3T magnetic resonance for evaluation of adult pulmonary tuberculosis. (April 2020)

• Record Type: Journal Article
• Authors: Yan, Qinqin; Yang, Shuyi; Shen, Jie; Lu, Shuihua; Shan, Fei; Shi, Yuxin
• Is Part Of: International journal of infectious diseases. Volume 93(2020)
• Page Start: 287
• Page End: 294
• Publication Date: 2020-04
• Subjects: Pulmonary tuberculosis -- 3T MRI -- Image quality -- Detection; Communicable diseases -- Periodicals; Communicable Diseases -- Periodicals; Communicable diseases; Periodicals; Electronic journals; 616.9
• DOI: 10.1016/j.ijid.2020.02.006
• Languages: English
• ISSNs: 1201-9712

Abstract

Highlights: MRI is comparable to CT in detection of tree-in-bud signs, nodules (≥5 mm) and consolidations. MRI has more advantages in presenting caseous necrosis, liquefaction, active cavities and abnormalities of lymph nodes and pleura. Compared to CT, MRI has lower sensitivity for detecting calcified nodules and non-calcified nodules (＜5 mm). Free-breathing T1-weighted Star-VIBE and T2-weighted 2D-fBLADE TSE are suitable for chest imaging, especially for patients with poor breath-holding. MRI with non-radiation is an alternative tool for follow-up examinations, especially for children, young women and pregnant women. Abstract: Objectives: To evaluate image quality and detection rate of four 3T magnetic resonance imaging (MRI) sequences and MRI performances in pulmonary tuberculosis (TB) when compared to computed tomography (CT). Methods: Forty patients with pulmonary tuberculosis separately underwent CT and 3T-MRI with T1-weighted free-breathing star-volumetric interpolated breath-hold examination (Star-VIBE) and standard VIBE, T2-weighted two-dimensional fast BLADE turbo spin-echo (2D-fBLADE TSE) and three-dimensional isotropic turbo spin-echo (3D-SPACE). Four MRI sequences were compared in terms of detection rate and image quality, which consisted of signal to noise ratio (SNR), contrast to noise ratio (CNR) and 5-point scoring scale. The total sensitivity was also compared between CT and MRI. Inter-observer agreement on 5-point scoring scale was calculated by Cohen's kappa (k). SNR, CNR and 5-point scoring scale were compared using two-tailed pared t-test. Using CT as a reference, the MRI detection rate of pulmonary abnormality was evaluated by Pearson's Chi-square test. Furthermore, the sizes of the nodules (≥5 mm) were compared using intraclass correlation coefficient. Results: In this study, Free-breathing Star-VIBE had significantly better SNR and identical CNR compared with standard VIBE. 2D-fBLADE TSE had statistically higher SNR but uniform or inferior CNR compared with 3D-SPACE. Inter-observers showed excellent agreement on 5-point scoring scale. The average score of Star-VIBE and VIBE had no difference. The average score of 2D-fBLADE TSE was higher than 3D-SPACE. There were no statistical differences in the detection rates of non-calcified parenchymal lesions between Star-VIBE and standard VIBE, 2D-fBALDE TSE and 3D-SPACE. MRI is comparable to CT in detecting consolidation, cavity, non-calcified nodules of ≥5 mm and tree-in-bud signs compared to CT. MRI detected non-calcified nodules of ＜5 mm, 5–10 mm, ≥10 mm and calcified nodules with sensitivity of 69.6%, 90.6%, 100% and 89.5% respectively. In addition, the sizes of the nodules (≥5 mm) had statistical consistency. MRI is more sensitive in detecting caseous necrosis, liquefaction, active cavity, abnormalities of lymph nodes and pleura. Conclusions: T1-weighted free-breathing Star-VIBE and T2-weighted 2D-fBLADE TSE, both with satisfactory image quality, are suitable for patients with pulmonary TB who need long-term follow-ups in clinical routine, especially for children, young women and pregnant women.