Validation of mNGS results using extensive lab and clinical data

Abstract

Purpose: Interpreting the results of metagenomic next-generation sequencing (mNGS) presents a significant challenge in both clinical and laboratory contexts. Methods: A retrospective analysis was conducted to validate mNGS findings, with a particular emphasis on Mycobacterium tuberculosis, Mycoplasma pneumoniae, and Pneumocystis jirovecii as representative pathogens, examined from both clinical and laboratory perspectives. Results: Based on a comprehensive clinical analysis, the mNGS demonstrated detection accuracies for M. tuberculosis, M. pneumoniae, and P. jirovecii of 87.0% (60 out of 69; 95% confidence interval [CI], 77.04%–92.99%), 97.6% (81 out of 83; 95% CI, 91.63%–99.34%), and 78.9% (45 out of 57; 95% CI, 66.72%–87.53%), respectively. Conversely, when incorporating laboratory confirmation from a variety of detection methodologies, the accuracy rates for mNGS in identifying M. tuberculosis, M. pneumoniae, and P. jirovecii were 92.7% (51 out of 55; 95% CI, 82.74%–97.14%), 82.3% (51 out of 62; 95% CI, 70.96%–89.80%), and 83.9% (26 out of 31; 95% CI, 67.36%–92.91%), respectively. Additionally, our analysis revealed no statistically significant difference in read counts and relative abundances between mNGS results deemed clinically as false positives and those considered true positives (P < 0.05). Conclusion: In contemporary clinical practice, the detection of positive results from mNGS is notably high from both laboratory and clinical standpoints. Nonetheless, the interpretation of results with low read counts presents significant challenges for both clinical and laboratory environments under current conditions.