Matrix Assisted Laser Desorption
Ionization Time-of-Flight Mass
Spectrometry
Nathan A Ledeboer
Assistant Professor of Pathology
Medical College of Wisconsin and
Medical Director, Microbiology and Molecular Pathology
Dynacare Laboratories and Froedtert Hospital
Milwaukee, WI
1
Outline
• Overview of Technology
• Are MALDI-TOF instruments all they described to
be?
• How can I justify the cost?
• Applications in Development
• What is the Impact for Patient Care
• The Future of Mass Spectrometry
• Conclusions
• Questions
2
Disclosures
• Dr. Ledeboer will discuss products that are not
FDA cleared
• Financial Disclosures
– Consultant:
•
•
•
•
•
Nanosphere, Inc
ThermoFisher Scientific, Inc
LabCorp
Cepheid (Scientific Advisory Board)
iCubate
– Honoraria
• Bruker Daltonics
• Meridian
3
OVERVIEW OF TECHNOLOGY
4
Bruker Biotyper vs Vitek MS
5
Step 1: Target Preparation, continued
Direct Smear Method:
• Touch colony with transfer device, such as toothpick
• Transfer a small amount onto spot
• Let air dry
• Cover with 1 µL of MALDI matrix, let air dry
• Analyze
Research use only – not for use in diagnostic procedures
Ethanol-Formic Acid Extraction
(if required – low score value after direct smear)
Inactivation of
pathogens
dH2O
Ethanol
Ethanol
10 min
Research use only – not for use in diagnostic procedures
Formic acid,
acetonitrile
Analyze
supernatant
Matrix Assisted Laser Desorption/Ionization
Matrix: HCCA (a-Cyano-4-hydroxycinnamic acid)
Solvent: Acetonitrile, TFA (trifluoroacetic acid)
•
•
Lyses cell walls and extract protein
Separates protein molecules (proteins are “sticky”)
1 µL Matrix
Analyte (organism)
Target plate
Research use only – not for use in diagnostic procedures
Matrix Assisted Laser Desorption/Ionization
• Laser light pulses
• Matrix molecules readily absorb laser light (photon energy), creating an excited
energy state
• The matrix is acidic, and donates positive charge to the analytes
Research use only – not for use in diagnostic procedures
Matrix Assisted Laser Desorption/Ionization
• Localized heating causes micro-explosion of material
• Collisions with neutral sample facilitate charge transfer to/from excited matrix
molecules
• Ions “desorb” from the target surface
Matrix
Research use only – not for use in diagnostic procedures
TOF – Time of Flight
Detector
Intensity
•
Following acceleration, the
charged ions are allowed to
drift through a free field
toward the detector
•
The speed of travel (time of
flight) is proportional to the
ion’s mass (smaller ions
reach the detector first)
Research use only – not for use in diagnostic procedures
m/z
Drift region
MALDI: Results output
Raw profile spectrum
Refined profile spectrum
Results are analyzed by a computer, cleaned-up and the spectrum is
12
searched against a database with known spectra.
MALDI identification result
2.0-3.00 Secure genus and species identification
1.7-1.99 Probable genus identification
0.0- 1.69 Unreliable identification
ARE MALDI-TOF INSTRUMENTS ALL
THEY ARE DESCRIBED TO BE?
15
MALDI publications:
More than 200 peer reviewed publications, mainly in high-ranking
microbiology journals as of September 2012
Topics include:
- Particular groups of bacteria
- (e.g. anaerobes, Listeria, Neisseria, Yeast...)
- Routine application
- Rare and difficult to analyse microorganisms (e.g. Prothotheca)
- Highly pathogenic bacteria (e.g. Francisella, Brucella)
- Blood culture direct analysis
- Urine direct analysis
- Sub-typing
- Resistance or Virulence Factors
-Mycobacteria
-Filamentous Fungi
In a study by Benagli
et al., the authors
compared
performance of
MALDI-TOF to
biochemical ID and
resolved discrepancies
with sequencing. The
results follow.
Benagli C et al. PLoS One. 6(1).
17
Clinical Application - Bacteroides species
277 Clinical Bacteroides Isolates from a European study:
270 isolates (97,5%) identified with significant score, 7 isolates not in Reference
Library (e.g. Bacteriodes distasonis)
MALDI ID
discrepant
Bacteroides fragilis
179
5
Bacteroides thetaiotaomicron
43
5
Bacteroides ovatus
15
6
Bacteroides vulgatus
20
1
Bacteroides uniformis
5
3
Bacteroides eggerthii
1
0
Bacteroides nordii
4
3
Bacteroides salyersiae
1
0
Bacteroides massiliensis
2
0
-----------------------------------------------------------------------------------------------------270
23
sequenced*
2
4
1
1
2
1
11
*only IDs with log(score)<2.5
16S rDNA Sequenceing Confirmed 10 of 11 Discrepant MALDI Results, 1 Case
Only “Bacteroides spec.“
Nagy et al., Clin Microbiol Infect 2009; 15: 796–802
Clinical Application: Yeast
No incorrect Yeast
Identification by
the Respective
Molecular
Fingerprint
Marklein et al., JCM, Vol. 47 2009
Bruker Biotyper vs Vitek MS
20
Saramis
System and application (no. of isolates
tested)
Biotyper
% Correctly identified
P value
% Correctly identified
P value
83.8 vs 92.7
<0.001
NAd
NA
86.9 vs 95.5
<0.001
NA
NA
12.8 vs 3.2
<0.001
NA
NA
0.3 vs 1.2
<0.01
NA
NA
83.8 vs 93.2
<0.001
92.7 vs 93.2
0.608
86.9 vs 93.6
<0.001
95.5 vs 93.6
<0.05
12.8 vs 5.8
<0.001
3.2 vs 5.8
<0.01
0.3 vs 0.4
1
1.2 vs 0.4
<0.05
Biotyper
Routine (986)a
Vitek MS
Routine (986)a
Martiny, et al, JCM, 2012, 50
21
Bruker Biotyper vs Vitek MS
Property
User friendliness
Microflex LT
Vitek MS RUO
Vitek MS IVD
Ready-to use Matrix solution
No
Yes
Yes
Facility of preparing smear
Very easy
Easy
Easy
Disposable targets
Yes
Yes
Yes
Reusable targets
Yes
No
No
Software
Easy to use
Not easy to use
Very easy to use
5–10
NDa
Remarks
For Vitek-MS systems, matrix
solution must be deposed each two
spots
Time for 96 identifications
Time to prepare work list (min) <5
Time to load target and make
2
5
40
55
Time for 16 identifications (min) ND
ND
15
IVD
Yes
No
Yes
RUO
Yes
Yes
No
Quality management
Easy
Easy
Very easy
+
+++
++
NAb
NA
NA
++
+
+++
Silent
Smaller
Via LIS
1 × 96
Noisy
Bulkier
NA
4 × 48
Noisy
Bulkier
Via Myla
4 × 48
vacuum
Time for analysis (min)
No ID before success of QC at end of
run (each 16 IDs)
Quality
Need for validation before clinical
reporting
Costc
Device
Reactants
Maintenance
Implementation
Noise
Size
Connectivity
Capacity
Based on catalog prices
Martiny, et al, JCM, 2012, 50
22
Time/ test
(hour)
FTE
Cost/test*
Supply
Cost/test
Total
Cost
Rapid Biochemicals
0.10
$4.14
$0.29
$4.43
Automated
Biochemicals
0.14
$5.79
$9.59
$15.38
Long Biochemicals
0.33
$13.65
$5.32
$18.97
Sequencing
0.73
$30.19
$20.02
$50.21
Mass Spectrometry
0.05
$2.07
$0.24
$2.31
*FTE cost/hour $41.35
Slide courtesy of Robin Patel, MD
23
Cost Savings Comparison
Instrument
Included
* Assumes a 3 year depreciation of instrument and an instrument cost of $200,000
5000
10000
20000
30000
40000
50000
$13.33
$6.66
$3.33
$2.22
$1.66
$1.33
$2.31
$2.31
$2.31
$2.31
$2.31
$2.31
$15.38
$15.38
$15.38
$15.38
$15.38
$15.38
-$0.26
$6.41
$9.74
$10.85
$11.41
$11.74
Cost Savings per year
-$1,300.00
$64,100.00
$296,337.55
$448,005.18
$471,128.03
$484,753.99
Return on Investment (in years)
6.06
3.12
0.67
0.45
0.42
0.41
5000
10000
20000
30000
40000
50000
Tests per Year
Instrument
Add*
MALDI-TOF ID Consumables and
Labor
Biochemical ID Consumables and
Labor
Difference per
test
No Instrument
Tests per Year
MALDI-TOF ID Consumables and
Labor
Biochemical ID Consumables and
Labor
Difference per
test
$2.31
$2.31
$2.31
$2.31
$2.31
$2.31
$15.38
$15.38
$15.38
$15.38
$15.38
$15.38
$13.07
$13.07
$13.07
$13.07
$13.07
$13.07
Cost Savings per year
$65,350.00
$130,700.00
$397,652.14
$539,670.76
$539,670.76
$539,670.76
Return on Investment (in years)
3.06
1.53
0.50
0.37
0.37
0.37
24
Cost-effectivenesss of switching to
MALDI-TOF MS for routine bacterial identification
Galliot O, Blondiaux N, Lorez C, Wallet F, Lemaitre N, Herwegh S and Courcol R
 September 2009

Switched from conventional
biochemicals (Vitek 2 and API)
to MALDI-TOF MS (Bruker)
 Cost analysis performed
Annual Savings = $177, 090
“allowed decrease of 89.3% of the cost of bacterial
identification in the first year.”
In addition:
Decreased waste from 1,424kg to 44kg
Decreased subculture media of $1,102
Decreased sequencing cost of $1,650
JCM epub ahead of print
October 2008September
2009
October 2009September
2010
Isolates Tested
33,320
38,624
Biochemical
Costs
$193,754
$5,374
MALDI-TOF
-
$15,836
TOTAL
$193,754
$21, 210
Avg Cost/ID
$5.81
$.54
COMING APPLICATIONS
26
MALDI– blood culture direct analysis
27
MALDI – blood culture direct analysis
Detailed results and effect of adopted processing
77 samples
ORIGINAL - Thresholds
NEW - Thresholds
93 samples
ORIGINAL - Thresholds
3000
4000
3000
4000
> 2.0
52
49
64
66
> 1.8
1.7 - 2.0
16
21
4
8
< 1.7
9
7
9
3
3000
4000
3000
4000
> 2.0
67,5%
63,6%
83,1%
85,7%
> 1.8
1.7 - 2.0
20,8%
27,3%
5,2%
10,4%
< 1.7
11,7%
9,1%
11,7%
3,9%
NEW - Thresholds
3000
4000
3000
4000
> 2.0
57
68
77
80
> 1.8
1.6 - 1.8
1.7 - 2.0
25
14
9
8
1.6 - 1.8
< 1.6
< 1.7
11
11
7
5
< 1.6
3000
4000
3000
4000
> 2.0
61,3%
73,1%
82,8%
86,0%
> 1.8
1.6 - 1.8
1.7 - 2.0
26,9%
15,1%
9,7%
8,6%
1.6 - 1.8
< 1.6
< 1.7
11,8%
11,8%
7,5%
5,4%
< 1.6
No false positive result in the “yellow“ and “green“ log(score) range
3000/4000 –detection boundary increased for protein to reduce background
Proof Point: Blood Culture Innovative Extraction
99% of tested strains were correctly identified directly
from Blood Culture
Sample Preparation
Positive culture
Water
75%
EtOH
Water
Water
50ul
95°C
30 min
2 ml
100%
EtOH
50ul
ACN + silica
beads, vortex 1
min.
FA
50ul
~60 min.
Sample Preparation
Positive culture
Water
75%
EtOH
Water
Water
50ul
95°C
30 min
2 ml
100%
EtOH
50ul
ACN + silica
beads, vortex 1
min.
FA
50ul
~60 min.
Heat Kill Test
Following heat inactivation step, TREK Myco bottles were inoculated and held 6 weeks.
Days to positivity
18
16
15 Min
14
30 Min
12
10
8
6
4
2
0
1
2
3
4
M. che/abs gr
5
1
2
3
4
5
M. fortuitum gr
1
2
3
MAI
4
5
1
2
3
MTB
4
5
Sample Preparation
Positive culture
Water
75%
EtOH
Water
Water
50ul
95°C
30 min
2 ml
100%
EtOH
50ul
ACN + silica
beads, vortex 1
min.
FA
50ul
~30 min.
Intens. [a.u.]
Effect of Bead-Beat on MALDI Spectrum
8000
AFB15-A 0:A9 MS, BaselineSubtracted, Smoothed
No Bead-beat step
6000
4000
2000
Intens. [a.u.]
0
5000
Mycobacterium sp AFB015 MCW 0:H8 MS, BaselineSubtracted, Smoothed
Bead-beat step
4000
3000
2000
1000
0
5400
5600
5800
6000
6200
6400
6600
6800
m/z
Mycobacteria v1.0 Database
HPLC
MAI (14)
Sequencing ID
M. avium
#
14
Middlebrook 7H10
Trek Myco
< 1.7
1.7 – 2.0
> 2.0
< 1.7
1.7 – 2.0
> 2.0
0
4
10
0
4(1)
10
M. intracellularae
MTBC (6)
M. chelonae/
abcessus
(9)
M. fortuitum
(16)
M. tuberculosis*
6
0
3
3
0
0
6
M. abscessus
5
0
1
4
0
1
4
M. chelonae
3
0
1
2
0
1
2
M. immunogenum
1
0
1
0
0
0
1
M. fortuitum
13
1
9
3
0
4
9
M. peregrinum
2
1
1
0
1
0
1
M. porcinum
1
1
0
0
1
0
0
M. peregrinum
M. conceptionense
M. xenopi (6)
M. xenopi
6
1
2
3
2
2
2
M. szulgai (1)
M. szulgai
1
0
1
0
1
0
0
HPLC
M. gordonae
Sequencing ID
#
Middlebrook 7H10
Trek Myco
< 1.7
1.7 – 2.0
> 2.0
< 1.7
1.7 – 2.0
> 2.0
M. gordonae
3
0
0
3
0
0
3
M. mucogenicum
1
0
1
0
0
0
1
2
0
2
0
0
0
2
5
0
3
2
0
1
4
1
0
0
1
0
0
1
5
2
2
1
0
1
4
M. phocaicum
M.
mucogenicum M. llatzerense
(14)
M. peregrinum
M.muco/phocaicum
M. phocaicum
M. malmoense M. malmoense
M. phlei (1)
M. phlei
M. smegmatis M. porcinum
Total
Agreement (species)
1
0
0
1
0
0
1
1
1
0
0
1
0
0
1
0
0
1
0
0
1
72
7
(9.7%)
31
(43.1%)
34
(47.2%)
6
(8.3%)
14
(19.4%)
52
(72.2%)
80.0%
96.6%
100%
83.3%
92.9%
100%
HPLC
M. gordonae
Sequencing ID
#
Middlebrook 7H10
Trek Myco
< 1.7
1.7 – 2.0
> 2.0
< 1.7
1.7 – 2.0
> 2.0
M. gordonae
3
0
0
3
0
0
3
M. mucogenicum
1
0
1
0
0
0
1
2
0
2
0
0
0
2
5
0
3
2
0
1
4
1
0
0
1
0
0
1
5
2
2
1
0
1
4
M. phocaicum
M.
mucogenicum M. llatzerense
(14)
M. peregrinum
M.muco/phocaicum
M. phocaicum
M. malmoense M. malmoense
M. phlei (1)
M. phlei
M. smegmatis M. porcinum
Total
Agreement (species)
1
0
0
1
0
0
1
1
1
0
0
1
0
0
1
0
0
1
0
0
1
72
7
(9.7%)
31
(43.1%)
34
(47.2%)
6
(8.3%)
14
(19.4%)
52
(72.2%)
85.7%
96.4%
100%
83.3%
92.3%
100%
Data Summary
Middlebrook 7H10
• 90.3% (65/72) with acceptable ID score (>1.7)
 98.3% (59/60) correct species ID
Trek Myco
• 91.7% (66/72) with acceptable ID score (>1.7)
 98.4% (60/61) correct species ID
Clinical Impact
Time:
t = 3-5 days
t=0
t = 1 hr
t = 1 day
t = 1-2 days
ID Filamentous Fungi workflow
1. Direct Transfer of “Front Mycelium“ (1 min)
if successful: ID is FINISHED
6585,21
6744,05
4000
6016,93
e.g. A.niger
Intens. [a.u.]

3000
8049,55
2000
9074,52
4536,96
1000
0
4000
5000
6000
7000
8000
9000
10000
11000
12000
m/z
2. Ethanol Extraction of “Front Mycelium“ (10 min)
 if successful: ID is FINISHED
3. Broth Cultivation (approx. 1 additional day) & extraction
adhering filamentous fungi
 ID is possible for slow or fast sporulating fungi
 ID is possible for every kind of filamentous
 fungi
ALL matches against the SAME Filamentous Fungi DB
 ID is possible for agar
IMPACT ON PATIENT CARE
42
Work flow and delay for matrix-assisted laser desorption ionization time-of-flight (MALDITOF) mass spectrometry identification of bacteria in this study.
Seng P et al. Clin Infect Dis. 2009;49:543-551
© 2009 by the Infectious Diseases Society of America
Tan KE, et al, JCM, In Press – Kindly provided by K. Carroll, MD
44
n
Proportion identified earlier by MALDI-protocol,
identified earlier
by number of days of workup
<0da
(days)
S. aureus
(%)
0db
1d
2d
3d
4d
5d
6d
>6d
(%)
(%)
(%)
(%)
(%)
(%)
(%)
(%)
2.8
0.9
16.7
16.7
109
1.35
1.8
66.1
28.4
Other Staphc
26
1.19
7.7
65.4
26.9
BHSd
72
0.60
38.9
58.3
1.4
VGSe
7
0.57
42.9
57.1
17
1.12
41.2
29.4
S. pneumoniae
6
0.33
66.7
33.3
Other GPCf
6
3.33
33.3
16.7
16.7
78
1.64
1.3
51.3
34.6
9.0
2.6
1.3
284
1.34
2.8
69.4
23.2
2.1
1.1
1.1
0.4
P. aeruginosa
77
1.82
41.6
49.4
2.6
1.3
2.6
2.6
Other NF GNBg
39
2.59
30.8
35.9
2.6
15.4
5.1
5.1
Haemophilus sp.
10
1.40
Other GNCBh
7
0.14
85.7
14.3
Corynebacterium sp.
9
1.67
22.2
33.3
Other GPRi
8
4.13
Anaerobic GNj
26
2.54
3.8
Anaerobic GPk
14
2.64
21.4
14.3
C. albicans
52
0.04
92.3
1.9
S. anginosus
Enterococcus sp.
Enterobacteriaceae
1.4
2.6
5.9
80.0
12.5
3.8
Tan KE, et al, JCM, In Press – Kindly provided by K. Carroll, MD
Organism-group
Mean # of days isolate
23.5
2.6
20.0
22.2
22.2
37.5
12.5
37.5
65.4
7.7
19.2
3.8
28.6
7.1
7.1
14.3
1.9
7.1
45
Patient Impact
• Retrospective chart review of patients with
positive blood cultures obtained during
validation of MALDI-TOF Biotyper and
Sepsityper Kit
• Time to ID (MALDI-TOF v traditional culture
methods) and start/stop times of antibiotics
reviewed
• Theoretical reduction in empiric antibiotic
duration and cost difference (AWP) calculated
Paul J., et al. IDSA. 2011
46
Time to Antimicrobial DeEscalation
Traditional ID
MALDI-TOF
Direct ID
Difference (in
hours)
P
Emperic
Antibiotic
Duration ( IQR,
hrs)
66.6
(44.4-87.5)
15.5
(1.4-21.3)
51.6
(41.8-71.6)
<0.01
Antibiotic Cost
(USD)
$245.24
$88.48
$156.76
<0.01
Paul J., et al. IDSA. 2011
47
THE FUTURE OF MASS
SPECTROMETRY
48
Electrospray Mass Spectrometry
Van Belkum, et al, JCM, 2012, 50
Characteristic
MALDI-TOF
Ionization
Soft ionization with matrix
Fragmentation
No (intact molecules)
Sample
Solid form (or liquid allowed to dry on target)
Molecules
Turnaround time
Mainly proteins, large glycopeptides,
oligonucleotides, carbohydrates
20-30 s per sample at laser frequency of 50 Hz to
generate a spectrum
Minutes from sample preparation to result
Throughput
Disposable target with multiple spots (48 to 96)
Reagent
Chemical matrix
Material
External calibrant
Disposable target
Quantification
Not well suited
Specificity
Depending on MS specificity and proteins tested
Sensitivity
Bacterial ID: 105 CFU using fingerprint approach
Urine sample after purification without culture: 105
CFU/ml
Integration into microbiology
lab workflow
Today's clinical microbiology
applications
LC-ESI-QqQ-MS
Soft ionization with solvents and
electronebulization
Yes
Liquid form (downstream of a liquid
chromatography step)
Different molecules, especially peptides
Minutes or hours depending on liquid
chromatography adsorption/elution times
No batch mode at the moment due to LC step
Chemical reagents for chromatographic
separation and elution
Internal calibrant
Chromatographic column and precolumn
Vials for LC injection
Fully adapted
Usually higher than MALDI when selected
reaction monitoring (MS2) is used
To be explored
Yes (IVD-compliant systems)
No (research applications)
Microbiology: identification of bacteria, yeast, and
molds
None
Quantitative assays for small molecules, such
as vitamin D (outside microbiology field) 50
Microcantilevers
Conclusions
• Data demonstrates excellent performance of
MALDI-TOF MS for identification of bacteria and
yeast from plates and from positive blood cultures.
• High Capital Cost can be overcome by consumable
savings
• Use of technology can result in a significant
reduction in laboratory turnaround and significant
antimicrobial cost savings
• MALDI-TOF and current technologies represent the
beginning of protein revolution
52
Questions
?