Clinical evaluation of probe capture based targeted next generation sequencing for pulmonary infection in immunocompromised patients: a cross-sectional diagnostic accuracy study.

Abstract

Background: Timely aetiological diagnosis of pulmonary infection in immunocompromised patients (ICPs) remains challenging because clinical presentations may be atypical and conventional microbiological tests (CMTs) have limited sensitivity. Probe capture based targeted next generation sequencing (ptNGS) has emerged as a potential alternative to metagenomic next generation sequencing (mNGS), but its clinical performance in this population remains incompletely defined.

Methods: In this cross-sectional diagnostic accuracy study, immunocompromised adults undergoing bronchoalveolar lavage for suspected pulmonary infection were enrolled. Bronchoalveolar lavage fluid (BALF) samples were analysed using CMTs, mNGS, and ptNGS. Composite clinical adjudication served as the reference standard. Diagnostic performance was compared at the case level, and pulmonary microbiota characteristics were explored.

Results: Among 78 enrolled patients, 60 were classified as having pulmonary infection. Causative pathogens were identified in 52 cases, and fungal pathogens, particularly Pneumocystis jirovecii, were the most frequently detected. At the case level, ptNGS and mNGS demonstrated higher sensitivity than CMTs (80.0% vs 80.0% vs 26.7%) and showed high concordance in microorganisms identified (91.7%). Specificity was 72.2% for CMTs, compared with 44.4% for mNGS and 38.9% for ptNGS. Positive sequencing results were also observed in patients without pulmonary infection (n = 18), predominantly involving viral or opportunistic microorganisms. Microbiota analysis of 65 samples revealed reduced microbial alpha diversity and altered community composition in patients with pulmonary infection.

Conclusions: In ICPs with suspected pulmonary infection, ptNGS substantially increases pathogen detection compared with CMTs and demonstrates diagnostic performance comparable to mNGS. Sequencing results require careful clinical interpretation, given the difficulty in distinguishing infection from colonisation in respiratory specimens. Exploratory microbiota analyses suggest infection associated alterations in lung microbial ecology that warrant further validation.