Growth Media on Performance of Mycobacteria Identification Using Matrix-Assisted Laser Desorption Ionization–Time of Flight Mass Spectrometry

Abstract

Identification of mycobacterial infections for both Mycobacterium tuberculosis and non-tuberculosis mycobacteria is important for effective patient care. Matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) is a promising tool that is used in many clinical laboratories for the identification of bacteria and yeast. This study evaluates the impact of growth media on the performance of the MALDI Biotyper^® MBT smart MS for mycobacteria identification. Increased rates of identification, particularly in non-rapid growers and pigment producers, and higher confidence scores were generated in mycobacteria isolated from solid agar, rather than liquid broth. Testing each isolate in triplicate can increase yield of detection. Using the Bruker MBT Mycobacteria Kit to process our samples for testing on the Bruker MALDI Biotyper^® instrument generated precise and accurate mycobacteria identification. These findings emphasize the importance of optimizing mycobacterial specimen processing workflows to include appropriate culture media, which can enhance mycobacterial identification and improve diagnostic accuracy.

1. Introduction

Acid-fast bacilli (AFB) are a group of bacteria that resist decolorization by acid alcohol due to their lipid-rich, waxy cell walls containing mycolic acids. The genus of Mycobacterium makes up a majority of the pathogenic AFB known to infect humans. In general, Mycobacterium can be divided into two groups: Mycobacteria tuberculosis complex (MTBC) and non-tuberculous mycobacteria (NTM) [1,2,3]. MTBC includes M. tuberculosis, M. bovis, M. caprae, M. pinnipedii, M. africanum, M. microti, and M. canettii [2,3]. The most encountered NTM species that can cause human infections are the Mycobacterium avium complex (MAC), which includes M. avium and M. intracellulare (also referred to as M. avium-intracellulare and M. chimaera), M. kansasii, M. marinum, M. ulcerans, M. abscessus complex, M. chelonae, and M. fortuitum complex [1].

Accurate and prompt identification of mycobacterial infections is the cornerstone for successful patient care and the containment of disease. Traditionally, for most clinical microbiology laboratories, mycobacteria can be identified on agar plates by their characteristic colony morphology [4]. The Runyon classification is a system used to categorize NTM based on their growth rate (rapid growers defined as growth within 7 days) and ability to produce pigment under light or dark conditions, separating the AFB into groups such as photochromogens, scotochromogens, and non-photochromogens. The caveat is that some AFB may require weeks for visible colonies, and then follow-up biochemical testing detecting for catalase activity, nitrate reduction, niacin accumulation, or tween 80 hydrolysis are performed to obtain a presumptive identification. Molecular approaches, such as line-probe assays, PCR-based hybridization, and sequencing of the 16S rDNA, rpoB, and hsp65 genes can be used to detect AFB [5]. However, many clinical laboratories do not have the resources to routinely perform sequencing because it is labor-intensive and technically complex [4,6,7]. The AccuProbe test, an FDA-approved direct DNA-probe method (Gen-Probe, San Diego, CA, USA) used to detect and identify certain mycobacteria species was discontinued in 2022.

To overcome the challenges of identification of mycobacteria, the use of matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) may be beneficial for clinical laboratories [8]. Mass spectrometry allows for analysis of the mass of ions in the sample. The desorbed ionized molecules are accelerated through a high vacuum and free flight tube, which allows the ions to separate according to their mass-to-charge ratios. The spectrum generated can be compared to spectra from a comprehensive library or database to determine the most likely organism. MALDI-TOF MS has been reliably and broadly used for the identification of many bacteria and yeast. The MALDI- TOF MS technology brings revolutionary benefits to the clinical laboratory in terms of its sensitivity, specificity, rapid turnaround time, low cost, and ability to be high throughput [8,9].

Currently, the common MALDI-TOF MS systems in clinical laboratories used to detect AFB organisms include the Vitek MS (bioMerieux, Craponne, France) and the Biotyper^® (Bruker Daltonics, Bremen, Germany) systems [10,11]. The VITEK MS uses an FDA-approved database that includes only 45 Mycobacterium species. Meanwhile, the Bruker Biotyper^® MBT smart MS system has the MBT Mycobacteria RUO Library that covers 182 Mycobacterium species, but the database is not FDA-approved. Here, we evaluated the effect of the growth media on the performance of the Bruker MALDI Biotyper^® MBT smart MS to identify mycobacteria.

2. Materials and Methods

2.1. Mycobacterial Isolates

Previously identified mycobacteria isolates (n = 124) from commercially available repositories or reference laboratories were used in this study. The breakdown of the isolates is as follows: 10 M. abscessus complex, 38 M. avium complex, 3 M. chelonae, 17 M. fortuitum, 12 M. gordonae, 3 M. immunogenum, 5 M. kansasii, 2 M. marinum, 2. M. mucogenicum phocaicum group, 1 M. simiae, 6 M. smegmatis, 24 M. tuberculosis complex, and 1 M. arupense. The organisms were supplied in frozen stocks and plated onto solid-phase medium, such as the Löwenstein–Jensen (LJ) agar or the Middlebrook 7H11 agar. The plates were incubated at 35 °C in 5% CO₂ atmosphere up to 8 weeks. To generate cultures in liquid broth, a 10 μL inoculation loop of biomass taken from mycobacteria growth on agar media was inoculated into the BACTEC™ MGIT 960™ (Becton-Dickinson, Franklin Lakes, NJ, USA) broth and incubated in the BACTEC™ MGIT 960™ TB System. Growth detected in the MGIT bottles was flagged by the instrument. A subsequent acid-fast stain was performed on positive MGIT bottles to confirm growth in liquid broth.

2.2. MALDI-TOF MS Identification

Mass spectrometry was performed following the manufacturer’s operational procedures for the Bruker MALDI Biotyper^® MBT smart MS instrument, Bruker MBT Mycobacteria Kit [12], and the Bruker Daltonik MBT Compass 4.1 version with the MBT Mycobacteria Module [13,14].

Briefly, for positive liquid cultures, 1.2 mL of liquid medium from the bottom of the tube was collected and transferred into a Suspension Vial. After centrifugation at maximum speed, the pellet was mixed with Washing Solution (300 μL) and Inactivation Reagent (900 μL), followed by incubation for 30 min at room temperature. The suspension was centrifuged at max speed for 2 min and the residual pellet was first mixed with acetonitrile (50 μL) followed by 70% formic acid (50 μL). Suspension was centrifuged at max speed and 1 μL of the supernatant was spotted onto the MALDI target plate to let dry.

Briefly, for isolates grown on solid media, 1 μL inoculation loopful of organisms was transferred into Washing Solution (300 μL) that was placed into a Suspension Vial. Inactivation Reagent (900 μL) was added into the Suspension Vial followed by incubation for 30 min at room temperature. The suspension was centrifuged at max speed for 2 min and the residual pellet was first mixed with acetonitrile (100 μL) followed by 70% formic acid (100 μL). Suspension was centrifuged at max speed and 1 μL of the supernatant was spotted onto the MALDI target plate to let dry.

After the sample was dried on the MALDI target plate, Bruker Matrix HCAA (1 μL) was overlaid onto each spot on the MALDI target plates within 30 min. Bruker HCAA (alpha-Cyano- hydroxyl 4 cinnamic acid) is a prepared solution that is resuspended in Standard Solvent (Acetonitrile 50%, water 47.5% and trifluoroacetic acid 2.5%). Each isolate was spotted three times to test in triplicate. Prepared MALDI target plates were read within 24 h of preparation.

The MBT Compass version 4.1 software with the MBT Mycobacteria Module was used to obtain the identification of the Mycobacterium species. The mass range used for ribosomal protein detection was 2000 to 21,000 daltons. Manufacturer-recommended cutoff scores were used for identification. A ‘high confidence identification’ score ranged from 1.80–3.00; a ‘low confidence identification’ score ranged from 1.60–1.79; ‘no organism identification possible’ had scores from 0.00–1.59.

2.3. Data Analysis

All statistical analyses were performed using GraphPad Prism (version 6.02). Continuous variables were analyzed as mean ± standard deviation and results were reported with corresponding 95% confidence intervals (CIs). Categorical variables were presented as frequencies and percentages. For categorical variables, Fisher’s exact test was employed to evaluate differences between groups in any contingency table cell. Comparisons between two independent groups for continuous variables were conducted using the Student’s t-test for normally distributed data. A p-value of <0.05 was considered statistically significant. To assess variability and precision within datasets, the coefficient of variance (CV) was calculated with formula CV = (σ/μ)*100, where σ represents the standard deviation and μ represents the mean of the dataset. The CV was expressed as a percentage to facilitate comparisons of relative dispersion across different datasets. A CV of <20% suggests that the data exhibits low variability relative to the mean.

3. Results

3.1. Mycobacteria Identification Rates from Growth on Different Media Types

A total of 124 isolates covering 13 unique Mycobacterium species was tested among the three different media types. The diversity of isolates was selected based on the epidemiology of the isolates commonly recovered in our patient population. Specifically, we evaluated mycobacteria belonging to different Runyon classification groups, such as the pigment producers consisting of photochromogens (M. kansasii, M. marinum) and scotochromogens (M. gordonae), as well as rapid growers (M. fortuitum, M. chelonae, M abscessus), the MAC complex and MTBC.

An increase in identification rate was seen when isolates were cultured in agar media compared to liquid broth cultures (Figure 1,

Table 1. Percent of mycobacteria identified from cultured growth on liquid broth versus solid agar (Löwenstein–Jensen, LJ and Middlebrook 7H11 agar).

	Liquid Broth (n/Total, %)	LJ Agar (n/Total, %)	7H11 Agar (n/Total, %)
Rapid Grower	10/13 (76.9)	9/10 (90.0)	6/7 (85.7)
Non-rapid Grower	33/47 (70.2)	26/30 (86.7)	16/17 (94.1)
Pigment Producer	5/11 (45.5)	5/5 (100)	3/3 (100)
Non-pigment Producer	38/49 (77.5)	30/35 (85.7)	19/21 (90.1)
All	43/60 (71.7)	35/40 (87.5)	22/24 (91.7)

). The identification rate from all the isolates grown on solid media was 89.1% compared to the 71.7% from isolates grown from liquid media (p = 0.02). When broken down to different mycobacteria groups, a difference of 19% in identification rate was observed in non-rapid growers (89.3% in solid media versus 70.2% in liquid broth, p = 0.04) and 54.5% was observed in pigment producers (100% in solid media versus 45.5% in liquid broth, p =0.02). A 10% increase in identification rate was observed in non-pigment producers, although this was insignificant. Of note, M. chelonae and M. immunogenum were two species that resulted in no detection after growth on broth media despite detection from agar media. There was no statistically significant difference observed in the identification rate between the two agar media and between liquid broth versus either of the agar media (Table 1,

Table A1. Percent of mycobacteria identified from cultured growth on liquid broth versus Löwenstein–Jensen, LJ and Middlebrook 7H11 agar.

Mycobacteria Species	Liquid Broth	LJ Agar		7H11 Agar
	Detection Rate	Detection Rate	p-Value	Detection Rate	p-Value
Mycobacterium abscessus complex	2/4 (50)	3/3 (100)	0.43	2/3 (66)	1.0
Mycobacterium avium complex	17/22 (77.3)	7/11 (63.6)	0.44	5/5 (100)	0.55
Mycobacterium chelonae	0/1 (0)	2/2 (100)	0.33	-	n/a
Mycobacterium fortuitum	8/8 (100)	4/5 (80)	0.38	4/4 (100)	1.0
Mycobacterium gordonae	3/8 (37.5)	3/3 (100)	0.19	1/1 (100)	0.44
Mycobacterium immunogenum	0/1 (0)	2/2 (100)	0.33	-	n/a
Mycobacterium kansasii	1/2 (50)	1/1 (100)	1.0	2/2 (100)	1.0
Mycobacterium marinum	1/1 (100)	1/1 (100)	1.0	-	n/a
Mycobacterium mucogenicum phocaicum group	2/2 (100)	-	n/a	-	n/a
Mycobacterium simiae	1/1 (100)	-	n/a	-	n/a
Mycobacterium smegmatis	2/2 (100)	2/2 (100)	1.0	2/2 (100)	1.0
Mycobacterium tuberculosis complex	6/8 (75)	9/9 (100)	0.21	6/7 (85.7)	1.0
Mycobacterium arupense	-	1/1 (100)	n/a	-	n/a
Overall	43/60 (72)	35/40 (88)	0.08	22/24 (92)	0.08

Abbreviations: n/a, not applicable for analysis due to low sample size.

).

3.2. Cut-Off Score Analysis of Different Mycobacteria Species

Using the confidence score thresholds as recommended by the manufacturer, the mycobacteria species that were detected with high-confidence scores in >90% of all the tests from broth media were M. kansasii, M. marinum, and M. smegmatis, while high-confidence scores for M. gordonae and MTBC were seen in >80% of the tests (

Table 2. Comparison of mean confidence scores detected for each mycobacteria species isolated from liquid broth or solid agar.

Mycobacteria Species	Liquid Broth (Mean, 95% CI)	CV	Solid Agar (Mean, 95% CI)	CV	p-Value
Mycobacterium abscessus complex	1.7 (1.6,1.8)	3.6%	2.0 (1.8, 2.2)	11.3%	0.007
Mycobacterium avium complex	1.9 (1.8, 1.9)	7.7%	2.0 (1.9, 2.0)	8.6%	0.03
Mycobacterium chelonae	-	-	2.0 (1.8, 2.2)	8.8%	n/a
Mycobacterium fortuitum	1.9 (1.8, 2.0)	9.2%	2.1 (2.0, 2.2)	10.9%	0.001
Mycobacterium gordonae	1.9 (1.8, 1.9)	5.0%	2.0 (1.9, 2.0)	5.4%	0.11
Mycobacterium immunogenum	-	-	1.7 (1.6, 1.8)	4.7%	n/a
Mycobacterium kansasii	2.1 (1.9, 2.4)	4.2%	2.0 (2.0, 2.1)	7.6%	0.13
Mycobacterium marinum	2.2 (2.1, 2.4)	3.4%	2.2 (2.2, 2.3)	2.3%	0.46
Mycobacterium mucogenicum phocaicum group	1.9 (1.6–2.2)	12.5%	-	-	n/a
Mycobacterium simiae	-	-	-	-	n/a
Mycobacterium smegmatis	2.1 (2.0, 2.2)	5.3%	2.0 (2.0, 2.2)	12.7%	0.37
Mycobacterium tuberculosis complex	1.9 (1.8,2.1)	6.4%	2.2 (2.1, 2.2)	8.5%	0.0000030
Mycobacterium arupense	-	-	1.9 (1.8, 2.1)	5.3%	n/a

Abbreviations: CV, coefficient of variance; CI, confidence interval; n/a, not applicable for analysis due to low sample size.

). More mycobacteria species were detected with high-confidence scores when isolated on agar media. The mycobacteria species that were detected with high-confidence scores in >90% of all the tests from agar media were M. fortuitum, M. gordonae, M. kansasii, M. marinum, and MTBC, while high-confidence scores for M. chelonae and MAC isolates were seen in >80% of the tests.

Comparing the actual numerical confidence score revealed that several species show statistically significant higher ranges (95% confidence interval) when tested from agar media compared to broth media (Table 2). Better performance was seen in M. abscessus (1.85–2.15 from agar versus 1.61–1.81 in broth, p = 0.007), MAC (1.91–2.02 from agar versus 1.84–1.93 in broth, p = 0.03), M. fortuitum complex (2.01–2.22 from agar versus 1.82–1.97 from broth, p = 0.002), and MTBC isolates (2.10–2.21 from agar versus 1.87–2.00 from broth, p = 0.0003). Every single species besides M. smegmatis and M. kansasii achieved higher confidence scores when the isolate was grown in agar media, although the difference was not statistically significant.

To further evaluate the effect of the media type on MALDI-TOF MS performance, we compared the confidence scores between isolates grown on LJ and Middlebrook 7H11 agar media (

Table 3. Comparison of mean confidence scores detected for each mycobacterial species isolated from LJ agar versus Middlebrook 7H11 solid agar.

Mycobacteria Species	LJ Agar (Mean, 95% CI)	CV	7H11 Agar (Mean, 95% CI)	CV	p-Value
Mycobacterium abscessus complex	2.0 (1.8, 2.3)	10.3%	1.7 (1.6,1.9)	5.9%	0.01
Mycobacterium avium complex	2.0 (1.9, 2.1)	9.4%	1.9 (1.8, 2.0)	7.0%	0.34
Mycobacterium chelonae	2.0 (1.8, 2.2)	8.8%	-	-	n/a
Mycobacterium fortuitum	2.1 (1.9, 2.3)	13.8%	2.1 (2.0, 2.3)	6.7%	0.82
Mycobacterium gordonae	2.0 (1.8, 2.1)	6.2%	1.9 (1.7, 2.1)	3.7%	0.93
Mycobacterium immunogenum	1.7 (1.6, 1.8)	4.7%	-	-	n/a
Mycobacterium kansasii	2.2 (2.0, 2.4)	2.8%%	1.9 (1.8, 2.0)	4.0%	0.003
Mycobacterium marinum	2.2 (2.0, 2.3)	2.3%	-	-	n/a
Mycobacterium mucogenicum phocaicum group	-	-	-	-	n/a
Mycobacterium simiae	-	-	-	-	n/a
Mycobacterium smegmatis	2.3 (2.1, 2.3)	3.2%	1.8 (1.6, 2.0)	9.8%	0.001
Mycobacterium tuberculosis complex	2.2 (2.2, 2.3)	3.9%	2.0 (1.9, 2.1)	10.7%	0.001
Mycobacterium arupense	1.9 (1.6, 2.1)	5.3%	-	-	n/a

Abbreviations: CV, coefficient of variance; CI, confidence interval; n/a, not applicable for analysis due to low sample size.

). Higher confidence scores were achieved from the LJ media compared to Middlebrook 7H11 agar, especially for M. abscessus (1.8–2.3 from LJ agar versus 1.6–1.9 from 7H11 agar, p = 0.01), M. kansasii (2.0–2.3 from LJ agar versus 1.8–2.0 from 7H11 agar, p = 0.003), M. smegmatis (2.1–2.3 from LJ agar versus 1.6–2.0 from 7H11, p = 0.001), and MTBC isolates (2.2–2.3 from LJ agar versus 1.9–2.1 from 7H11, p = 0.001). Of note, 11/13 (85%) of the mycobacteria species grown on LJ agar compared to the 4/7 (57%) of species grown on Middlebrook 7H11 agar had a confidence score (95% CI range) that was within the ‘high confidence’ range as determined by the manufacturer.

3.3. Precision of MALDI-TOF Identification for Mycobacteria

Across all three media types, reproducibility was the highest for M. chelonae, M. kansasii, M. marinum, M. immunogenum, M. mucogenicum, M. smegmatis, and M. arupense as seen by the correct identification achieved in all three triplicates (

Table 4. Recovery and identification of mycobacteria in replicate MALDI-TOF MS runs.

Mycobacteria Species	Detection in Technical Replicates (n, %)
	1/3 Spot (#/Total, %)	2/3 Spots (#/Total, %)	3/3 Spots (#/Total, %)
Mycobacterium abscessus complex	2/7 (28.6)	2/7 (28.6)	3/7 (42.8)
Mycobacterium avium complex	2/30 (6.7)	7/30 23.3)	21/30 (70)
Mycobacterium chelonae	0/2 (0)	0/2 (0)	2/2 (100)
Mycobacterium fortuitum	1/16 (6.3)	1/16 (6.3)	14/16 (87.4)
Mycobacterium gordonae	0/7 (0)	1/7 (14.3)	6/7 (85.7)
Mycobacterium immunogenum	0/2 (0)	0/2 (0)	2/2 (100)
Mycobacterium kansasii	0/4 (0)	0/4 (0)	4/4 (100)
Mycobacterium marinum	0/2 (0)	0/2 (0)	2/2 (100)
Mycobacterium mucogenicum phocaicum group	0/2 (0)	0/2 (0)	2/2 (100)
Mycobacterium simiae	1/1 (100)	0/1 (0)	0/1 (0)
Mycobacterium smegmatis	0/6 (0)	0/6 (0)	6/6 (100)
Mycobacterium tuberculosis complex	1/21 (4.8)	0/21 (0)	20/21 (95.2)
Mycobacterium arupense	0/1 (0)	0/1 (0)	1/1 (100)

). For mycobacteria grown on agar media, isolates of M. abscessus (n = 2), M. gordonae (n = 1), M. fortuitum (n = 1), and MAC (n = 4) were identified in <2 replicates, whereas for mycobacteria grown in liquid broth, isolates of M. abscessus (n = 2), M. fortuitum (n = 1), M. simiae (n = 1), MAC (n = 5), and MTBC (n = 1) were identified in <2 replicates (Table 4 and

Table 5. Comparison of recovery and identification of mycobacteria in replicate MALDI-TOF mass spectrometry runs from isolates grown on different growth media.

Mycobacteria Species	Liquid Broth Technical Replicate (#/Total, %)			LJ Agar Technical Replicate (#/Total, %)			7H11 Agar Technical Replicate (#/Total, %)
	1/3 spot	2/3 spots	3/3 spots	1/3 spot	2/3 spots	3/3 spots	1/3 spot	2/3 spots	3/3 spots
Mycobacterium abscessus complex		2/2 (100)		1/3 (33)		2/3 (67)	1/2 (50)		1/2 (50)
Mycobacterium avium complex	2/17 (11.8)	3/17 (17.6)	12/17 (70.6)		2/8 (25)	6/8 (75)		2/5 (40)	3/5 (60)
Mycobacterium chelonae						2/2 (100)
Mycobacterium fortuitum		1/8 (12.5)	7/8 (87.5)			4/4 (100)	1/4 (25)		3/4 (75)
Mycobacterium gordonae			3/3 (100)		1/3 (33)	2/3 (67)			1/1 (100)
Mycobacterium immunogenum						2/2 (100)
Mycobacterium kansasii			1/1 (100)			1/1 (100)			2/2 (100)
Mycobacterium marinum			1/1 (100)			1/1 (100)
Mycobacterium mucogenicum phocaicum group			2/2 (100)
Mycobacterium simiae	1/1 (100)
Mycobacterium smegmatis			2/2 (100)			2/2 (100)			2/2 (100)
Mycobacterium tuberculosis complex	1/6 (16.7)		5/6 (83.3)			9/9 (100)			6/6 (100)
Mycobacterium arupense						1/1 (100)

). Of note, some isolates of M. abscessus, M. fortuitum, M. simiae, MAC, and MTBC were only identified in one of the three replicates. There were no significant differences between the two agar media types. Our data suggest that to increase the yield of detection, each isolate should be run multiple times during the same run. The coefficients of variance for all species detected regardless of growth on different media type were all <20%, suggestive of high consistency and stability (Table 2 and Table 3).

It must be noted that in our study, we did not observe any incorrect identifications despite low-confidence scores. All the mycobacteria identified were concordant with expected result. However, on one of the spots for one of the MAC isolates, an identification of M. timonense with a confidence score of 1.72 was made, while the other spots resulted in MAC with a score of 1.75. Similarly, an identification of M. senegalense was reported with a confidence score of 1.64 on one spot, while the other two resulted in the M. fortuitum group with confidence scores < 1.70. In all these scenarios, the reported species is a part of the respective mycobacteria group/complex in the expected result.

4. Discussion

In this study, we accurately identified MTBC and NTM of pathogenic concern using the Bruker MBT Mycobacteria Kit on the Bruker MALDI Biotyper^® MBT smart MS. The performance appeared to be better on isolates cultured from solid agar media, particularly for non-rapid growers and pigment-producing species. Higher confidence scores obtained from isolates grown on solid media further emphasize the critical role of media selection in ensuring accurate identification. This is particularly relevant for organisms such as Mycobacterium abscessus, the MAC complex, M. fortuitum, and M. tuberculosis, where statistically significant improvements in MALDI-TOF MS scores were observed from isolates grown on agar media. Between the two agar media tested, M. abscessus and M. kansasii showed improved identification on LJ agar, while M. smegmatis and MTBC yielded better results on Middlebrook 7H11. Overall, isolates grown on LJ agar were identified with higher confidence scores.

Other studies have compared the effects of growth media on the performance of the MALDI-TOF MS for mycobacteria identification, and results are mixed. Similar to ours, several have shown a decrease in identification from 13–20% when isolates were tested from broth compared to solid media, while Buchan et al. demonstrated higher confidence scores for mycobacteria when tested directly from broth cultures [10,11,15]. Such discrepancies can be attributed to the broth culture used. In our study, we used the BACTEC™ MGIT 960™ culturing system, which has been shown to produce lower identification rates. Additionally, variations in protein extraction protocols (with formic acid and acetonitrile, bead-beating), matrix formulation (e.g., sinapinic acid versus HCCA), and the exact biomass of the specimen varied among the studies [13,15,16,17]. In this study, we did not evaluate the actual quantity of cells as we followed the manufacturer’s instruction-for-use, stating that positive MGIT liquid broth can be processed for MALDI-TOF MS; however, further studies evaluating the actual quantity of cells may help optimize the performance of MALDI-TOF MS from MGIT systems. For example, a 1 µL inoculation loopful of organisms used for cultured isolates on solid agar contains approximately 10⁷–10⁸ CFU, while positivity of liquid cultures on the BACTEC™ MGIT 960™ System detects approximately 10⁵–10⁶ CFU/mL. Other variables that can affect MALDI-TOF performance include incubation time, growth stage of the organisms, presence of biofilms or organism clumping, and other confounding interfering substances in enriched broth media [18,19].

Advantages of using the MALDI-TOF MS for mycobacteria identification include rapid turnaround time, lower cost, and ease of use [8]. A particular benefit of the Bruker MALDI Biotyper MBT HT Mycobacteria Module and Bruker MBT Mycobacteria Kit used here is that the manufacturer developed an optimized kit and software protocol allowing for streamlined identification of mycobacteria species. One of the primary functions of the Bruker MBT Mycobacteria Kit is to clean up the sample using metal cylinders to break up cellular clumps before using a stringent washing solution followed by full extraction protocol. These tools mitigate any issues due to contamination or user designed acquisition parameters, thereby creating a standardized methodology that can be easily implemented in clinical laboratories without the expertise. However, most of the technical details are proprietary.

For MALDI-TOF MS to generate an identification, not only is biomass important, but misidentification can occur if the culture is not pure. Given the slow growth rate of mycobacteria, a technologist may need to sweep quite a bit of the colonies, assuming the culture is pure [20]. In comparison to PCR, which may only require a scant amount of organism, MALDI-TOF MS needs a much greater amount of organism because of the lack of a preamplification step. In the event of a mixed culture consisting of both rapid and slow growers or contamination with bacteria, incorrect or lack of detection may occur. However, in our study, it is worth noting that we did not encounter any instance of incorrect identification, but we did note that there were some Mycobacterium species that were only detected in one or two of three spots. This suggests that to ensure detection of the organism, laboratories should include at least three technical replicates to capture all species. Not only will this increase the potential for diagnostic yield, but in the event where differing identifications are obtained as results, correlations with the confidence score and repeatability of the result can aid in the final diagnosis. Another limitation is that the identification of new isolates is possible only if the spectral database contains peptide mass fingerprints of the type strains of specific genera, species, subspecies, etc. Another well-known challenge with the use of MALDI-TOF MS for mycobacteria is the need to render the organism nonviable before testing outside a BSL3 laboratory to prevent possible lab staff exposure. A streamlined, standardized protocol consisting of an inactivation and extraction step, such as using the Bruker MBT Mycobacteria Kit on the Bruker MALDI Biotyper^® MBT smart MS, that has a library of >180 mycobacteria organisms enables the production of a higher-quality spectra for diverse identification and ensures the safety of our laboratory staff [21,22].

While the manufacturer recommends confidence scores of >1.8, our study and other studies conducted on yeast and bacteria suggest that manufacturer-recommended score thresholds may be too stringent and scores as low as 1.6 show concordance with identification generated by sequencing [10,11,23,24]. Given that mycobacteria are slow-growing organisms with big clinical implications, laboratories should evaluate as part of their validation if they are comfortable with lowering the scores and reporting criteria to benefit their patient population. Another suggestion is resulting species in the low-confidence identification score can be reported with consultation of the attending pathologist on a case-by-case basis. In our study, more reads in the ‘high confidence’ range were seen when the isolates were grown on LJ agar compared to Middlebrook 7H11, which warrants further studies into media effects. Additionally, laboratories should decide if they are comfortable with combining certain species within a particular subgroup as a group and how reporting affects downstream clinically actionable changes.

Limitations of this study include the use of banked clinical isolates rather than prospective clinical specimens. Additionally, the diversity of the isolates and species tested was relevant for our patient population and may not be applicable to other institutions and geographical demographics. Future studies should explore additional variables that may influence MALDI-TOF MS performance, including incubation times, growth conditions, sample preparation approaches, protein extraction techniques, and spectral database expansion. Incorporating a broader range of clinically relevant mycobacterial species into the databases may further improve its diagnostic utility. Additionally, the potential to support real-time identification of mixed mycobacterial populations warrants further investigation, particularly in polymicrobial infections.

5. Conclusions

The ability to accurately identify mycobacteria is critical for guiding appropriate antimicrobial therapy and facilitating infection control measures. MALDI-TOF MS offers a rapid and cost-effective alternative to traditional methods, such as phenotypic testing and molecular assays, if media optimization is incorporated into the workflow. The growth media used for mycobacteria significantly impact the performance of MALDI-TOF MS for mycobacterial identification. Growth on solid media as opposed to liquid broth cultures lead to higher identification rates and higher confidence scores. By optimizing media selection and workflow protocols, MALDI-TOF mass spectrometry technology can serve as a powerful tool for the rapid and reliable identification of mycobacteria, ultimately improving patient care and public health outcomes. Standardized and optimized procedures offered by commercially available MALDI-TOF MS industries, such as the Bruker MBT Mycobacteria Kit, can be easily implemented in clinical laboratories to aid mycobacteria identification.