Abstract
Metagenomic next-generation sequencing (mNGS) is increasingly employed for the diagnosis of lower respiratory tract infections (LRTIs). However, the relative diagnostic performance of long-read versus short-read sequencing platforms remains incompletely defined. For this systematic review, a search was conducted in PubMed, Embase, Scopus, Web of Science, and Google Scholar to identify studies directly comparing long-read (e.g., Oxford Nanopore, PacBio) and short-read (e.g., Illumina, Ion Torrent, BGISEQ) metagenomic sequencing for the diagnosis of LRTI. Eligible studies reported diagnostic accuracy or comparative performance between platforms. Risk of bias was evaluated using the QUADAS-2 tool. Thirteen studies met inclusion criteria. Reported platforms included Illumina, Oxford Nanopore, PacBio, Ion Torrent, and BGISEQ-500. A total of 13 studies met inclusion criteria. Across studies reporting sensitivity, average sensitivity was similar for Illumina (71.8%) and Nanopore (71.9%). Specificity varied substantially, ranging from 42.9 to 95% for Illumina and 28.6 to 100% for Nanopore. Concordance between platforms ranged from 56 to 100%. Illumina consistently produced superior genome coverage (approaching 100% in most reports) and higher per-base accuracy, whereas Nanopore demonstrated faster turnaround times (<24 h), greater flexibility in pathogen detection, and superior sensitivity for Mycobacterium species. Risk of bias was frequently high or unclear, particularly in patient selection (6 studies), index test interpretation (5), and flow and timing (4), limiting the robustness of pooled estimates. Long-read and short-read mNGS platforms exhibit comparable strengths in the diagnosis of LRTIs. Illumina remains optimal for applications requiring maximal accuracy and genome coverage, whereas Nanopore offers rapid, versatile pathogen detection, particularly for difficult-to-detect organisms such as Mycobacterium. However, there are certain limitations of the review, including a lack of comparable outcomes reported in all studies; therefore, further research is warranted to address this.