast-enhanced computed tomography (CT). However, using CT to differentiate malignant and benign necrotic pulmonary lesions is challenging, as these lesions have similar peripheral enhancement. With the introduction of dual-energy spectral CT (DESCT), more quantitative parameters can be obtained and the ability to differentiate material compositions has been highly promoted. This study investigated the use of kVp-switching DESCT in differentiating malignant from benign necrotic lung lesions.MethodsFrom October 2016 to February 2019, 40 patients with necrotic lung cancer (NLC) and 31 with necrotic pulmonary mass-like inflammatory lesion (NPMIL) were enrolled and underwent DESCT. The clinical characteristics of patients, CT morphological features, and DESCT quantitative parameters of lesions were compared between the two groups. Binary logistic regression analysis was performed to identify the independent prognostic factors differentiating NPMIL from NLC. Receiver operating characteristic (ROC) curves were used to assess the diagnostic performance of single-parameter and multiparametric analyses.ResultsSignificant differences in age, C-reactive protein concentration, the slope of the spectral curve from 40 to 65 keV (K40–65 keV) of necrosis in non-contrast-enhanced scanning (NCS), arterial phase (AP) and venous phase (VP), effective atomic number of necrosis in NCS, and iodine concentration (IC) of the solid component in VP were observed between groups (all p < 0.05). The aforementioned parameters had area under the ROC curve (AUC) values of 0.747, 0.691, 0.841, 0.641, 0.660, 0.828, and 0.754, respectively, for distinguishing between NLC and NPMIL. Multiparametric analysis showed that age, K40–65 keV of necrosis in NCS, and IC of the solid component in VP were the most effective factors for differentiating NLC from NPMIL, with an AUC of 0.966 and percentage of correct class of 88.7%.ConclusionsDESCT can differentiate malignant from benign necrotic lung lesions with a relatively high accuracy.

【 授权许可】

CC BY