Applicability of Bronchoalveolar Lavage Fluid and Plasma Metagenomic Next-Generation Sequencing Assays in the Diagnosis of Pneumonia

Abstract

Background Metagenomic next-generation sequencing (mNGS) provides innovative solutions for predicting complex infections. A comprehensive understanding of its strengths and limitations in real-world clinical settings is necessary to ensure that it is not overused or misinterpretation. Methods 209 cases with suspected pneumonia were recruited to compare the capabilities of two available mNGS assays (i.e., BALF mNGS and plasma mNGS) to identify pneumonia-associated DNA/RNA pathogens and predict antibiotic resistance. Results Compared to clinical diagnosis, BALF mNGS demonstrated a high positive percent agreement (PPA) (95.3%) but a low negative percent agreement (NPA) (63.1%). Plasma mNGS revealed a low proportion of true negatives (TN) (30%) in predicting pulmonary infection. BALF mNGS independently diagnosed 65.6% (61/93) of co-infections and had a remarkable advantage in detecting caustic, rare or atypical pathogens. Pathogens susceptible to invasive infection or bloodstream transmission, such as Aspergillus spp, Rhizopus spp. Chlamydia psittaci, and human herpesviruses are prone to be detected by plasma mNGS. BALF mNGS tests provided a positive impact on the diagnosis and treatment of 128 (61.2%) patients. Plasma mNGS, on the other hand, turned out to be more suitable for diagnosing patients who received mechanical ventilation, developed severe pneumonia or sepsis (all p<0.01). BALF mNGS was able to identify resistance genes that matched the phenotypic resistance of 69.4% (25/36) multidrug-resistant (MDR) pathogens. Conclusions Our data reveal new insights into the (dis)advantage of two different sequencing modalities in pathogen identification and antibiotic resistance prediction for patients with suspected pneumonia.