Introduction
Mycobacterial diseases remain a substantial cause of morbidity and mortality in the United States, where the prevalence of tuberculosis (TB) is 4.4–4.8%. Reference Mancuso, Diffenderfer, Ghassemieh, Horne and Kao1 The prevalence of non-tuberculous mycobacteria remains more infrequent but persists at 11.70 per 100,000. Reference Winthrop, Marras, Adjemian, Zhang, Wang and Zhang2 Establishing the diagnosis in these cases promptly is challenging and often requires the evaluation of multiple, time-consuming, quality-dependent tests that may not definitively rule out the infection. Reference Datta, Shah, Gilman and Evans3
Next-generation sequencing (NGS) of microbial cell-free DNA (mcfDNA) has become increasingly common as an open-ended method for testing against a broad range of pathogens through non-invasive plasma sampling. Reference Blauwkamp, Thair and Rosen4 Previous studies of mcfDNA assays have characterized the sensitivity and specificity of the test in TB cases. Reference Yu, Shen, Ye and Shi5 Less clear is the role mcfDNA NGS plays compared to other diagnostic modalities, its performance, and its impact on management. In this retrospective case series, we characterize the mcfDNA role in mycobacterial diagnosis and management in one quaternary medical center.
Methods
The performance of mcfDNA NGS was compared to mycobacterial cultures among patients who had both tests performed between 2017 and 2024. The electronic medical records for patients with positive mcfDNA NGS for mycobacteria were examined for clinical symptoms, radiology, diagnostic tests, and clinical management.
mcfDNA NGS tests were sent to Karius, Inc. (Redwood City, CA), the sputum acid-fast bacilli (AFB) PCR was sent to Quest Diagnostics ® (Secaucus, NJ), while the mycobacterial cultures were performed in-house.
The primary outcome was the sensitivity of the mcfDNA NGS tests compared with mycobacterial cultures. The secondary outcomes included the turnaround time (TAT) and whether mcfDNA NGS led to a change in antimicrobial management. The clinical TAT was defined as the time from physician order and collection to reporting results (days). In contrast, the laboratory TAT was defined as the time from specimen receipt by the laboratory to reporting results. This study was approved by the Baylor College of Medicine Institutional Review Board.
Results
149 patients had both mcfDNA NGS and mycobacterial cultures performed, of which 135 were positive for neither, and 15 were positive for at least one. Five patients with Mycobacterium tuberculosis (MTB) on mcfDNA NGS had confirmatory cultures. Four had non-tuberculous mycobacteria (NTM) on both methods, and three had positive cultures but negative NGS. Three others had positive NGS for NTM but negative AFB cultures, with two deemed false positives and one with a prior history of Mycobacterium avium complex (MAC). Sensitivity and specificity for mcfDNA NGS were 75.0% and 97.8%; for MTB, both were 100%, while NTM sensitivity was 57.1% and specificity 97.8%.
Nine patients with positive mcfDNA NGS and microbiologic data were reviewed: five for MTB and four for NTM. The median age was 69 years, with six males and three females. Seven had risk factors for mycobacterial infection, including two from endemic regions, one on adalimumab, a lung transplant recipient, and three with AIDS (CD4 < 50). Chronic fever and weight loss were common, with lymphadenopathy seen in three MTB and two MAC cases.
On average, each patient received 27 diagnostic studies for infectious diseases (range 14–78) before diagnosis (Table 2). All patients except one had pending tests for mycobacterial disease at the time mcfDNA NGS was sent, one already had a positive culture for Mycobacterium abscessus, and one had a previously diagnosed MAC infection.
Table 2. Imaging and laboratory studies for study patients
Abbreviations: CSF, Cerebrospinal fluid; AFB, Acid-fast bacilli; MTB, Mycobacterium tuberculosis; MAC, Mycobacterium avium complex; PPD, Purified Protein Derivative; IGRA, Interferon Gamma Release Assay; NGS, Next-generation sequencing; MPM, Molecules per microliter; RR, Reference range; EBV, Epstein-Barr virus; CMV, Cytomegalovirus.
a Dates are reported from day of hospital admission. Dates in which tests were received are reported for send-out studies.
b First positive mycobacterial diagnostic result received for a patient.
c Patients received multiple cultures and repeat PCR tests after the ones included in the table, as reflected in the counts of laboratory studies. For conciseness, only the first set are shown in this table.
d In a 684 healthy subject cohort used in the Karius test validation process, the 97th percentile of both Mycobacterium tuberculosis and Mycobacterium avium complex mcfDNA concentration was 10 MPM (Blauwkamp et al., 2019)
e Some patients seen in 2017 received an earlier version of mcfDNA NGS studies that did not quantify the MPM of mcfDNA. Instead, the samples were tested with a negative buffer control.
f Concentration above quantifiable range for the assay.
The median clinical TAT for mcfDNA NGS was 3.3 days (range 2–6), while the median laboratory TAT was one day (range 1–2). Mycobacterial cultures had a median clinical and laboratory TAT of 27 days (range 9–46). Three acid-fast stains were positive: a lymph node biopsy and two sputum specimens with a median clinical and laboratory TAT of one day. Two patients had a positive sputum MTB PCR with a clinical TAT of nine days for both and a median laboratory TAT of 4.5 days (range 2–7).
The median time from admission to diagnosis was 11 days (range 4–15). In four patients, NGS results were the first to be positive. In three, alternative testing resulted in positive for MAC and TB before NGS results were received. The diagnosis was already established in two when mcfDNA NGS testing was sent.
Anti-mycobacterial therapy was started empirically in three patients: two with tuberculosis and one with NTM. One was already receiving therapy for a prior MAC diagnosis. In two, anti-mycobacterial therapy was started once positive NGS had resulted. In two, treatment was started after positive mycobacterial cultures. One was discharged and lost to follow-up before the start of treatment. Additionally, five patients discontinued empiric antibiotics (vancomycin, meropenem, amphotericin B, micafungin, and azithromycin) once positive NGS results were received. Four of nine patients required intensive care unit admissions and intubation. One patient expired, one was discharged to hospice care, and another to a skilled nursing facility, while the rest were discharged home.
Discussion
Our results showed that when clinical suspicion is high, mcfDNA NGS testing can be an additional and valuable tool to achieve an accurate and timely diagnosis. Despite being a send-out test collected after other mycobacterial studies, mcfDNA NGS testing established the diagnosis before other laboratory studies in four out of nine patients. It also led to changing clinical management in patients due to earlier initiation of antimycobacterial treatment and discontinuation of other antimicrobials.
There were significant issues encountered by clinicians when using traditional tests, including the time required for laboratory diagnosis using conventional mycobacterial culture and send-out TB PCRs and the limited sensitivity of acid-fast smears. NGS of mcfDNA allowed clinicians to discontinue unnecessary antibiotics in some patients and start anti-mycobacterial therapy in others. The observed performance of the test was greater for MTB than seen in a prior study, Reference Yu, Shen, Ye and Shi5 but it was markedly less for NTM than MTB. Notably, the delay in ordering the mcfDNA NGS compared to traditional tests could have negatively impacted its performance.
The test results must be carefully correlated with clinical presentation, as it can detect low concentrations of organisms like MAC that may colonize airways without causing disease. Additionally, mcfDNA NGS remains costly at $5,494 compared to $231 for MTB PCR and $4,325 for daily ICU charges. Until costs decrease, this test should be reserved for cases where diagnostic delays may lead to clinical deterioration and increased spending.
The greatest benefit of mcfDNA NGS was observed when traditional tests were done alongside NGS to maximize microbial identification and expedite results or when the disease presentation was nonspecific and broad antimicrobial therapies were used. Patients where mcfDNA NGS did not change outcomes were often immunocompromised, with the disseminated disease diagnosed by sampling sputum, blood, or stool. A prior study showed no clinical benefit in over nine of ten tests. Clinicians should carefully assess available diagnostics and their potential impact on outcomes. Reference Hogan, Yang and Garner7
While this series does not definitively answer how diagnostic approaches for such complex presentations should be adjusted, it does illustrate the necessity for further investigation to best utilize this open-ended diagnostic tool in such challenging encounters. Our study is limited by the small number of patients included. Also, the TAT of different laboratory tests varies by the institution and microbiology laboratory capacity, affecting the generalizability of our findings.
Table 1. Clinical characteristics of study patients
Abbreviations: TB, Tuberculosis; AIDS, Acquired immunodeficiency syndrome; CD4, Cluster of differentiation 4; HIV, Human immunodeficiency virus.
Acknowledgements
None.
Financial support
None reported.
Competing interests
All authors report no conflicts of interest relevant to this article. This research did not receive any grant from the public, commercial, or not-for-profit sectors. This study was approved by the institutional review board of Baylor College of Medicine and complies with journal ethics policies.
Open access
