Testing Performance of mNGS in Febrile Patients with Hematological Disease and Its Guiding Value for Fever Management in Clinical Treatment

Abstract Objective To explore the testing performance of metagenomic next-generation sequencing (mNGS) in identifying pathogenic microbes in febrile patients with hematological disease and its significant role in guiding clinical treatment. Methods The data of pathogens from the blood cultures of neutropenic patients with hematological disease and/or febrile patients with bloodstream infection (BSI) were summarized and the features of infection were analyzed, through a retrieval of the WoS, PubMed, CNKI, Wanfang, and VIP databases. A retrospective study was conducted on 96 febrile patients with hematological disease (104 specimens) presented to our hospital between May 2022 and May 2024. These patients underwent both routine and mNGS tests for a comparison of the testing performance, and were assigned to mNGS-positive and mNGS-negative groups according to the mNGS results, respectively. Based on an analysis of the data and indexes of the two groups, the contributing factors of mNGS positivity were determined using the univarient and the logistic regression while a prediction model was developed to assess predictive value and summarize the prognosis information using the Receiver Operating Characteristic Curve (ROC). Results According to the included six papers, among totally 3614 isolates from positive blood cultures, Gram-negative bacteria, Gram-positive bacteria, and fungi accounted for 66.66%, 31.02%, and 2.05%, respectively. In this retrospective study, out of 104 peripheral blood tests using mNGS technology, pathogens were detected in 71 tests, with a positive detection rate of 68.27%, which was substantially higher than that of blood culture (7.69%) and routine test (16.35%). Out of 131 isolated pathogenic microbes, viruses held the maximum ratio (60.74%). The identification rate of combined infections by the mNGS test exceeded those by complete blood count (CBC) and routine test. When clinical diagnosis was employed as the gold standard, mNGS test had greater values of sensitivity, positive prediction and negative prediction than those of routine test. The univariant analysis revealed that the mNGS-positive group had higher incidences of pulmonary infection and neutropenia and lower natural killer (NK) cell levels, compared with the mNGS-negative group. The multivariate analysis result of logistic regression model showed that contributing factors of mNGS positivity in febrile patients with hematological disease included pulmonary infection [Odds ratio (OR): 2.389; 95%confidence level (CI): 1.199–4.763)], neutropenia (OR: 4.092; 95% CI: 1.179–14.209), and low NK cell levels (OR: 1.127; 95% CI: 1.117–1.139). According to the analysis result of the ROC curve, the area under curve (AUC) values used for single and combined predictive values of mNGS positivity in febrile patients with hematological disease ranged from 0.623 to 0.849, with the maximum value of 0.849 for combined prediction. The sensitivity and specificity values were 78.79% and 93.33%, respectively. Following relevant guidelines, physicians modified medication for patients with poor disease management based on mNGS results, achieving a survival rate of 75.0%. Conclusion mNGS allows a more comprehensive and accurate identification of pathogenic microbes in patients with hematological disease. However, the positive rate of mNGS test may be affected by pulmonary infection, neutropenia, and low NK cell levels. Therefore, appropriate timing of the mNGS use can provide critical information for the formulation of clinical protocol and enables individualized and accurate treatment.