肉芽肿
活检
分枝杆菌
外科病理学
病理
病因学
医学
生物
肺结核
作者
Wenwen Sun,Zheng-wei Dong,Yiming Zhou,Feng Jin,Hongcheng Liu,Lin Fan
标识
DOI:10.1016/j.micinf.2023.105185
摘要
Metagenomic Next-Generation Sequencing (mNGS) has been validated to have an important role in the diagnosis of mycobacterium infection. The study aimed to further explore the mycobacteria identification ability of mNGS on formalin-fixed paraffin-embedded(FFPE)tissues from postoperative specimens. Patients who underwent surgical biopsy or resection for clarifying the diagnosis and whose initial postoperative pathology indicated granulomatous lesions were included. Fresh tissues were sent for mycobacterium culture and Xpert MTB/RIF (Xpert) to establish the diagnosis. FFPE specimens were sent for mNGS and molecular pathology,the diagnostic values were compared between the two methods. A total of 65 cases with definite diagnoses were finally included in the study. 31 cases were confirmed as mycobacterium granuloma using the fresh specimen etiology as diagnostic criteria. The overall sensitivity and specificity of mNGS on FFPE specimens in the diagnosis of mycobacterium granuloma were 100% and 88.24%, respectively. In 19 cases diagnosed as tuberculous granulomas, the sensitivity (100% vs47.37%) and negative predictive value (NPV, 100%vs 82.14%) of mNGS were both significantly higher than that of molecular pathology on the FFPE section(both p 0.00)while the positive predictive value (PPV) and specificity were not significantly different. In 12 cases diagnosed as Non-tuberculous mycobacterium (NTM)granuloma, the sensitivity of mNGS was also significantly higher than that of molecular pathology on FFPE section (100% vs 66.67%, p 0.00) while the specificity, PPV and NPV were all not significantly different. The mNGS could be used for one-time detection of pathogens on FFPE sections with high sensitivity. It could be recommended as a supplementary method for the identification of pathogenic bacteria in the diagnosis of postoperative granuloma lesions.
科研通智能强力驱动
Strongly Powered by AbleSci AI