二代测序技术应用于脑脊液检测在结核性脑膜炎中的早期诊断价值

Evaluation of next-generation sequencing for early diagnosis of tuberculous meningitis

目的评价二代测序技术应用于脑脊液检测在结核性脑膜炎(TBM)患者中的早期诊断价值。方法前瞻性纳入2018年2月2日至2018年8月2日于山东省胸科医院就诊的临床怀疑TBM的患者共50例,并跟踪随访其诊疗结局。送检脑脊液标本均进行二代测序,测序所得原始序列与病原微生物数据库进行对比得到最终结果。二代测序结果以检测到结核分枝杆菌复合群唯一比对序列为阳性,未检测到唯一比对序列为阴性。以符合脑脊液结核分枝杆菌培养阳性、涂片阳性、Xpert MTB/RIF检测阳性及结核分枝杆菌核酸检测阳性等4项中至少1项即为确诊TBM患者;临床可疑TBM且抗结核治疗有效为临床诊断患者;有其他病原学依据或临床排除TBM者为非TBM患者。分析二代测序在TBM早期诊断中的敏感性和特异度。结果确诊为TBM患者22例中Xpert MTB/RIF检测阳性13例,培养阳性6例,结核分枝杆菌核酸PCR检测阳性5例,临床诊断为TBM患者12例,非TBM患者16例。在确诊及临床诊断患者中,二代测序技术检测到结核分枝杆菌复合群系列20例,敏感性为58.8%(20/34),特异度为100%(16/16)。在确诊患者中,二代测序的敏感性为63.6%(14/22);在同步进行结核分枝杆菌培养、Xpert MTB/RIF检测与二代测序的50例标本中,以临床诊断为标准,3种方法的特异度均为100%(16/16);传统方法、Xpert MTB/RIF检测及二代测序的敏感性分别为29.4%(10/34)、38.2(13/24)和58.8(20/34),前两种检测方法与二代测序敏感性差异均有统计学意义(McNemar检验:χ^2=8.333、P=0.013,χ^2=8.333、P=0.065)。传统方法与二代测序联合检测的敏感性高达82.4%(28/34)。结论二代测序技术能够较快速地检测脑脊液中的结核分枝杆菌复合群,且其敏感性和特异度均较高,可作为TBM的早期诊断指标。二代测序联合传统检测方法可提高检出率。

Objective To evaluate the early diagnostic value of next-generation sequencing technology applied to cerebrospinal fluid detection in patients with tuberculous meningitis.Methods A total of 50 patients with clinically suspected tuberculous meningitis who were treated in Shandong Provincial Chest Hospital from February 2nd,2018 to August 2nd,2018 were collected prospectively,and the diagnosis and treatment outcomes of those patients were followed up.The submitted cerebrospinal fluid specimens were all subjected to next-generation sequencing,and the obtained original sequence was compared with the pathogenic microorganism database to get the final results.The next-generation sequencing results showed positive when detecting the unique alignment sequence of the Mycobacterium tuberculosis complex,and negative when no unique alignment sequence was detected.Tuberculous meningitis were confirmed with at least one of the four items:cerebrospinal fluid Mycobacterium tuberculosis culture positive,smear positive,Xpert MTB/RIF test positive and Mycobacterium tuberculosis nucleic acid test positive;clinically diagnosed tuberculous meningitis patients were with clinically suspected tuberculous meningitis and effective antituberculosis treatment;non-tuberculous meningitis patients were with other etiological evidence or clinical exclusion of tuberculous meningitis.The sensitivity and specificity of next-generation sequencing in early diagnosis of tuberculous meningitis were analyzed,respectively.Results Among the 22 patients with confirmed tuberculous meningitis,13 cases were positive for Xpert MTB/RIF test,6 cases were positive for culture,and 5 cases were positive for Mycobacterium tuberculosis nucleic acid by PCR.There were 12 cases clinically diagnosed as tuberculous meningitis and 16 cases as non-tuberculous meningitis patients.Among the confirmed and clinically diagnosed patients,20 cases of Mycobacterium tuberculosis complex series were detected by next-generation sequencing technology,with a sensitivity of 58.8%(20/34)and a specificity of 100%(16/16).Among the confirmed patients,the sensitivity of next-generation sequencing was 63.6%(14/22).Among the 50 specimens that were simultaneously submitted for Mycobacterium tuberculosis culture,Xpert MTB/RIF test and next-generation sequencing,the specificity of the three methods was 100%(16/16)with clinical diagnosis as the standard.The sensitivity of traditional method,Xpert MTB/RIF test and next-generation sequencing were 29.4%(10/34),38.2(13/24)and 58.8(20/34),respectively.The sensitivity differences between the first two detection methods and next-generation sequencing were significantly different(McNemar test:χ^2=8.333,P=0.013;χ^2=8.333,P=0.065).The sensitivity of the combined detection of traditional method and next-generation sequencing was as high as 82.4%(28/34).Conclusions Next-generation sequencing technology could quickly detect the Mycobacterium tuberculosis complex in the cerebrospinal fluid,with significant sensitivity and specificity,and could be used as an early diagnosis index for tuberculous meningitis.Next-generation sequencing combined with traditional detection method could increase the detection rate.