Introduction
Metagenomic next-generation sequencing (mNGS) has the potential to identify a variety of pathogens from a single blood sample, to aid in the diagnosis of deep-seated and bloodstream infections. One commercially available assay, the Karius test, reportedly detects 1,250 species of pathogens. Reference Blauwkamp, Thair and Rosen1 Sensitivity (51–90%) and specificity (86%) of mNGS have been reported to outperform culture in prior studies. Reference Miao, Ma and Wang2,Reference Rossoff, Chaudhury and Soneji3 However, there are known limitations to mNGS, particularly identification of genetic material from transiently translocated organisms, the current high cost of testing, and lack of antimicrobial susceptibility results. Furthermore, only a few studies have evaluated the impact of mNGS on clinical decision-making when compared to more conventional diagnostic methods. Reference Hogan, Yang and Garner4,Reference Shishido, Noe, Saharia and Luethy5 In this study we assess the clinical impact of the Karius test on patient care at an Iowa Academic Medical Center.
Methods
We conducted a retrospective cohort study of patients at a single center, the University of Iowa Hospitals & Clinics, spanning from January 2020 to June 2022. All patients who underwent mNGS testing with the Karius assay were included in this review. Any provider has the option to request Karius testing, but approval from a microbiology director is necessary, and the testing is authorized only when conventional tests, like blood culture and biopsies, yield negative results or are infeasible. A team of infectious disease physician reviewers (MO, TK, DI, KP, LN) conducted a comprehensive chart review focusing on patient characteristics, clinical progression, and outcomes. The results of mNGS and conventional tests were compared for concordance. Conventional tests encompassed cultures, serology, antigen testing, and targeted polymerase chain reaction (PCR). We assessed the clinical impact of mNGS testing using criteria established in prior studies. Reference Hogan, Yang and Garner4 These were classified into positive impact, negative impact, no impact, or indeterminate (Supplementary Table 1 (online)). This study was approved by the institutional review board.
Results
A total of 37 patients had the Karius test performed during our study window. Fever was the most commonly observed symptom in 65% of patients, followed by respiratory symptoms at 57%. Among these 37 patients, 29 (78%) had an immunocompromising condition, with the most common being hematologic malignancy (21 patients, 58%, as shown in Table 1). Additionally, 34 patients (92%) had received a formal infectious disease (ID) consultation prior to the performance of the Karius test. The most common finding before testing was the presence of pulmonary infiltrates on imaging (30 patients, 81%). Thirteen patients had positive results for both mNGS and conventional testing, of which 12 had completely or partially identical organisms. The Karius test results were determined to have a positive clinical impact in six cases (16%, Table 2, Supplemental Table 1 (online)), an indeterminate impact in one case (3%), and no impact in 30 cases (81%).
mNGS, metagenomic next-generation sequencing.
AFB, acid-fast bacilli, mNGS: metagenomic next-generation sequencing, TB, tuberculosis, HSV, herpes simplex virus, HHV, human herpesvirus, EBV, Epstein-Barr virus, LFT, liver function test, ID, infectious diseases, JC virus, JC polyomavirus virus, HIV, human immunodeficiency virus, HPV, human papillomavirus, CMV: cytomegalovirus.
In two of the six patients for whom the Karius test had a positive clinical impact, mycobacterial pathogens were identified, which led to the initiation of empiric therapy faster than would have been possible with AFB cultures (Table 2). In one case, the test result prompted the initiation of antifungal therapy for probable pulmonary aspergillosis. In another patient with a diffuse rash and pulmonary infiltrates, the mNGS results were positive for JC virus, Epstein-Barr virus, and human papillomavirus, supporting the reduction of immunosuppression. Antifungal therapy was de-escalated based on a negative result in another patient. In one case where a valvular echodensity was observed on echocardiography, with no risk factors or clinical stigmata of infective endocarditis, a negative result provided reassurance to move forward with lung transplantation. It is noteworthy that one of the 30 patients categorized as having ‘no impact’ based on the criteria used was clinically considered inaccurate. This patient had been experiencing renal and pulmonary lesions for months, along with a cardiac mass and a positive beta-D-glucan test result. This patient had been on antifungal therapy with mold-active agents for an extended period. Despite the Karius test returning a negative result, tissue cultures and blood cultures eventually grew Fusarium sp. These results, taken together with the identification of fungal hyphae in a biopsy of the cardiac mass, prove the Karius result was a false negative.
Discussion
We conducted a manual chart review of 37 patients with Karius testing to assess the impact of metagenomic cell-free DNA testing on patient care using previously established criteria from the literature. We considered 16% to have a positive clinical impact, while 81% had no impact, and 3% had an indeterminate impact.
The positive impact rate of 16% in our patients is similar to previous reviews which demonstrated an impact of 7.3%, Reference Hogan, Yang and Garner4 however is lower than others which demonstrated a positive impact of >40%. Reference Shishido, Noe, Saharia and Luethy5,Reference Han, Yu and Zhang6 The variability in reported clinical impact across the literature may be attributed to the use of nonstandard criteria for classifying clinical impact. Notably, some studies classify a test that confirms a previously known diagnosis as having a positive clinical impact while the criteria we used did not. In our review, most cases in which the diagnosis was confirmed were determined to have no impact on patient care. Furthermore, variations in practices among different institutions, including differences in the timing of testing (early vs. late) and selection criteria for employing mNGS, could also contribute to differences in clinical impact.
Interestingly, in two out of the six patients with a positive clinical impact, mNGS identified a mycobacterial pathogen more rapidly than conventional testing. The test not only provided faster results than conventional cultures but also yielded a positive result for disseminated Mycobacterium avium complex in myelofibrosis patient and M. tuberculosis in one other case, despite negative sputum smears and GeneXpert MTB/RIF PCR results. This suggests that patients suspected of having slow-growing pathogens, such as Mycobacteria species, may be a subgroup in which mNGS is more likely to have a positive impact. Further studies with a larger sample size are necessary to confirm this observation.
Additionally, in another two of the six patients with a positive clinical impact, a clinical decision was made based on the negative results of mNGS. However, extreme caution should be exercised when utilizing a negative test result, especially in patients with a high pretest probability of infection. There is a wide range of reported sensitivity for mNGS testing in the literature (51–90%). Reference Blauwkamp, Thair and Rosen1,Reference Miao, Ma and Wang2 We also identified one case in the ‘no impact’ group where a patient with a high burden of disease, specifically Fusarium sp. growing directly within the bloodstream, returned a negative mNGS test result.
The outcomes of mNGS require meticulous interpretation, irrespective of a positive or negative outcome. A positive outcome may indicate colonization or contamination, while a negative test cannot guarantee the absence of infection in a patient. It might be prudent for hospitals to limit the authorization of test orders to ID physicians or microbiologists exclusively. In this present study, the majority of patients underwent ID consultations prior to mNGS submission. This underscores the crucial contribution of infectious disease expertise in ensuring responsible test ordering, nuanced interpretation of mNGS results, and providing direction for subsequent management decisions.
In conclusion, our retrospective study found that mNGS had a positive impact on 16% of the patients for whom it was performed. While mNGS results may provide valuable supplementary information, it is essential to interpret them within the broader clinical context and evaluation. mNGS results could be beneficial in selected cases, particularly when conventional testing, such as for Mycobacterium infection, has a prolonged turnaround time. Prospective studies are needed to better determine the clinical scenarios in which mNGS has the greatest potential to improve patient outcomes.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/ash.2024.31.
Acknowledgments
None.
Financial support
None.
Competing interests
None.