关键词: 体层摄影术,螺旋计算机, 结核, 气管疾病, 支气管疾病, 图像处理,计算机辅助, 诊断

Abstract:

Objective To explore the value of bronchial image post-processing technique in the diagnosis of tracheobronchial tuberculosis by multi-slice CT (MSCT) scanning. Methods A total of 117 patients with high suspicion of TBTB due to clinical symptoms or MSCT axial images who underwent MSCT scans and bronchoscopy from June to December 2018 in the Tuberculosis Hospital of Shannxi Province were retrospectively collected. The axial plane images of all patients were post-processed with MSCT bronchial images. After bronchoscopy, tissue samples were obtained under the microscope and sent to the pathology department for pathological biopsy. Lastly, 69 patients with TBTB and 48 patients with non-TBTB were diagnosed. The diagnostic results of bronchoscopy and biopsy pathology were used as the reference standard to evaluate the sensitivity, specificity and accuracy of MSCT bronchial image post-processing technology for the diagnosis of TBTB, and a consistency test was used to evaluate the consistency of the diagnosis by two imaging physicians. Results Based on the results of bronchoscopy and biopsy pathology, the sensitivity, specificity, and accuracy of MSCT bronchial image post-processing technology for the diagnosis of TBTB were 97.10% (67/69), 43.75% (21/48), and 75.21% (88/117), respectively, and the consistency analysis result showed that the Kappa value was 0.44 (95%CI: 0.29-0.60). In addition, two imaging physicians had better consistency in the diagnosis of the disease through this method (Kappa=0.81, 95%CI: 0.68-0.93). Conclusion MSCT bronchial image post-processing technology has better diagnostic value for TBTB.

Key words: Tomography,spiral computed, Tuberculosis, Tracheal diseases, Bronchial diseases, Image processing,computer-assisted, Diagnosis