Purpose. Metagenomic next-generation sequencing (mNGS) has been widely used in the diagnosis of infectious diseases, while its performance in diagnosis of tuberculous meningitis (TBM) is incompletely characterized. The aim of this study was to assess the performance of mNGS in the diagnosis of TBM, and illustrate the sensitivity and specificity of different methods. Methods. We retrospectively recruited TBM patients between January 2021 and March 2023 to evaluate the performance of mNGS on cerebrospinal fluid (CSF) samples, in comparison with conventional microbiological testing, including culturing of Mycobacterium tuberculosis (MTB), acid-fast bacillus (AFB) stain, reverse transcription PCR and Xpert MTB/RIF. Results. Of the 40 enrolled, 34 participants were diagnosed with TBM, including 15(44.12 %) definite and 19(55.88 %) clinical diagnosis based upon clinical manifestations, CSF parameters, brain imaging, pathogen evidence and treatment response. The mNGS method identified sequences of Mycobacterium tuberculosis complex (MTBC) in 11 CSF samples. In patients with definite TBM, the sensitivity, specificity, positive predictive value, negative predictive value and accuracy of mNGS were 78.57, 100, 100, 66.67 and 85 %, respectively. Compared to conventional diagnostic methods, the sensitivity of mNGS (78.57 %) was higher than AFB (0 %), culturing (0 %), RT-PCR (60 %) and Xpert MTB/RIF (14.29 %). Conclusions. Our study indicates that mNGS of CSF exhibited an overall improved sensitivity over conventional diagnostic methods for TBM and can be considered a front-line CSF test.

中文翻译：

---

脑脊液宏基因组新一代测序诊断结核性脑膜炎的性能

目的。宏基因组下一代测序（mNGS）已广泛应用于传染病的诊断，但其在结核性脑膜炎（TBM）诊断中的表现尚不完全表征。本研究的目的是评估 mNGS 在 TBM 诊断中的性能，并说明不同方法的敏感性和特异性。 方法。我们回顾性招募了 2021 年 1 月至 2023 年 3 月期间的 TBM 患者，以评估 mNGS 对脑脊液 (CSF) 样本的性能，并与传统微生物检测（包括结核分枝杆菌(MTB) 培养、抗酸杆菌 (AFB) 染色、逆转录 PCR 和 Xpert MTB/RIF。 结果。在 40 名入组者中，34 名参与者被诊断为 TBM，其中 15 名（44.12％）确诊为 TBM，19 名（55.88％）根据临床表现、脑脊液参数、脑成像、病原体证据和治疗反应进行临床诊断。 mNGS 方法鉴定了 11 个脑脊液样本中的结核分枝杆菌复合体 (MTBC) 序列。在明确TBM的患者中，mNGS的敏感性、特异性、阳性预测值、阴性预测值和准确性分别为78.57%、100%、100%、66.67%和85%。与传统诊断方法相比，mNGS（78.57%）的敏感性高于AFB（0%）、培养（0%）、RT-PCR（60%）和Xpert MTB/RIF（14.29%）。 结论。我们的研究表明，与传统的 TBM 诊断方法相比，CSF mNGS 的敏感性总体提高，可以被视为一线 CSF 测试。