Abstract
Background
An early diagnosis and treatment of invasive fungal disease (IFD) is associated with improved outcome, but the moderate sensitivity of noninvasive diagnostic tests makes this challenging. Invasive diagnostic procedures such as bronchoalveolar lavage (BAL) have a higher yield but are not without risk. The detection and sequencing of microbial cell-free DNA (mcfDNA) may facilitate a noninvasive diagnosis.
Materials
In a prospective observational study, we collected plasma in the 120 hours preceding or following a BAL in patients with hematological malignancies suspected for a pulmonary IFD. The EORTC/MSGERC2020 criteria were used for IFD classification. Sequencing was performed by Karius (Redwood City, CA) using their Karius Test (KT) on plasma and a “research use only test” on BAL fluid if available. Cases with a probable/proven IFD were identified based on standard diagnostic tests on serum and BAL (microscopy, polymerase chain reaction, galactomannan, culture) and used to calculate the sensitivity, specificity, and additional diagnostic value of the KT.
Results
Of 106 patients enrolled, 39 (37%) had a proven/probable invasive aspergillosis, 7 (7%) a non-Aspergillus IFD, and 4 (4%) a mixed IFD. The KT detected fungal mcfDNA in 29 (28%) patients. Compared with usual diagnostic tests, the sensitivity and specificity were 44.0% (95% confidence interval [CI], 31.2–57.7) and 96.6% (95% CI, 88.5%–99.1%). Sensitivity of the KT was higher in non-Aspergillus IFD (Mucorales:2/3, Pneumocystis jirovecii: 3/5). On BAL, the sensitivity was 72.2% (95% CI, 62.1–96.3), and specificity 83.3% (95% CI, 49.1–87.5).
Conclusions
Sequencing of mcfDNA may facilitate a noninvasive diagnosis of IFD in particular non-Aspergillus IFD. However, on plasma and similar to currently available diagnostics, it cannot be used as a “rule-out” test.
Keywords: hematological malignancy, invasive fungal disease, invasive fungal infection, microbial cell-free DNA, next generation sequencing
Microbial cell-free DNA sequencing may facilitate a non-invasive diagnosis of invasive fungal disease (IFD) in particular non-Aspergillus IFD. However, similar to currently available diagnostics, it cannot be used as a “rule-out” test. Sequencing was also successfully performed on bronchoalveolar lavage fluid.
Invasive fungal diseases (IFDs) pose a significant threat to immunocompromised patients, particularly those with hematological malignancies or after allogeneic stem cell transplantation [1–3]. An early and accurate diagnosis of invasive fungal infections (IFIs) allows for prompt initiation of antifungal therapy, ultimately reducing IFD-related morbidity and mortality [4, 5]. Currently, routine IFD diagnosis is based on conventional microscopy, fungal culture, biomarkers such as galactomannan (GM) and beta-D-glucan, and to a lesser extent also polymerase chain reaction (PCR)–based DNA detection methods. They are part of the 2020 EORTC/MSGERC consensus definitions that are frequently used to classify IFIs [6]. Unfortunately, the existing diagnostic armamentarium is far from perfect. Conventional microscopy and fungal culture have a low sensitivity, and in addition, fungal culture typically takes days but occasionally much longer to grow [4]. Although GM shows improved performance compared with fungal culture, its efficacy is influenced by the patient population and the material used. The use of serum GM is especially useful in patients with neutropenia but is less validated in other populations. In addition, validation studies showed that the sensitivity of GM in bronchoalveolar lavage (BAL) fluid is higher than in serum at a minor cost of specificity [7]. Similarly, beta-D-glucan displays reasonable sensitivity but lacks specificity because it is produced by many fungi, including Candida, and bacteria [8]. PCR-based tests were recently introduced in the EORTC/MSGERC consensus definitions but cannot distinguish airway colonization from invasive disease [9, 10]. Semi-invasive or invasive sampling methods, such as bronchoscopy or lung biopsy, respectively, have a higher diagnostic yield but are time-consuming and not always possible [11, 12]. It is clear that novel noninvasive diagnostics are needed that facilitate the timely detection of fungal pathogens.
Cell-free DNA (cfDNA) is freely circulating in body fluids including blood. Over the past 2 decades, cfDNA sequencing has been shown to enable noninvasive diagnosis of fetal abnormalities and characterization/monitoring of cancer [13]. More notably, cfDNA originating from microbes (microbial cell-free DNA [mcfDNA]) and extracted from blood plasma has been successfully sequenced as well [14].
Karius Inc., a clinical diagnostics company located in California, has developed an mcfDNA test (The Karius test [KT]) capable of detecting DNA of more than 1000 pathogens including bacteria, DNA viruses, and fungi from 1 mL of plasma [15]. A benefit of this technique is detecting DNA from multiple species simultaneously.
In this multicenter study, we compared the performance of the KT on plasma to standard IFD diagnostics in a cohort of patients with an underlying hematological malignancy in whom an IFD was suspected. Additionally, we extended our investigation to include the sequencing of mcfDNA on BAL fluid in a subgroup of patients to explore the broader potential of the KT in diagnosing IFIs in various bodily fluids.
METHODS
Study Design and Population
This was a multicenter study conducted at 3 tertiary care centers in the Netherlands and Belgium: Erasmus University Medical Center Rotterdam, University Hospital Ghent, and AZ Sint-Jan Bruges. The study involved patients with an underlying hematological malignancy and a clinical and radiological suspicion of an IFD. Written consent was obtained from all patients who planned to undergo or had already undergone a BAL in the previous 5 days and who were included in the ARPOS study (NL62004.078.17). Plasma samples were collected and together with any available leftover BAL fluid were stored at −80 °C for research on new IFD diagnostic methods. The inclusion of patients and the sample collection was therefore prospective, but the KT was performed at a later time. All patients included in the ARPOS study for whom material was available at that time were included in this study.
Diagnostic Criteria
The classification of IFD was based on the 2020 EORTC/MSGERC consensus definitions. All patients met the host factor criterion. Additionally, the radiological criterion for IFD diagnosis was determined based on the chest computed tomography scan closest to the BAL sampling. BAL was performed as part of routine clinical care. More details about the diagnostic tests that were done as part of routine care on serum and/or BAL are described in Supplementary Methods 1.
Microbial Cell-free DNA Sequencing (KT)
BAL fluid and plasma samples were sent under frozen conditions to Karius (Redwood City, CA) for mcfDNA sequencing using a certified courier (World Courier). Because the plasma and BAL samples had been collected in the context of the ARPOS study, testing was not performed in real time. For plasma, the KT details and performance have been previously described [15]. Briefly, following mcfDNA extraction, enriched library preparation, sequencing, discarding of human sequences, and alignment of the remaining sequences to a curated database of reference genomic sequences, test results were generated using the Karius version 3.13 analytic pipeline.
KT reports consisted of all microorganisms deemed to be significantly higher than real-time background control specimens [15]. These microorganisms were quantified in molecules per microliter as an equivalent to the number of DNA fragments per microliter. For BAL fluid, Karius used a research use only pipeline (RUO-KT) using the same process as in plasma cfDNA sequencing aside from a 1:10 predilution, and with a distinct Clinically Reportable Reference list that includes 981 bacteria, fungi, DNA viruses, and parasites that have the potential to cause pneumonia. The relative abundance of each reported potential pathogen was reported and calculated by dividing the number of pathogen DNA reads by median background flora DNA reads. A preliminary reference range for each pathogen was also provided and defined as the 97.5 percentile relative abundance detected in BALs from a cohort of 66 patients clinically adjudicated negative for a probable cause of pneumonia [16]. Finally, “status” is reported as an indicator of relative abundance compared with the reference range for each pathogen. An example of a report of the test on plasma and on BAL from the same patient is available in Supplementary Methods 2 and 3.
Data Collection and Analysis
Clinical data were extracted from electronic health records, including patient characteristics such as age, sex, underlying disease and treatment, neutrophil count, and antifungal treatment. Mycological evidence was collected as part of routine clinical care (BAL microscopy, culture, and galactomannan); Aspergillus PCR and serum galactomannan were also performed centrally at Erasmus University Medical Center Rotterdam. Results from the KT/RUO-KT were classified as positive, negative, or failed. Because the ARPOS study focuses on invasive fungal disease, the potential clinical impact of bacterial, viral, or parasitic DNA that was detected by the KT/RUO-KT was not evaluated in this study. Also, fungal species that were detected but not typically associated with invasive pulmonary disease (eg, Candida species, Penicillium species) were not considered as cases of IFD in this study.
Categorical variables were reported as frequencies, continuous variables were reported as median and quartiles. Results from the KT/RUO-KT were compared with the 2020 EORTC/MSGERC consensus definitions, proven and probable IFD were used as cases, and possible and no IFD were used as controls.
To assess the performance of the KT/RUO-KT, the sensitivity, specificity, positive and negative predictive value, and positive and negative likelihood ratio were calculated along with 95% confidence intervals (CIs). Subgroup analyses were conducted based on the presence of neutropenia, treatment with mold-active antifungals, and patients who recently underwent an allogeneic stem cell transplantation. Finally, the performance of the KT on plasma was calculated separately for the 3 most frequent IFDs: Aspergillus, Mucorales, and Pneumocystis jirovecii.
RESULTS
Plasma samples, diagnostic test results, and clinical data were collected from 109 patients between June 2018 and May 2022. After excluding 3 patients because of the long (>5 days) interval between BAL and plasma sampling, 106 patients were included in the analysis. Sufficient BAL fluid remained to perform the RUO-KT on 34 patients.
Baseline Characteristics
Of the 106 patients, 72 (68%) were male and the median age was 64 (interquartile range: 54–70) years. Fifty-one patients had underlying acute myeloid leukemia and 11 had underlying myelodysplastic syndrome. Twenty-eight (30%) had received a hematopoietic stem cell transplantation. At the time of plasma sampling, the neutrophil count was <500/mL in 52 (49%). Antifungal prophylaxis or therapy was given to 98 of the 106 (92%) patients but in only 25 (24%) was the agent was active against molds. More details are shown in Table 1.
Table 1.
Baseline Characteristics
| Total (N = 106) | Proven/probable Invasive Aspergillosis (N = 35) | Proven/probable Non-Aspergillus IFD (N = 7) | Proven/probable IA + non-Aspergillus IFD (N = 4) | Possible IFD (N = 48) | No IFD (N = 12) | |
|---|---|---|---|---|---|---|
| Age (y)—median (IQR) | 64 (54–70) | 65 (56–69) | 70 (54–72) | 65 (16–68) | 62 (51–70) | 65 (56–75) |
| Male—no. (%) | 72 (68%) | 25 (71%) | 6 (86%) | 2 (50%) | 33 (70%) | 9 (75%) |
| Underlying disease | ||||||
| AML—no. (%) | 51 (48%) | 17 (49%) | 4 (57%) | 1 (25%) | 22 (47%) | 2 (17%) |
| MDS—no. (%) | 11 (10%) | 1 (3%) | … | 3 (75%) | 4 (9%) | 1 (8%) |
| ALL—no. (%) | 8 (7%) | 2 (5%) | … | … | 2 (4%) | 1 (8%) |
| MM—no. (%) | 6 (6%) | 1 (3%) | … | … | 1 (2%) | 3 (25%) |
| CLL—no. (%) | 5 (5%) | … | … | … | … | 2 (17%) |
| DLBCL—no. (%) | 4 (4%) | 3 (8%) | … | … | 3 (6%) | 1 (8%) |
| Other—no. (%) | 21 (20%) | 11 (31%) | 3 (43%) | … | 16 (33%) | 2 (17%) |
| HSCT | ||||||
| Allogeneic HSCT—no. (%) | 22 (24%) | 10 (29%) | 2 (29%) | 2 (50%) | 9 (19%) | 2 (17%) |
| Autologous HSCT—no. (%) | 6 (6%) | 3 (9%) | … | … | 1 (2%) | 2 (17%) |
| Time between HSCT and plasma sampling (mo)—median (IQR) | 8 (1–31) | 3 (1–27) | 14 (14–14) | 16 (14–) | 11 (0–30) | 52 (0–) |
| Neutropeniaa (<500/mL)—no. (%) | 52 (49%) | 19 (54%) | 2 (29%) | 4 (100%) | 26 (54%) | 1 (8%) |
| Diagnosis (standard of care tests)b—no. (%) | ||||||
| No fungal diseasec | 60 (57%) | … | … | … | … | … |
| Probable IA | 35 (34%) | … | … | … | … | … |
| Probable PJP | 3 (3%) | … | … | … | … | … |
| Probable IM | 4 (4%) | … | … | … | … | … |
| Mixed probable IA and PJP | 2 (2%) | … | … | … | … | … |
| Mixed probable IA and IM | 2 (2%) | … | … | … | … | … |
Abbreviations: ALL, acute lymphatic leukemia; AML, acute myeloid leukemia; CLL, chronic lymphatic leukemia; DLBCL, diffuse large B-cell lymphoma; IA, invasive aspergillosis; IM, invasive mucormycosis; HSCT, hematopoietic stem cell transplantation; L-AmB, liposomal Amphotericin B; MDS, myelodysplastic syndrome; MM, multiple myeloma; PJP, Pneumocystis-related pneumonia.
aAround time of plasma sampling (± 48 h).
bDiagnosis is based on modified EORTC/MSGERC criteria. Diagnosis of IM could also be based on a single positive Mucorales PCR on bronchoalveolar fluid and/or plasma that is not implemented in the consensus criteria; 1 patient had a positive Mucorales culture.
cPatients with a possible IA were also accounted for the group “No fungal disease”: No IFD (N = 12), possible IA (N = 48).
Standard of Care Diagnosis
Thirty-nine (37%) patients had a probable IA. A probable mucormycosis was found in 6 (6%) and a probable P jirovecii–related pneumonia (PCP) in 5 (5%) patients. Additionally, a second IFD was diagnosed in 4 (4%) patients with a probable invasive aspergillosis (IA; 2 also had a probable mucormycosis and 2 a probable Pneumocystis-related pneumonia [PJP]).
From the 39 patients with a probable IA, BAL culture was positive in 7 (18%) and BAL GM in 27 (69%). Serum GM was ≥1.0 in only 7 (18%) and ≥0.5 in 14 (36%) patients with a probable IA.
The Aspergillus PCR was positive on BAL from 37/106 (35%) patients, of whom 24 (23%) fulfilled the “PCR positive in duplicate” criterion as required in the EORTC/MSGERC consensus definitions. In 8 patients, the diagnosis of probable IA was solely based on a positive Aspergillus PCR on BAL fluid. So, overall, the PCR and/or GM test on BAL was the only positive mycological test in the majority (25/39) of IA cases.
In the 6 patients with a probable mucormycosis, BAL culture and Mucorales PCR on plasma was positive in 1 (16%); the other patients were diagnosed based on a positive PCR test in plasma (N = 1) or BAL fluid (N = 2) or both (N = 2).
In all 5 patients with a probable PJP, diagnosis was based on a positive P jirovecii PCR on BAL. More details on mycological evidence for IFD diagnosis can be found in Supplementary Data 4 (Supplementary Table 1).
In 60 patients, diagnosis of a probable/proven IFD could not be established based on the EORTC/MSGERC consensus definitions. In 10 of these patients, the Aspergillus PCR was positive; however, the duplicate positive criterion was not met and spare BAL fluid was unavailable so we were unable to repeat the PCR. Finally, 1 patient with a positive Mucorales PCR on BAL was not classified as probable IFD because the patient survived without any antifungal treatment and the cycle threshold value at which the PCR became positive was 40.
Results of the KT
Plasma
The KT was successful in 105/106 (99%) of the plasma samples, meaning that human and/or microbial DNA could be sequenced. Overall, the test was positive for any bacterial, viral, or fungal DNA in 79 (75%) patients. Fungal DNA was detected and successfully sequenced in 29 (27%) of them. In 3 patients, 2 different species were found (1: Aspergillus flavus/oryzae and Rhizomucor pusillus, 2: Aspergillus fumigatus and A oryzae, 3: Penicillium nalgiovense and Debaryomyces hansenii). Fungal DNA from species known to cause invasive pulmonary disease were found in 24 patients. Aspergillus DNA in 17 and Mucorales and Pneumocystis DNA in 4 patients each (in 1 patient both Aspergillus and Mucorales DNA was detected). We refer to Tables 2 and 3 for details on all the detected fungal species, including fungi not typically associated with pulmonary disease.
Table 2.
Identification of Fungal Species by Next-generation Sequencing of Microbial Cell-free DNA (Karius Test on Plasma and Research-only-use Test on BAL Fluid)
| Fungal Pathogens | Plasmab (N = 32) | BALc (N = 20) |
|---|---|---|
|
Aspergillus species − A fumigatus − A oryzae − A flavus/oryzae |
18 (56%) − 15 (47%) − 2 (6%) − 1 (3%) |
12 (60%) − 11 (55%) − 0 − 1 (5%) |
|
Candida speciesa − C albicans − C glabrata − C tropicalis − D hansenii (C famata) |
4 (13%) − 2 (6%) − 1 (3%) − 0 − 1 (3%) |
5 (25%) − 3 (15%) − 1 (5%) − 1 (5%) − 0 |
| Pneumocystis jirovecii | 4 (13%) | 2 (10%) |
|
Mucorales species − Rhizomucor pusillus − Rhizopus microsporus |
4 (13%) − 2 (6%) − 2 (6%) |
1 (5%) − 0 − 1 (5%) |
| Penicillium nalgiovense a | 2 (6%) | 0 |
aIn general, these species are no causative pathogens for invasive pulmonary disease and are therefore not defined as cases of invasive fungal disease in this study.
bIn total, 32 fungal pathogens were found across 106 patients in plasma, in 3 of these patients, 2 fungal pathogens were found.
cIn total, 20 fungal pathogens were found across 106 patients in BAL fluid. In 1 patient, 2 fungal pathogens were found.
Table 3.
Overview of Patients in Which Sequencing of Microbial Cell-free DNA (Karius Test) Showed a Fungal Species + Associated Standard Diagnostics
| Karius testa Plasma |
RUO-Karius Testa BAL | Serum GM | BAL GM | BAL Culture | BAL Aspergillus PCRb | BAL Mucor PCR | BAL Pneumocystis PCRc | BAL Microscopy | Plasma Mucor PCRc | Radiology | ||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | A fum | A fum | 0 | NP | −e | − | NP | NP | − | NP | Noduli + halo | Possible IA |
| 2 | A fum | A fum | 0.1 | 1.52 | A fum | ++ | NP | NP | − | NP | Noduli + halo | Probable IA |
| 3 | A fum | A fum | 0.1 | 0.1 | − | ++ | NP | NP | − | NP | Noduli + halo | Probable IA |
| 4 |
A flavus/oryzae
R pusillus |
A flavus/oryzae | 0.6 | 4.7 | A fum | ++ | − | NP | − | + (35.91) | Nodulary consolidation + wedge-shaped consolidation | Mixed probable IA and IM |
| 5 | A fum |
A fum
C albicans |
0.2 | 2.22 | − | − | NP | NP | NP | NP | Nodule + halo | Probable IA |
| 6 | A fum |
A fum
C glabrata |
0.5 | 6.07 | − | ++ | NP | NP | − | NP | Noduli + halo | Probable IA |
| 7 | P jirovecii |
A fum
P jirovecii |
0.31 | 0.09 | − | ++ | NP | + (29.2) | − | NP | Wedge-shaped consolidation | Mixed probable IA and PJP |
| 8 | A fum | 0 | 0.12 | − | ++ | NP | NP | NP | NP | Cavity | Probable IA | |
| 9 | . fum | 0.1 | 0.79 | − | ++ | NP | NP | NP | NP | Noduli + halo | Probable IA | |
| 10 | P nalgiovense | A fum | 0.1 | 0.19 | − | + | NP | NP | − | NP | Noduli + halo | Possible IA |
| 11 | A fum | 0.1 | 2 | A fum | NP | NP | NP | + | NP | Noduli + halo | Probable IA | |
| 12 | A fum | 0 | 2.6 | − | ++ | NP | NP | − | NP | Noduli + halo | Probable IA | |
| 13 | P jirovecii | 0.1 | 0.25 | − | + | NP | + | − | NP | Centrilobular ground-glass opacities | Probable PJP | |
| 14 | A oryzae | R microsporus | 0.1 | 0.22 | − | − | + (33.5) | NP | − | − | Noduli + halo | Probable Mucormycosis (PCR) |
| 15 | C albicans | 0.1 | 0.06 | − | − | NP | NP | − | NP | Bilateral ground glass | No mycological evidence | |
| 16 | C albicans | 0.1 | 0.2 | − | − | NP | NP | Yeast | NP | Bilateral patchy consolidations | No mycological evidence | |
| 17 | C tropicalis | 0.12 | 0.1 | − | − | NP | NP | NP | NP | Lobar consolidation | No mycological evidence | |
| 18 | R pusillus | NP | 1 | 4.07 | − | ++ | + (45) | NP | − | + (32) | Noduli + halo | Mixed probable IA and IM |
| 19 | A fum | NP | 0.2 | 3.08 | A fum | ++ | NP | NP | − | NP | Nodule without halo | Probable IA |
| 20 | A fum | NP | 0.1 | 0.46 | − | ++ | NP | NP | − | NP | Nodule + air-crescent | Probable IA |
| 21 | C glabrata | NP | 1.6 | 6.5 | − | ++ | NP | NP | − | NP | Nodulary consolidation + halo | Probable IA |
| 22 | A fum | NP | 0.1 | 0.6 | − | ++ | NP | NP | − | NP | Noduli + halo + tree-in-bud | Probable IA |
| 23 | A fum | NP | 12.2 | 7.2 | − | ++ | NP | NP | − | NP | Consolidation + halo | Probable IA |
| 24 | A fum | NP | 0.1 | 1 | − | ++ | NP | NP | − | NP | Noduli + halo | Probable IA |
| 25 | C albicans | NP | 0 | 0 | − | − | NP | NP | NP | NP | Noduli + halo + tree-in-bud | Possible IA |
| 26 | P jirovecii | NP | 0 | 1.3 | − | − | NP | + (31.2) | Yeast | NP | Bilateral peribronchovascular ground glass noduli | Mixed probable IA and PJP |
| 27 | A fum | NP | 1.5 | 0.7 | − | − | NP | + (42.4) | − | NP | Nodulary consolidation + halo + tree-in-bud | Probable IA |
| 28 | P jirovecii | NP | 0.1 | 0.8 | − | Species + | NP | + (30) | − | NP | Wedge-shaped consolidation + bilateral ground-glass opacities | Probable PJP |
| 29 | R microsporus | NP | 0 | 0.08 | Mucor species | NP | NP | NP | + | + (36.4) | Reverse halo | Probable Mucormycosis (culture) |
| 30 |
A fum
A oryzae |
NP | 0.1 | 3.26 | − | ++ | NP | NP | − | NP | Noduli + halo | Probable IA |
| 31d | P jirovecii | NP | 0 | 0.07 | − | − | NP | + | − | NP | Bilateral ground-glass opacities | No IFD found |
| 32 | A fum | NP | 4.2 | 6.8 | − | NP | NP | NP | - | NP | Noduli + halo | Probable IA |
| 33 | R microsporus | NP | 0.1 | 0.3 | − | − | + (36.59) | NP | − | + (38.5) | Wedge-shaped consolidation + ground glass | Probable Mucormycosis (PCR) |
| 34 | A fum | NP | 0.2 | 8 | A fum | ++ | NP | NP | + | NP | Wedge-shaped consolidation + halo | Probable IA |
| 35 | C albicans | NP | 0.1 | 0.1 | − | NP | NP | NP | − | NP | Noduli + halo | Possible IA |
| 36 |
D hansenii
P nalgiovense |
NP | 0 | 0.3 | − | − | NP | NP | − | NP | Noduli + halo + wedge-shaped consolidation | Possible IA |
| 37 | A fum | NP | 0.8 | 2.9 | − | NP | NP | NP | − | NP | Noduli + halo + wedge-shaped consolidation | Probable IA |
Abbreviations: −, the mycological test being performed but the result being negative; A fum, Aspergillus fumigatus; IA, invasive aspergillosis; IFD, invasive fungal disease; IM, invasive mucormycosis; NP, not performed; PJP, Pneumocystis jirovecii pneumonia.
aIn patients 1 to 3, the Karius test on plasma shows a full agreement with the RUO-Karius test on BAL fluid, in patients 4 to 7, there is a partial agreement. In patients 8 to 14, there is no agreement between the Karius test on BAL and plasma. In patients 15 to 37, the Karius test was not performed on BAL fluid.
b Aspergillus PCR was performed in duplicate and this was deemed as positive mycological evidence according to the 2020 EORTC/MSGERC consensus definitions. This is depicted in the table with “++”. If only 1 PCR test was positive, this is depicted as a single “+” symbol.
cA positive Mucorales PCR was deemed as sufficient mycological evidence for the diagnosis of an invasive mucormycosis in combination with clinical and radiological signs.
dIn patient 31, Pneumocystis PCR was positive; however, there was a very low fungal load. There was no radiological suspicion for a PJP. The patient made a full recovery under broad spectrum antibiotic therapy. No treatment for PJP was initiated.
Bronchoalveolar Lavage
The RUO-KT was performed on BAL fluid from 34 patients and was successful for 33. It was positive in 31 (94%): in 17 (51%) fungal DNA was detected whereas in 14 of these the detected fungus was known to cause invasive pulmonary disease (Table 2). Aspergillus DNA was found in 12 patients. In 1 of these 12, P jirovecii DNA was also found, whereas in 2 of these 12 Candida species were found as well. Finally, Mucorales and Pneumocystis DNA was detected in 1 patient each. Results of DNA from fungi not associated with pulmonary disease that was detected by the RUO-KT can be found in Tables 2 and 3.
Comparison of mcfDNA Sequencing on Plasma and BAL
Refer to Table 3 for an overview of all patients in whom fungal DNA was detected and sequenced either in BAL and/or plasma. In 3 of the 14 patients in whom the RUO-KT detected DNA from pathogenic fungi in BAL fluid, the same DNA was detected on plasma (all A fumigatus). In 4 patients, there was a partial agreement; in 3 of them fungal DNA that was detected in plasma (Aspergillus in 3, P jirovecii in 1) was also detected in BAL but additional fungi were detected in BAL as well (Aspergillus in 1, Candida in 2). In 1 patient, both Aspergillus and Mucorales species were found in plasma but only Aspergillus was found in BAL. In 7 patients, no agreement was found between the RUO-KT in BAL compared with the KT in plasma. In 5 of these patients, A fumigatus was detected in BAL fluid, but not in plasma. In 1 patient, P jirovecii was found in BAL, but not in plasma, and in 1 patient BAL fluid showed Rhizomucor microsporus in contrast to plasma, which only showed A oryzae. In 69 patients, no fungal species were detected with mcfDNA sequencing.
Performance of mcfDNA Sequencing
To compare the performance of the KT on plasma and the RUO-KT on BAL with standard of care diagnostics typically available in a tertiary care center, we considered all probable/proven IFD cases according to the 2020 EORTC/MSGERC consensus definitions as true positives, whereas those classified as possible or no IFD were retained as IFD-negative cases (Table 4). Four patients fulfilled the definition for 2 different fungal causes of infection; in total, 109 events were therefore registered for performance analysis instead of 105. According to this analysis, the KT on plasma had a sensitivity of 44.0% (95% CI, 31.2–57.7) and a specificity of 96.6% (95% CI, 88.5–99.1). The sensitivity of the RUO-KT on BAL fluid was 72.2% (95% CI, 49.1–87.5) but the specificity was lower at 88.2% (95% CI, 65.7–96.7).
Table 4.
Diagnostic Performance of Next-generation Sequencing of Microbial Cell-free DNA (Karius Test on Plasma and Research-only-use Test on BAL Fluid) Compared With the Modified EORTC/MSGERC Definitions
| General Performance of the Karius Test | Na | Sensitivity (%) | 95% CI | Specificity (%) | 95% CI | PPV (%) | 95% CI | NPV (%) | 95%CI | LR+ | 95% CI | LR− | 95% CI |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| PLASMA | 109 | 44.0 | 31.2–57.7 | 96.6 | 88.5–99.1 | 78.6 | 60.5–89.8 | 70.3 | 59.7–79.2 | 13.0 | 3.2–52.5 | 0.6 | .45–.74 |
| BAL | 35 | 72.2 | 49.1–87.5 | 88.2 | 65.7–96.7 | 81.3 | 57.0–93.4 | 79.0 | 56.7–91.5 | 6.1 | 1.6–23.3 | 0.3 | .15–.68 |
| PLASMA + BAL (all patients) | 110 | 56.9 | 43.4–69.5 | 94.9 | 86.1–98.3 | 82.9 | 67.3–91.9 | 74.7 | 63.8–83.1 | 11.2 | 3.6–34.6 | 0.45 | .33–.63 |
| PLASMA + BAL (patients with both tests) | 36 | 73.7 | 51.2–88.2 | 88.2 | 65.7–96.7 | 82.4 | 70.2–98.8 | 79.0 | 56.7–91.5 | 6.3 | 1.7–23.7 | 0.30 | .14–.65 |
| Performance of the Karius testb per IFD | N | Sensitivity (%) | 95% CI | Specificity (%) | 95% CI | PPV (%) | 95% CI | NPV (%) | 95% CI | LR+ | 95% CI | LR− | 95% CI |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Probable IA | 105 | 38.5 | 24.9–54.1 | 97 | 89.6–99.2 | 88.2 | 65.7–96.7 | 72.7 | 62.6–80.9 | 12.9 | 3–52.6 | 0.6 | .49–.82 |
| Probable PJP | 105 | 60.0 | 23.1–88.2 | 99.0 | 94.6–99.8 | 75 | 30.1–95.4 | 98.0 | 93.1–99.5 | 60 | 7.5–479 | 0.4 | .14–.18 |
| Probable IM | 105 | 66.7 | 30–90.3 | 100 | 96.3–1 | 100 | 51–100 | 98 | 93.1–99.5 | - | - | 0.3 | .1–1 |
| Performance of the Karius testb in subgroups | N | Sensitivity (%) | 95% CI | Specificity (%) | 95% CI | PPV (%) | 95% CI | NPV (%) | 95% CI | LR+ | 95% CI | LR− | 95% CI |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Neutropenia | 56 | 44.8 | 28.4–62.5 | 96.3 | 81.7–99.3 | 72.2 | 49.1–87.5 | 68.4 | 52.5–80.9 | 12.1 | 1.7–86.4 | 0.6 | .41–.8 |
| AMT | 27 | 50.0 | 26.8–73.2 | 100 | 77.2–100 | 77.8 | 45.3–93.7 | 72.2 | 49.1–87.5 | - | - | 0.5 | .3–.84 |
| No AMT | 75 | 40.6 | 25.5–57.7 | 95.4 | 84.5–98.7 | 76.5 | 52.7–90.4 | 70.7 | 58.0–80.8 | 8.7 | 2.1–36.1 | 0.6 | .46–.84 |
| aSCT | 27 | 50.0 | 28.0–72.0 | 100 | 74.1–100 | 80.0 | 49–94.3 | 64.7 | 411.3–82.7 | - | - | 0.5 | .3–.82 |
Abbreviations: AMT, anti-mold therapy; aSCT, allogeneic stem cell transplantation; IA, invasive aspergillosis; IM, invasive mucormycosis; NPV, negative predictive value; PJP, P jirovecii–related pneumonia; PPV, positive predictive value.
aThe Karius test was successful in 105 patients on plasma and in 33 patients on BAL fluid. If multiple pathogens were found with the standard diagnostic tests or Karius test, patients were counted more than once; therefore, the values in the table are sometimes >105.
bOnly the results from the Karius test on plasma are accounted for in this table.
From the 60 patients without a diagnosis of a probable IFD based on the EORTC/MSGERC definitions, the KT found a possible cause for infection in 3 patients on plasma (Figure 1A). The KT therefore had a possible additional diagnostic value of 2.9% (3/105) in the overall population and 5% (3/60) in the nonprobable subgroup. For the 33 patients in whom the RUO-KT was also performed on bronchoalveolar fluid, the test had a possible additive diagnostic value of 6.1% (2/33) (Figure 1B).
Figure 1.
Additional diagnostic value (ADV) of the Karius Test on plasma and research-only-use test on BAL fluid compared to the EORTC/MSGERC definitions. A: Diagnostic value of the Karius test (mcfDNA) on plasma compared with the modified EORTC/MSGERC consensus definitions (Mucorales PCR added to the definitions for IM). NGS on plasma was successful in 105 patients. In 60 patients, EORTC/MSGERC criteria were not fulfilled for an invasive fungal infection disease (IFID). In 3 (2.9%) of these patients, the KT (mcfDNA test) showed a fungal pathogen with potential to cause an IFIIFD. This was categorized as a possible ADV. Importantly, 3 patients had a co-infection next to an IA; however, only the co-infection could be confirmed by the KT test (Pneumocystis in 2 patients, Mucorales in 1 patient). Additionally, 1 patient had a probable IM, but the KT only showed Aspergillus in plasma. These patients was also categorized in the “UC + mcfDNA+” group. B: Diagnostic value of the research-use-only Karius test (mcfDNA) on bronchoalveolar fluid (BALf) compared with the modified EORTC/MSGERC consensus definitions (Mucorales PCR added to the definitions for IM). NGS on BALf was successful in 33 patients. In 17 patients, EORTC/MSGERC criteria were not fulfilled for an invasive fungal infection disease (IFDI). In 2 (6.1%) of these patients, the KT (mcfDNA test) showed a fungal pathogen with potential to cause an invasive fungal infection disease (IFID). This was categorized as a possible ADV. Importantly, 1 patient was suspected for an Aspergillus and Mucorales co-infection based on the modified EORTC/MSGERC definitions. The KT test could only confirm the Aspergillus species. Abbreviations: ADV, additional diagnostic value; KT, Karius test; mcfDNA, microbial cell-free DNA; UC, usual care.
Subgroup analysis showed that in patients who had undergone an allogeneic stem cell transplantation, the sensitivity of the KT was 50.0%; it was 44.8% in patients with neutropenia. For the 14 patients with a serum galactomannan of ≥0.5 OD, the sensitivity and specificity was 53.3% and 100%. Importantly, the sensitivity of the KT was not impacted by the use of mold-active prophylaxis (sensitivity 7/14, 50%; specificity 13/13, 100%). Finally, the sensitivity of the KT on plasma was higher in patients with a Mucorales or P jirovecii infection compared with those with an IA (4/6 [66.7%], 3/5 [60.0%] versus 15/39 [38.5%], respectively), whereas the specificity was high across fungal species, ranging between 97% and 100%.
DISCUSSION
Diagnosing an IFD is challenging and often not possible without an invasive diagnostic procedure. In this study, we explored the potential of mcfDNA sequencing as a novel noninvasive diagnostic tool in 106 patients with underlying hematological malignancies with a suspected IFD. We compared the performance of the KT on plasma or BAL to a combination of diagnostic tests typically used in these patients. When performed on plasma, the overall sensitivity and specificity was 44.0% and 96.6%. In contrast to the impact that mold-active antifungal therapy can have on the yield of fungal cultures and serum galactomannan testing, this did not impact the yield of the KT [17]. Although the specificity was consistently high across all species and patient subgroups, the sensitivity varied substantially across fungal species. A low sensitivity (38.5%) was observed in those with an IA, whereas only 4 of the 11 cases of a non-Aspergillus IFD were missed (sensitivity 64%). This is in line with previous studies. In a cohort of hematopoietic stem cell recipients with a suspected IFD, Hill and colleagues reported an overall sensitivity of 51%, which is comparable to the 50% sensitivity we observed in this subgroup. Similar to our findings, they observed a higher sensitivity in non-Aspergillus mold infections (79%) than for IA (31%) [18]. Hong et al. also reported that in all 5 proven cases of non-Aspergillus mold infections, the KT was positive [14]. More recently, Hoenigl et al. showed a high sensitivity of the KT in patients with COVID-associated pulmonary aspergillosis [19]. In our study, we found that the KT on plasma potentially had an additional diagnostic value in only 3 (2.9%) patients compared with the already available diagnostic armamentarium based on culture, GM, and PCR. This increased to 6.1% when the RUO-KT was performed on BAL fluid. In a recent study by Bergin et al., plasma mcfDNA testing resulted in a higher additional diagnostic value of 12.1% (95% CI, 7.7–18.0). Because they also took the bacterial and viral DNA sequencing results into account, this higher additive diagnostic value is not surprising [16]. Additional diagnostic value will be higher in centers unable to perform PCR-based testing for Mucorales, Aspergillus, and P jirovecii DNA. Another strategy related to mcfDNA sequencing might be optimizing a cfDNA PCR on serum or plasma instead of sequencing. Currently, a plasma cfDNA fungal panel has also been created by the group of Banaei and colleagues that was able to identify Aspergilli, Mucorales, Pneumocystis, and other species. When a large plasma volume (>2 mL) was used for DNA extraction, sensitivity and specificity were 69.6% (95% CI, 57.9–79.2) and 99.5% (95% CI, 99.0–99.8), respectively. This panel additionally showed a high clinical impact [20] cfDNA PCRs targeting a single species had an even better performance, but at the cost of losing the ability to identify multiple species using 1 test [21, 22].
In this study, and for the first time, the diagnostic performance of the RUO-KT test was evaluated on BAL fluid as well. As expected, the sensitivity was substantially higher on BAL compared with plasma (72.2% vs 44.0%) and fungal DNA detected in plasma was only missed in BAL fluid in 2 patients. Therefore, in cases where conventional diagnostic tests have failed and BAL fluid sampling is pursued, performing the RUO-KT on BAL fluid could prove beneficial. The disadvantage of using BAL fluid is that DNA from commensal flora will often be detected and clinical adjudication will be essential to distinguish invasive pathogens from commensal flora. This may also explain the lower specificity we observed for the RUO-KT on BAL (88.2%) compared with plasma (96.6%). Furthermore, the abundance of commensal DNA could occasionally mask the presence of DNA from invasive pathogens [15].
Our study has limitations. First, the patients included in this study had been enrolled in an ongoing study with a case report form limited to the diagnosis and treatment of IFD. We were therefore unable to look at any additional diagnostic value for bacterial, viral, or parasitic pulmonary infections [16]. Despite being 1 of the largest published studies on the value of the KT in immunocompromised patients with a suspected IFD, the number of patients with an invasive pulmonary aspergillosis (n = 39) still remained relatively small and invasive mucormycosis or PJP was diagnosed in less than a handful of patients. This precludes definite conclusions on the latter pathogens. Second, the decision of the treating physician to test for the presence of non-Aspergillus species (eg, using a Mucorales or Pneumocystis PCR) was not standardized but was typically made in a patient with negative IA diagnostics. We can therefore not exclude that 1 or more invasive mucormycosis or PJP infections were missed.
From a practical point of view, the actual turnaround time of the KT will depend on the transportation time to and the location of the KT site as well as the time between sample arrival and the generated report. Currently, there is only 1 test site in the United States [15, 18]. As with any other fungal diagnostic test, measuring the actual clinical impact of the KT on the outcome of patients with a suspected IFD is not possible based on observational data. Ideally, this should be investigated in future prospective randomized clinical trials but it is unlikely that such studies will be performed in the (near) future. The moderate sensitivity of the KT on plasma in patients with an IA means that it cannot be used as a rule-out test for IA. However, the better sensitivity in non-Aspergillus mold infections suggests that in patients in which the readily available Aspergillus diagnostic tests turn out negative, it may be a sensible next step to use mcfDNA sequencing to look for the presence of DNA from other fungal pathogens as well as fungal mimickers like Nocardia or Legionella.
Future research is necessary to explore other patient populations in whom plasma mcfDNA sequencing may proof useful. These may be patients with a suspicion of deep-seated infections but negative cultures (eg, endocarditis, spondylodiscitis, suspected mycotic aortic aneurysms) [23]. In addition, a decrease in the DNA load over time may serve as a biomarker of treatment response as suggested in a small case series [24]. Finally, future studies should look into the utility of mcfDNA testing in other bodily fluids (eg, joint aspirate, cerebrospinal fluid).
In conclusion, detecting and sequencing fungal mcfDNA in plasma or BAL fluid had a moderate sensitivity but high specificity, with a better performance for non-Aspergillus IFD compared with IA. Further studies should assess its potential clinical impact in diverse clinical scenarios.
Supplementary Material
Contributor Information
Sammy Huygens, Department of Internal Medicine, Section of Infectious Diseases and Department of Medical Microbiology and Infectious Diseases, Erasmus University Medical Center, Rotterdam, The Netherlands; Department of Hematology, Ghent University Hospital, Ghent, Belgium.
Alexander Schauwvlieghe, Department of Hematology, AZ St-Jan Brugge-Oostende Hospital, Bruges, Belgium.
Nick Wlazlo, Department of Hematology, Erasmus University Medical Center, Rotterdam, The Netherlands.
Ine Moors, Department of Hematology, Ghent University Hospital, Ghent, Belgium.
Jerina Boelens, Department of Microbiology, Ghent University Hospital, Ghent, Belgium; Department of Diagnostic Sciences, Ghent University, Ghent, Belgium.
Marijke Reynders, Department of Laboratory Medicine, Medical Microbiology, AZ St-Jan Brugge-Oostende Hospital, Bruges, Belgium.
Ga-Lai Chong, Department of Medical Microbiology & Infectious Diseases, Erasmus University Medical Center, Rotterdam, The Netherlands.
Corné H W Klaassen, Department of Medical Microbiology & Infectious Diseases, Erasmus University Medical Center, Rotterdam, The Netherlands.
Bart J A Rijnders, Department of Internal Medicine, Section of Infectious Diseases and Department of Medical Microbiology and Infectious Diseases, Erasmus University Medical Center, Rotterdam, The Netherlands.
Supplementary Data
Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Notes
Acknowledgments : We acknowledge Ayten Karisli for her help with inclusion of patients at Erasmus Medical Center.
Author contributions. S.H.: data curation, formal analysis, investigation, project administration, writing—original draft paper and review & editing, visualization. A.S.: conceptualization, resources, writing—review & editing. N.W.: resources, writing—review & editing. I.M.: resources, writing—review & editing. J.B.: resources, writing—review & editing. M.R.: resources, writing—review & editing. G.-L.C.: resources, writing—review & editing. C.H.W.K.: resources, writing—review & editing. B.R.: project conceptualization and administration, supervision, writing—original draft paper and review & editing.
Patient consent statement. Written consent was obtained from all patients who were included in the ARPOS study (NL62004.078.17), from which the biomaterial was used for this project. Patients agreed that their biomaterial could be used for the validation of new diagnostic tests for detecting invasive fungal disease. The protocol of ARPOS study (NL62004.078.17) was approved by the Dutch competent authority (CCMO) and the institutional review board (METC) at Erasmus MC.
Financial support. No funding was received for the conduct of this study. Karius performed the test free of charge.
References
- 1. Latgé JP, Chamilos G. Aspergillus fumigatus and aspergillosis in 2019. Clin Microbiol Rev 2019; 33:e00140-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Bhatt VR, Viola GM, Ferrajoli A. Invasive fungal infections in acute leukemia. Ther Adv Hematol 2011; 2:231–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. van de Peppel RJ, Visser LG, Dekkers OM, de Boer MGJ. The burden of invasive aspergillosis in patients with haematological malignancy: a meta-analysis and systematic review. J Infect 2018; 76:550–62. [DOI] [PubMed] [Google Scholar]
- 4. Lestrade PP, Bentvelsen RG, Schauwvlieghe A, et al. Voriconazole resistance and mortality in invasive aspergillosis: a multicenter retrospective cohort study. Clin Infect Dis 2019; 68:1463–71. [DOI] [PubMed] [Google Scholar]
- 5. Ponnaiyan D, Anitha CM, Prakash PSG, et al. Mucormycosis diagnosis revisited: current and emerging diagnostic methodologies for the invasive fungal infection (review). Exp Ther Med 2022; 25:47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Donnelly JP, Chen SC, Kauffman CA, et al. Revision and update of the consensus definitions of invasive fungal disease from the European Organization for Research and Treatment of Cancer and the Mycoses Study Group Education and Research Consortium. Clin Infect Dis 2020; 71:1367–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. D'Haese J, Theunissen K, Vermeulen E, et al. Detection of galactomannan in bronchoalveolar lavage fluid samples of patients at risk for invasive pulmonary aspergillosis: analytical and clinical validity. J Clin Microbiol 2012; 50:1258–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Theel ES, Doern CD. beta-D-glucan testing is important for diagnosis of invasive fungal infections. J Clin Microbiol 2013; 51:3478–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Huygens S, Dunbar A, Buil JB, et al. Clinical impact of polymerase chain reaction-based aspergillus and azole resistance detection in invasive aspergillosis: a prospective multicenter study. Clin Infect Dis 2023; 77:38–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Lamberink H, Wagemakers A, Sigaloff KCE, van Houdt R, de Jonge NA, van Dijk K. The impact of the updated EORTC/MSG criteria on the classification of hematological patients with suspected invasive pulmonary aspergillosis. Clin Microbiol Infect 2022; 28:1120–5. [DOI] [PubMed] [Google Scholar]
- 11. Maertens J, Maertens V, Theunissen K, et al. Bronchoalveolar lavage fluid galactomannan for the diagnosis of invasive pulmonary aspergillosis in patients with hematologic diseases. Clin Infect Dis 2009; 49:1688–93. [DOI] [PubMed] [Google Scholar]
- 12. Imbert S, Meyer I, Palous M, et al. Aspergillus PCR in bronchoalveolar lavage fluid for the diagnosis and prognosis of aspergillosis in patients with hematological and non-hematological conditions. Front Microbiol 2018; 9:1877. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Bohers E, Viailly PJ, Jardin F. cfDNA sequencing: technological approaches and bioinformatic issues. Pharmaceuticals (Basel) 2021; 14:596. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Hong DK, Blauwkamp TA, Kertesz M, Bercovici S, Truong C, Banaei N. Liquid biopsy for infectious diseases: sequencing of cell-free plasma to detect pathogen DNA in patients with invasive fungal disease. Diagn Microbiol Infect Dis 2018; 92:210–3. [DOI] [PubMed] [Google Scholar]
- 15. Blauwkamp TA, Thair S, Rosen MJ, et al. Analytical and clinical validation of a microbial cell-free DNA sequencing test for infectious disease. Nat Microbiol 2019; 4:663–74. [DOI] [PubMed] [Google Scholar]
- 16. Bergin SP, Chemaly RF, Dadwal SS, et al. Plasma microbial cell-free DNA sequencing in immunocompromised patients with pneumonia: a prospective observational study. Clin Infect Dis 2024; 78:775–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Marr KA, Laverdiere M, Gugel A, Leisenring W. Antifungal therapy decreases sensitivity of the Aspergillus galactomannan enzyme immunoassay. Clin Infect Dis 2005; 40:1762–9. [DOI] [PubMed] [Google Scholar]
- 18. Hill JA, Dalai SC, Hong DK, et al. Liquid biopsy for invasive mold infections in hematopoietic cell transplant recipients with pneumonia through next-generation sequencing of microbial cell-free DNA in plasma. Clin Infect Dis 2021; 73:e3876–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Hoenigl M, Egger M, Price J, Krause R, Prattes J, White PL. Metagenomic next-generation sequencing of plasma for diagnosis of COVID-19-associated pulmonary aspergillosis. J Clin Microbiol 2023; 61:e0185922. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Senchyna F, Hogan CA, Murugesan K, et al. Clinical accuracy and impact of plasma cell-free DNA fungal polymerase chain reaction panel for noninvasive diagnosis of fungal infection. Clin Infect Dis 2021; 73:1677–84. [DOI] [PubMed] [Google Scholar]
- 21. Mah J, Nicholas V, Tayyar R, et al. Superior accuracy of aspergillus plasma cell-free DNA PCR over serum galactomannan for the diagnosis of invasive aspergillosis. Clin Infect Dis 2023; 77:1282–90. [DOI] [PubMed] [Google Scholar]
- 22. Moreno A, Epstein D, Budvytiene I, Banaei N. Accuracy of pneumocystis jirovecii plasma cell-free DNA PCR for noninvasive diagnosis of pneumocystis pneumonia. J Clin Microbiol 2022; 60:e0010122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Hogan CA, Yang S, Garner OB, et al. Clinical impact of metagenomic next-generation sequencing of plasma cell-free DNA for the diagnosis of infectious diseases: a multicenter retrospective cohort study. Clin Infect Dis 2021; 72:239–45. [DOI] [PubMed] [Google Scholar]
- 24. Heldman MR, Ahmed AA, Liu W, et al. Serial quantitation of plasma microbial cell-free DNA before and after diagnosis of pulmonary invasive mold infections in hematopoietic cell transplant recipients. J Infect Dis 2023; 229:576–587. 10.1093/infdis/jiad255 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.