Journal of Mass Spectrometry and Advances in the Clinical Lab 28 (2023) 82–90

Contents lists available at ScienceDirect

Journal of Mass Spectrometry and
Advances in the Clinical Lab
journal homepage: [Link]/journal/journal-of-mass-
spectrometry-and-advances-in-the-clinical-lab

Review

Quantitative tandem mass spectrometry in the clinical laboratory:

Regulation and opportunity for validation of laboratory developed tests
Judith A. Stone a, 1, J. Grace van der Gugten b, 1, *
a
Department of Clinical Laboratories, University of California San Francisco Health, San Francisco, CA, United States
b
Office of the Chief Medical Examiner, Government of Alberta, Edmonton, AB, Canada

A R T I C L E I N F O A B S T R A C T

Keywords: Tandem mass spectrometry is an important analytical tool for clinical laboratories, but tests developed and
Liquid chromatography-tandem mass spec­ validated in-house (laboratory developed tests, or LDTs) require special consideration. In late 2022, the forecast
trometry for United States (U.S.) federal regulation of LDTs changed unexpectedly when the VALID Act was not passed by
Method validation
the U.S. Congress. This Act would have modified the Food and Drug Administration’s (FDA’s) role to increase
Laboratory diagnosis
regulatory oversight for LDT providers. In this revised context, we review optimization of quantitative mass
spectrometry LDT validation and suggest avenues other than an additional FDA mandate to achieve uniform best
practice. Common challenges, logistical barriers, and recommendations for easing the burden of best-quality
quantitative mass spectrometry LDT method validation are discussed.

LDT method validation and regulation in the U.S. premarket approval [4,5]. There are additional requirements for vali­
dation of LDTs as compared to verification of FDA-approved assays
The FDA defines a laboratory developed test (LDT) as “an in-vitro (Table 1). Organizations with deemed status from CMS for medical
device (IVD) that is intended for clinical use and designed, manufac­ laboratory accreditation (e.g., College of American Pathologists,CAP)
tured, and used within a single laboratory” [1]. Most medical laboratory may have additional validation criteria for quantitative liquid
tests performed in the U.S. are commercially available from IVD com­ chromatography-tandem mass spectrometry (LC-MSMS) LDTs
panies, and have been regulated by the FDA as medical devices since (QntLCMS-LDTs) that are not specified for LDT by Clinical Laboratory
passage of the 1976 Medical Device Amendments (MDA) to the 1938 Improvement Amendments of 1988 (CLIA’88). One example is the LC-
Federal Food, Drug, and Cosmetic Act (FFDCA) [2]. The FDA did not MSMS matrix effect validation protocol in the CAP Chemistry/Toxi­
define the FFDCA/MDA as applicable to LDT regulation until 1992 and cology checklist [6]. The 2021 CAP checklist question CHM.18825
then chose to exercise a policy of “enforcement discretion” [2]. In 2014, specifies that “10 different sources of native patient matrix should be
however, the FDA changed course and issued draft guidance that would tested. If the mean matrix effect is > +/-25% or the matrix effect CV is
have ended enforcement discretion and required premarket FDA >15% – data demonstrating that accuracy is not affected is required”.
approval of LDTs, among other provisions [3]. . The complex history of LDT regulation in the U.S. has been reviewed
Significantly, CLIA’88 regulations administered by Centers for in the scientific literature [2,3] and discussed from a clinical diagnostics
Medicare & Medicaid Services (CMS) to govern U.S. medical labora­ regulatory and patient advocacy perspective [7–12]. A simplified
tories only require the laboratory director offering the LDT to grant overview is that medical laboratory stakeholders have successfully

Abbreviations: AACC, American Association for Clinical Chemistry; AMP, Association for Molecular Pathology; CAP, College of Amercian Pathologists; CLIA’88,
Clinical Laboratory Improvement Amendments of 1988; CLSI, Clinical Laboratory Standards Institute; CMS, Centers for Medicare & Medicaid Services; CoA, Cer­
tificate of Analysis; FDA, Federal Drug Administration; FFDCA, Federal Food, Drug and Cosmetics Act; GC–MS, Gas Chromatography-Mass Spectrometry; HELP, U.S.
Senate Committee on Health, Education, Labor & Pensions; IA, Immunoassay(s); IVD, In-vitro Device; LC-MSMS, Liquid Chromatography-Tandem Mass Spec­
trometry; LDT, Laboratory Developed Test; MDA, Medical Device Amendments; MLS, Medical Laboratory Scientist(s); MSACL, Mass Spectrometry & Advances in the
Clinical Laboratory; QntLCMS-LDT, Quantitative LC-MSMS LDT; SOP, Standard Operating Procedure; VALID, Verifying Accurate Leading-edge IVCT Development
Act of 2021; VITAL, Verified Innovative Testing in American Laboratories Act of 2021.
* Corresponding author.
E-mail address: [Link]@[Link] (J.G. van der Gugten).
1
Authors contributed equally to this work.

[Link]
Received 14 December 2022; Received in revised form 28 February 2023; Accepted 1 March 2023
Available online 5 March 2023
2667-145X/© 2023 THE AUTHORS. Publishing services by ELSEVIER B.V. on behalf of MSACL. This is an open access article under the CC BY-NC-ND license
([Link]
J.A. Stone and J.G. van der Gugten Journal of Mass Spectrometry and Advances in the Clinical Lab 28 (2023) 82–90

Table 1 automated instrumentation or software for intermediate or final inter­

Validation/Verification required by CLIA’88 and CAP to perform LDT versus pretation.” [7].
FDA approved assays.
Analytical FDA CLIA’88 LC-MSMS LDT (no LC-MSMS Background and goals of the review
Parameter approved LDT specifics in CLIA’88)

Precision X X X This review addresses a narrow segment of the “traditional” LDT

Accuracy X X X universe in depth. We describe infrastructure gaps and list common
Sensitivity X X X problems (and suggested solutions) encountered while developing and
Specificity X X
validating traditional QntLCMS-LDT. New concepts are proposed to
Reportable Range X X X
Carryover X X support the laboratory medicine community and to enhance feasibility
Reference Range Verification Establish Establish of best quality QntLCMS-LDT method development and validation.
Sample/Materials X X Ongoing validation of every batch (series) after a method is in use for
Stability patient testing [17], as well as revalidation of methods in use following
Matrix Effect X (CAP)
modifications, are important related issues outside the scope of this
review.
opposed measures put forward since 2014 that would have authorized We approach the topic from the stance of 15–24 years of practice
FDA regulation of LDT. The contrasting perspective, in favor of LDT with QntLCMS-LDT validation at community hospital, academic medi­
regulation by the FDA, is supported by public health, IVD cal center, regional health organization, CRO, and reference laboratory
manufacturing, and patient advocacy groups [8]. The Verifying Accu­ venues. We have also taught introductory QntLCMS-LDT to MLS, pa­
rate Leading-edge IVCT Development (VALID) Act, reintroduced in the thology residents, clinical chemistry fellows, and laboratory directors in
U.S. Senate (S.2209) and House (H.R.4128) in 2022, would have training programs and at MSACL/AACC workshops. Our knowledge is
authorized FDA oversight of LDT [11]. The VALID Act was thought inevitably anecdotal, but represents a reasonable cross-sampling of
likely to pass in 2022 due to its linkage to other “must pass” legislation, QntLCMS-LDT validation practice. We hold the activist view that addi­
but it was unlinked and did not pass [10,13]. The Verified Innovative tional requirements for QntLCMS-LDT validation and approval are
Testing in American Laboratories (VITAL) Act is a contrasting bill sup­ appropriate. Method development and method validation are addressed
ported by many LDT providers that defines LDT regulatory control to be in tandem, as we view them as inextricably linked.
through CMS under CLIA’88, not by the FDA [14]. However, the most
recent congressional action recorded for the VITAL Act was in 2021. Are QntLCMS-LDT Necessary?
Comprehensive U.S. federal regulation of all IVDs, addressing both LDT
providers and IVD manufacturers, is still a work in progress as of The primary goal of any clinical laboratory test is to help diagnose
February 2023. and correctly treat patients whose samples are being tested; therefore,
LDT providers and associated professional organizations have argued accuracy of laboratory tests is critical. Superior accuracy is one reason
against FDA regulation of LDTs as unnecessary and unfeasible, with little why LC-MSMS has become a key piece of equipment in the clinical
consensus on an alternative [3,12,15]. Some support an enhanced laboratory. Notoriously, low-level testosterone is inaccurate when
version of CLIA’88 rules as an achievable and sufficiently rigorous measured by immunoassay (IA) [18], but excellent accuracy is achieved
pathway [16]. Academic medical center laboratories are seen as when measuring by LC-MSMS [19]. Aldosterone, when measured by IA
particularly vulnerable to more regulation because they tend to offer in patients with chronic kidney disease, is overestimated by as much as
LDTs and may not have the financial and personnel resources for FDA two-fold, a phenomenon not seen when measured by LC-MSMS [20].
pre-market submissions [2]. LDT providers’ specific objections to FDA Thyroglobulin (Tg), a thyroid cancer biomarker, is incorrectly under­
oversight include the following [3]. It is notable that the current VALID estimated by IA when autoantibodies are present. This is a significant
Act has been modified in several ways to address input from the medical problem when monitoring for cancer recurrence, and LC-MSMS Tg as­
laboratory community on these issues [12]: says can overcome this problem [21]. Clinical LC-MSMS methods have
elucidated problems with IA that have caused misleading results inter­
1. Insufficient resources of LDT providers to create and maintain pretation [22] and have led to assay improvements by the manufacturer
“medical device manufacturing quality system requirements” and [23].
supply additional validation testing, documentation and payment of Additionally, LC-MSMS assays can be developed quickly and vali­
fees required for FDA premarket submission. dated in a reasonable timeframe for a wide range of compounds. Com­
2. Inadequate staffing at the FDA to perform premarket review of LDT mercial FDA-approved tests may not be available for niche or new
in a timely manner. compounds, or compounds with low volume workloads. In these cir­
3. Laboratories might have to discontinue LDT and/or avoid developing cumstances, LC-MSMS has been used where small-volume, esoteric tests
new LDT with potentially negative consequences for patients. are not profitable for diagnostic companies (IVD manufacturers) to
develop, but which are clinically useful, such as therapeutic drug
The contrasting viewpoint that supports FDA involvement focuses on monitoring for antifungals or anticoagulant therapies, or for cancer
increased risk to patient safety from unreliable LDTs, seen as a conse­ therapies, such as busulfan [24].
quence of insufficient regulation and enforcement. One driver for this While LC-MSMS offers advantages over more traditional IA-based
perspective was the expansion of for-profit nucleic acid and new laboratory techniques, LDTs are a necessity for implementation of
biomarker LDT testing in the 2000s [8]. Genetic and direct-to-consumer clinical LC-MSMS tests. Manufacturers of LC-MSMS platforms do not
LDTs have a much larger market share compared to the LDT menu typically supply FDA-approved tests; they may provide information and
available when CLIA’88 was first implemented. Legacy LDTs ordered by materials such as LC and tandem mass spectrometry (MSMS) parameter
a medical provider for diagnosis or monitoring and performed in a starting points, reagents, calibrators, and QC materials, but these are
healthcare setting appear to be of less concern, as the FDA has proposed limited to a small test menu. Furthermore, manufacturers often do not
grandfathering approval and exemption from most oversight for so- have access to native matrix clinical patient samples, and vendor
called “traditional” tests that “use components that are legally mar­ methods are typically developed using a simplified matrix that does not
keted for clinical use and whose output is the result of manual inter­ contain endogenous interferences seen in genuine patient samples. As
pretation by a qualified laboratory professional, without the use of such, method development work must still be done in-house for most LC-
MSMS assays. This includes defining, optimizing, and validating

83
J.A. Stone and J.G. van der Gugten Journal of Mass Spectrometry and Advances in the Clinical Lab 28 (2023) 82–90

commercially available or home brew reagents, calibrators, QC mate­ the scientist is not well trained in the technique or has insufficient time.
rials, extraction procedures, LC and MSMS parameters. Collaborations However, broader use of this experiment and related regulatory lan­
between individual laboratories and manufacturers can generate guage could lead to more robust QntLCMS-LDT methods.
methods published by the vendor as application or technical notes; Pre-analytical variables should not be ignored. Clinical laboratories
however, these publications are typically not ready to use as-is and use a wide variety of sample collection tubes. Depending on the analyte,
suggested parameters must be verified or re-developed for the specific tube type can impact analytical results. Sample processing and storage
instrument set-up on which the assay will be run. conditions should also be assessed. For example, samples for plasma
There are ongoing developments to provide automated LC-MSMS renin activity determination must be collected cold in EDTA tubes,
platforms with FDA-approved kit-based assays for clinical laboratories immediately centrifuged, and plasma frozen at once to prevent con­
(see Supplementary Material 1). While these may be welcome additions version of angiotensin I to angiotensin II. Laboratories often receive
to high-workload laboratories, they cannot replace all QntLCMS-LDT at already-processed samples from other locations, and procedures to
this time. confirm tube type may be necessary [33]. Interferences from gel sepa­
rator tubes are a known issue for QntLCMS-LDT steroid assays, and
Practical recommendations for robust QntLCMS method should be investigated and sample type acceptance criteria defined as
development and validation part of the method SOP [34]. Knowing the appropriate sample type, pre-
analytical stability, and sample storage requirements will inform which
The development of an LC-MSMS assay is a critical stage where samples are chosen for method comparison, matrix effects studies, sta­
sample preparation, chromatographic, and mass spectrometric param­ bility experiments, and blank matrix experiments.
eters are determined and optimized [25–30]. This is an empirical Pre-validation tests should be performed to estimate imprecision and
endeavor that requires time and patience. Perceived setbacks and fail­ assess accuracy during method development. Pre-validation flags should
ures during method development should be seen as learning opportu­ be monitored, and any issues should be addressed to improve the pro­
nities, with the ultimate goal of improving the assay. cess. Method validation is a major undertaking, and ignoring pre-
Trade-offs between laboratory resources and optimal method pa­ validation flags can lead to a failed validation. Materials and supplies
rameters may be necessary. Laboratory staffing and available equipment for method validation and routine patient analysis should be inventoried
may mean that some sample preparation procedures are not viable, and and ordered at the time of development and pre-validation, especially
other method aspects, such as liquid chromatography, may require more given the current supply chain challenges. If possible, it is beneficial to
attention in order to mitigate a lack of robustness that may be inherent test consumables (e.g., tubes, vials, 96-well plates, solvents, pipette tips)
in a simpler sample preparation. Laboratory limitations must be from multiple vendors (Fig. 1).
considered from the outset to avoid wasting time and resources pursuing The goal of method development is to create a robust method that
method parameters that are not feasible in the long run. The high can be successfully validated and used routinely with minimal problems.
complexity of LC-MSMS is challenging but delivers powerful analytical In this overview and the following table, we emphasize the stress points
value as a medical laboratory technique. Multiple options exist to and practical challenges of method development and validation, which
modify sample preparation, chromatography, mass spectrometry, and IT are discussed in more depth in existing QntLCMS method validation
features, which means there is almost always an alternative available to literature [25–30]. This process is highly empirical and usually involves
overcome method problems and add robustness. steps backward along the way. We recommend that method develop­
Assay development should include thorough investigation of matrix ment, at a minimum, include the following pre-validation tests to
effects. Matrix effects testing is required for QntLCMS-LDT by some thoroughly assess and stress-test the method before validation:
agencies with CMS deemed accreditation status (CAP) but not by
CLIA’88 [5,6]. A known risk in QntLCMS-LDT methods is sample- - Thorough matrix effects evaluation
specific inaccuracy beyond acceptable limits due to matrix effect - Imprecision estimates
[31,32]. Due to the nature of ionization in the MSMS source, sample - Accuracy assessment
matrix can suppress or enhance analyte ionization. Although it is not
possible to reduce matrix effects entirely, they should be sufficiently Personnel tactics for successful QntLCMS-LDT method validation
understood and mitigated to the extent that is feasible. Ideally, matrix
effects should be minimized and standardized between patient speci­ Although clinical LC-MSMS use has grown in prevalence over the
mens. Typically, this is achieved by sample preparation which varies in past 15–20 years [35,36], QntLCMS-LDT method development practice
its ability to remove sample matrix, and by the addition of stable
isotope-labeled internal standards (SIL). Appropriate use of SIL usually,
but not always, compensates for variable ionization due to matrix effect.
Thorough assessment of matrix effects should be performed during
method development so that mitigation measures (i.e., changes to
sample preparation procedures, SIL) can be made prior to embarking on
method validation.
Two types of matrix effect experiments are described in guidance
documents: quantitative, involving spike-and-recovery type experi­
ments, and qualitative post-column T-infusion experiments (see Sup­
plementary Material 2). Post-column infusion is less effort than the
quantitative matrix effect experiment, but it is not required by CLIA’88
or CAP regulations. It is a valuable tool during method development for
comparing sample preparation techniques and LC gradients to reduce
ion suppression risk. The qualitative matrix effect experiment should be
thought of as a method development tool (visualizing ion suppression)
rather than a method validation experiment with pass/fail criteria.
Deriving acceptance criteria for qualitative data (the timeframe of ma­
trix effect) is more difficult than for quantitative data (how much matrix Fig. 1. Successful QntLCMS-LDT Method Development is not a linear process!
effect). This experiment may be omitted during method development if [Image credit: This Is A Book, by Demetri Martin].

84
J.A. Stone and J.G. van der Gugten Journal of Mass Spectrometry and Advances in the Clinical Lab 28 (2023) 82–90

remains a specialized technique, not included in typical MLS training. Facilitate Hands-on method development Training:
MLS who perform routine patient testing usually rotate between
different laboratory ‘benches”, potentially with infrequent exposure to a. Medical laboratories with a track record of successful QntLCMS-LDT
QntLCMS-LDT. Laboratories may employ scientists who specialize in performance could offer intensive, short term, hands-on training for
assay QntLCMS-LDT development and validation. However, validation scientists and physicians who first complete in person or online a
performed by a single individual is not a true representation of daily didactic QntLCMS-LDT method development training. Mentorship
patient testing, as multiple MLS are likely to rotate through a QntLCMS- by a laboratory medicine scientist with QntLCMS-LDT method
LDT bench. As discussed elsewhere, sample preparation often involves development experience is key for training quality assurance.
multiple manual steps – a source of variance between days and between b. Scientists from medical laboratories with QntLCMS-LDT expertise
analysts. A robust validation of QntLCMS-LDT imprecision and bias could travel to provide short term, on-site, training and method
should include sample preparation by multiple MLS who will be per­ development support for start-up QntLCMS-LDT services.
forming patient testing once the method is approved by the laboratory c. Barriers to offering either type of training include funding, limited
director (Table 2). access to instruments and the time-consuming task of developing a
curriculum, training documents and the competency assessments
Infrastructure gaps affecting QntLCMS-LDT development and required by medical laboratories. Assembling and providing online
validation access to such materials would be a first step towards standardizing
and expanding hands-on medical laboratory QntLCMS-LDT method
Three topics are presented for consideration by the laboratory development training.
medicine community to promote more uniform and robust technical
validation of QntLCMS-LDT. A common thread is cooperativity between
QntLCMS-LDT providers to achieve these goals. In 2015 The Association
Exp and Availability and Training for Automation of QntLCMS-LDT
for Molecular Pathology published a proposal for modernizing CLIA’88
Workflows:
regulation of LDTs with a focus on molecular testing [16]. Recommen­
Once the hurdles of initial method development, validation, and
dations included pre-introduction external review of high- and
training of operational personnel have been overcome, a primary
moderate-risk LDTs and directing CMS to stipulate a minimum level of
obstacle to expanding medical laboratory QntLCMS-LDT services is the
standards for LDT analytical and clinical validity. We suggest similar
intensive manual labor necessary for sample preparation, instrument
concepts for QntLCMS-LDT.
operation, and data review of larger workloads. Solutions from multiple
1. Need for QntLCMS-LDT method development training and vendors exist for modular automation of liquid handling (ALH), media-
certification based sample preparation, automated multiplexing of LC mobile phases,
As of this publication, there are no specific degrees, training, expe­ LC column switching, and MSMS detection modes, rules-based data
rience, or licensure prerequisites for medical laboratory QntLCMS-LDT analysis, and auto-verification [47–53]. Although cost is a factor, the
method development scientists, other than federal and state re­ primary limitation is the knowledge gap and personnel shortages for
quirements that apply for generalist medical laboratory scientists (MLS) implementing and validating these options. Potential solutions include:
or laboratory directors. As stated previously, QntLCMS-LDT method
development practice is highly complex, but unlike routine LCMSMS a. Delineate and promote templates for LC-MSMS and related instru­
operation, it is rarely included in MLS training programs [29,45,46]. ment configurations that are most suitable for modular semi-
Laboratory inspectors and directors who are authorized to review or automated LDT at medical laboratories.
approve LDT validations are not necessarily trained in the technical b. Advocate for co-marketing by LC-MSMS, ALH, and middleware
details of QntLCMS-LDT validation. Learning how to perform QntLCMS- vendors of comprehensive order-to-result workflows for modular
LDT method development can be easier at non-medical versus medical semi-automated QntLCMS-LDT. Such resources could assist medical
laboratories, as more instruments and mentors are likely to be available laboratories with smart purchasing and implementation of their
and there are typically no medical laboratory licensure constraints at initial QntLCMS-LDT system to include suitable automation rather
pharmaceutical (pharma), contract research organizations (CROs), or than struggling to fund upgrades after workloads have increased.
clinical research laboratories. However, scientists who follow the non- c. Include the basics of ALH programming, best practices, and valida­
medical training route may be ineligible to develop methods at medi­ tion in QntLCMS-LDT online and hands-on training.
cal laboratories that require licensure for that function. d. Include the basics of instrument connectivity (interfacing) and data
transfer in QntLCMS-LDT online and hands-on training.
Define a knowledge base for QntLCMS-LDT method development:
e. Launch a collaboration between medical laboratory, IT and vendor
We advocate for the U.S. laboratory medicine community to define
stakeholders to standardize information transfer between LC-MSMS,
curriculum and competency criteria for personnel who perform
ALH, middleware and LIS platforms.
QntLCMS-LDT method development and manage validations. Inspectors
who accredit laboratories using this technique, as well as supervisors,
directors, and external reviewers who approve the method validations
and standard operating procedures (SOPs) should have a more limited, More Training – Then Certification:
task-specific characterization of competency. Didactic resources for this A voluntary examination-based certificate for QntLCMS-LDT method
specialty training are available (see Supplementary Material 4). We development scientists with hands-on training or experience is a logical
believe there is an additional, absolute requirement for scientists who next step. Any progress towards certification must include the existing
will perform QntLCMS-LDT method development to have hands-on cohort of qualified QntLCMS-LDT method development scientists in
training. Didactic training alone is insufficient [29]. The instruments medical laboratories. An additional option would be an accelerated path
used for this technique are more demanding to maintain in optimal for medical laboratory subspecialty training and certification of scien­
operating condition, and method development/validation is more tists who already have QntLCMS-LDT method development expertise
complex than for many other medical laboratory techniques [45]. from other applied markets (e.g. pharma, CRO, environmental and food
Vendor support and training for QntLCMS-LDT is available, but instru­ testing). Ideally, this voluntary QntLCMS-LDT method development
ment vendor personnel may not be authorized to perform method certificate would first benefit job applicants. Subsequent QntLCMS-LDT
development on behalf of medical laboratory clients [4]. specialty certification through ASCP and/or other entities, leading to
licensure, could follow. Consultants who provide fee-for-service

85
J.A. Stone and J.G. van der Gugten Journal of Mass Spectrometry and Advances in the Clinical Lab 28 (2023) 82–90

Table 2
Method Validation parameter challenges, problems and mitigation strategies.
Method Validation Shortcoming/Problem and why it might occur Implication for routine testing/Implication for patient results.
Recommendation

[Link] Method Validation Plan (MVP), Inadequate MVP. An MVP is a key document that should contain a summary of all tests to be performed along with acceptance
criteria for each test. This plan should be in place before method validation is started, and acceptance criteria
should not be adjusted based on the outcome of the validation experiments. Generic acceptance criteria
guidelines are found in numerous documents [4,25–30,37] but may need to be modified based on, for example,
biological variance or total allowable error [38].
Laboratories may have an existing generic MVP SOP. The lack of an analyte specific MVP could lead to
deficiencies in method validation experiments or inappropriate acceptance criteria. If a generic MVP SOP is used,
it should be sufficiently detailed to cover all required QntLCMS-LDT Method Validation experiments and include
minimum acceptance criteria.

Imprecision Studies
[Link] or excluding outliers in imprecision studies Under-estimation of the true imprecision of the assay may result from exclusion of outliers without justification.

[Link] day imprecision conducted over too short of interval Performing method validation experiments too quickly – for example over only 5 days rather than the minimum
recommended 20 days – can easily yield underestimates of the method’s imprecision (as well as bias and
quantitation limits).

[Link] studies performed by only one person Unless a laboratory has an automated liquid handler (ALH), sample preparation methods for QntLCMS-LDT
involve manual pipetting of patient samples, internal standards and reagents, and done by a single person will not
reflect the true imprecision of the assay when in the hands of multiple MLS. A robust validation of QntLCMS-LDT
imprecision and bias should include sample preparation by multiple MLS who will be performing patient testing
once the method is approved by the laboratory director.

Accuracy Assessment
[Link] external assessment of accuracy – lack of reference materials Accuracy is critically important and is established during method validation but must also be maintained for the
or proficiency testing samples. Reference materials or lifetime of the method. The superior selectivity of mass spectrometry is of little value if the results generated are
participation in harmonization programs can be costly – a inaccurate: this depends on calibrator accuracy. According to CLSI C62A, ‘Calibrator value assignments should be
significant hurdle for smaller hospital laboratories with limited traceable to a reference measurement procedure, whenever possible’ [26]. This is not always (or often) feasible –
budgets. typically due to cost or availability issues. The suggested hierarchy of accuracy assessment is:
1. Method comparison to a higher order reference measurement procedure (RMP) [39]
2. Analysis of materials with assigned values – such CRMS from JCTLM listed organizations ([Link]
[Link]/#/app/home), or proficiency testing (PT) materials with target values assigned using an RMP
3. Spike and recovery experiments
Spike and recovery as an accuracy assessment tool should be used as a last resort, and variables introduced in
order to mitigate bias. A second source of standard material (preferably with a C of A) should be used and a second
person should perform the spiking experiments.

[Link] few samples used for accuracy assessment Accuracy assessment should be done across the calibration range of the assay to adequately assess accuracy at all
levels. When this is not an option, accuracy should be assessed at the LLOQ and ULOQ and at clinically relevant
levels (see Supplementary Material 3).

Method Comparison
[Link] to the wrong method Method comparisons are performed in order to:
Assess accuracy
Assess bias to existing method and determine if changes to reference intervals are necessary.
Limitations of the comparator assay should be well understood. For example – IA suffers from poor accuracy for
low levels of many endogenous steroid hormones and therefore LC-MSMS vs. IA comparisons may not be valuable
for assessing accuracy or reference intervals. It is beneficial to compare any new LC-MSMS assay with an existing
LC-MSMS assay if possible. The clinical LC-MSMS community is often willing to share samples for method
comparison purposes. Calibrator materials should also be shared in order to rule out bias introduced by calibrator
source differences.

[Link] outliers in patient comparison studies Outliers in a patient comparison study could indicate an issue with improperly mitigated matrix effects, or
inadequately investigated interferences, and could indicate that the method should be further developed. A
prompt reanalysis of outliers with both the candidate and comparator methods, attempting to mitigate any
sample stability issues, is the optimal first step.

[Link] few samples used for method comparison study. This may Robustness of clinical laboratory tests is important. A sufficient number of comparison samples is required to give
occur because: confidence that the method will perform as expected for every sample tested. A minimum of forty patient samples
a) too few samples are available from the laboratory workload for between methods testing is recommended [40]. Testing too few samples may add risk of:
for esoteric QntLCMS-LDT missing significant positive interferences from interfering metabolites and/or drugs.
b) insufficient budget allocation for buying value assigned missing unacceptable negative ion suppression or enhancement (aka matrix effects) not compensated for by the
specimens from other laboratories SIL IS.
c) insufficient time and personnel allowed for archiving and Both issues can cause unacceptable inaccuracy.
analyzing patient samples for validation

(continued on next page)

86
J.A. Stone and J.G. van der Gugten Journal of Mass Spectrometry and Advances in the Clinical Lab 28 (2023) 82–90

Table 2 (continued )
Method Validation Shortcoming/Problem and why it might occur Implication for routine testing/Implication for patient results.
Recommendation

Matrix Effects
[Link] studies done during development Thorough matrix effects (ME) investigations should be done during method development, so that any method
This may occur because: modifications can occur prior to starting method validation. If matrix effects are not thoroughly mitigated,
a) time allotted for development is insufficientb) relying on reporting of inaccurate results is a risk. The calculations required for quantitative matrix effects [41] can be
previous test matrix effects results which are not LC-MSMS complex and easy to get wrong (as the authors have experienced), and these should be checked and tested during
based method development. Matrix Effects experiments done during method validation should be primarily for record/
(for example, relying on stated matrix effects by an documentation purposes.
immunoassay kit manufacturer for a specific analyte reagent
kit)

System suitability level and testing

[Link] System Suitability Test (SST) in use during validation, The system suitability test (SST) is typically of a mixture of pure analyte and internal standard in solution (i.e. not
wrong SST concentration (too high) or threshold (too low). extracted) which is injected and used to determine adequate LC and MSMS performance prior to starting an
analytical run containing patient sample extracts. If the SST does not meet acceptance criteria, troubleshooting
steps can be taken prior to running patient sample extracts.
A chief performance limitation of LC-MSMS systems is that detection capability (achieved signal to noise (S:N))
decreases unpredictably as more biological matrix specimens are introduced to the instrument. Residual matrix in
sample extracts will contaminate the MSMS hardware over time.

Fig. 2. Loss and recovery of analyte signal over time.

MSMS cleaning becomes necessary to restore acceptable S:N. Daily quality assurance measures (system suitability
testing or SST) should be in place to verify adequate S:N during validation as well as during subsequent patient
testing. However, such SST acceptance criteria are usually also in development during validation, and could be
less robust at detecting insufficient S:N. A healthy paranoia about SST tracking and decreasing S:N during
validation is useful.

Interference studies
[Link] of purchasing interference materials. The intended use of the assay will help inform what interferences to test. Known prescription medications used to
Insufficient time allotted/inappropriate scheme used for treat the illness or disease state for which patients are being tested should be assessed, as well as common OTC
interference testing. drugs that could potentially interfere with the analyte of interest. Isobaric and structurally similar compounds
What substances to test? should also be tested. For example, steroid hormones share a common structural backbone, and fragment
similarly in the mass spectrometer, and some compounds are indistinguishable by MSMS. Qualitative mixes of
OTC and common prescribed drugs are available from some standard reference material suppliers (ie Cerilliant).
It is impossible to test for every possible scenario. Taking into consideration the structure of the compound and
testing similar compounds that fall into the OTC and commonly prescribed category is a good place to start.
Routine clinical chemistry (ie IA) interference testing typically includes the evaluation of hemolysis, icterus and
lipemia effects on analytical results. These investigations should also be done for QntLCMS-LDT methods,
although they are often overlooked as they are not expected to have an impact. This may be true for many
analytes, but such effects may be both analyte and sample preparation dependent and should not be ignored.

[Link] state and potential related interferences There may be endogenous interferences in diseased patients that are not seen in healthy subjects. If possible,
samples from patients with syndromes known to cause interference or extreme values for the analyte(s) under
validation – such as renal failure, hepatic failure, autoimmune disease, or specific malignancies – should be
included in the patient comparison study.

Stability studies
[Link] time allocated for stability studies Stability should be tested at all storage times and conditions (i.e. temperatures) as such variance reflects what will
Wrong storage times/conditions tested happen to actual patient samples, including taking into consideration freeze–thaw cycles and re-analysis. If any
point of sample processing requires special treatment (i.e. thaw in fridge), this should be clearly stated in relevant
SOPs.
Stability studies should include pre-analytical steps to determine analyte stability in whole blood prior to
centrifugation as well as stability in serum or plasma.
If pre-analytical analyte stability is critical, strict instructions for sample collection and processing should be
communicated clearly to phlebotomy and included in relevant SOPs.

(continued on next page)

87
J.A. Stone and J.G. van der Gugten Journal of Mass Spectrometry and Advances in the Clinical Lab 28 (2023) 82–90

Table 2 (continued )
Method Validation Shortcoming/Problem and why it might occur Implication for routine testing/Implication for patient results.
Recommendation

[Link] or incorrect stability materials used Stability should be assessed for the analyte in solution, in all matrices including the blank matrix used for
calibration preparation, and in the final extract. Commercial, lyophilized QC materials should not be used for
stability studies in place of real patient samples. If possible, individual patient samples should be used for stability
testing, with patient sample pools as an alternative if sample volumes are insufficient.

[Link] made about reagent stability Most QntLCMS-LDT reagents are prepared in house, and stability can be easily overlooked. Critical reagents
(extraction solutions, mobile phases) should undergo stability testing, especially those prepared with known
volatile additives such as ammonia or acetic acid.

Reference interval studies

[Link] number of samples. The recommendation for clinical laboratory reference interval determinations is a minimum of 120 healthy
Not being able to perform reference interval study ‘reference’ subjects [42]. It is not always possible to recruit and collect this many samples. In many laboratories,
reference intervals are set based on previously published data and existing reference intervals. This may be
problematic if methodologies have changed, which is often the case with new QntLCMS-LDT assays [43].
Incomplete data could result in a reference interval that is inappropriately narrow or wide, with increased risk of
missing patients who need follow up or indicating a need for unnecessary treatment. There is no perfect, feasible
solution – these challenges are common to all reference interval studies. Using a well-defined reference interval
from a previous study, or using a method comparison study as a surrogate may be the best option available.

[Link] population for reference interval The intended use for the test must be considered when obtaining reference interval specimens. This may be
difficult, such as obtaining the samples necessary for testosterone reference intervals for pediatric tanner stages.
In this case, it may be reasonable to adopt the reference interval based on previous work [44].

QntLCMS-LDT method development and validation services should be Implementing modular automation solutions for QntLCMS-LDT re­
subject to the same certification as in-house personnel. In this time of a quires additional expense and expertise in ALH, LC, MSMS, and IT.
critical shortage of generalist MLS, the suggestion of additional, There are long-term financial returns to increased automation of
resource-intensive training is challenging; however, there will never be QntLCMS-LDT, but no regulatory incentives. This is true despite the
an ideal time to expand medical laboratory QntLCMS-LDT expertise. knowledge that complex, iterative manual laboratory processes, as listed
below, carry a greater risk of error compared to automation [47–51,54].
2. Variable Scope of Validation Experiments and Validation
Review.
a. Manual preparation and tracking of reagents (sample preparation
CLIA’88 regulations leave many of the details for method validation
materials, mobile phases, LC columns)
to the discretion of the laboratory director, such as the number and
b. Manual pipetting and container transfers during sample preparation
concentration of samples and duration of validation testing for preci­
c. Manual sample identification and placement for instrument loading
sion, accuracy, matrix effect and method comparisons [4,5]. Our ob­
d. Manual, instead of rules based, checks and tracking of LC-MSMS
servations as trainers indicate that QntLCMS-LDT validation practice
calibration and metadata during data analysis (e.g. ion ratios, peak
varies widely between laboratories. If the laboratory director or super­
areas, retention times)
visor reviewing validations does not have either QntLCMS-LDT method
e. Manual pre-analytical and post-analytical data transfer between in­
development experience or a designated expert, such variability is not
struments, middleware and LIS
surprising. Matrix effect experiments, as previously discussed, are not
f. Emphasis on minimizing reagent instead of labor costs
required by CLIA’88 rules. With such regulatory ambiguity, there is a
greater likelihood that some QntLCMS-LDT methods may have insuffi­
Regulatory incentives, such as reduced fees, leaner documentation,
cient validation and pose a greater risk to patient safety.
and accelerated validation review could promote enhanced automation
Enhance Validation Criteria: of QntLCMS-LDT methods. The latest version of the VALID Act con­
Rationale for more detailed criteria exists to promote site-neutral tained a technology certification process, which provided a pathway for
validation practice. The goal is to achieve uniform, high-quality vali­ permitting LDTs that fell within the same analytical scope of a previ­
dations across the entire spectrum of QntLCMS-LDT medical laboratory ously FDA-approved representative test to gain FDA-approved status
providers (academic and non-academic hospital laboratories, clinics, without going through FDA review for each subsequent test [11].
and reference laboratory variants). At a minimum, matrix effect testing
External Review of QntLCMS-LDT Method Validations:
and acceptance criteria, as outlined in CAP guidance, should be required
An understandable patient safety concern for all U.S. LDT testing is
for all QntLCMS-LDT used for patient care [6]. Establishing minimums
that external review, presumed to be more objective than internal re­
for the number of patient samples required for method comparison and
view, is not currently required for method validations. The latest version
accuracy studies, the minimum number of days over which precision,
of the VALID Act had provisions for qualification and use of external
accuracy, and patient sample comparison testing must be performed,
reviewers to meet the increased FDA workload expected if the legislation
and requiring clinically oriented justifications of acceptance criteria that
went into effect [11]. An alternative would be creating a non-FDA
deviate significantly from established norms could help prevent the
infrastructure for external approval of validations by reviewers from
most egregious examples of insufficient validation. To paraphrase Dr.
within the QntLCMS-LDT medical laboratory community.
Russell Grant, a leader in training for medical laboratory QntLCMS-LDT
Such an infrastructure could include an online validation document
method development and validation, less validation equals more risk, a
database for submission, review, approval, and archiving. The Associ­
higher likelihood of corrected results, and greater probability that
ation for Molecular Pathology (AMP) 2015 publication on enhanced
method revision and revalidation will be necessary.
CLIA LDT oversight suggests a similar resource, designed to provide LDT
Facilitate modular semi-automation of sample and data management: information for physicians and patients as well as regulators and LDT

88
J.A. Stone and J.G. van der Gugten Journal of Mass Spectrometry and Advances in the Clinical Lab 28 (2023) 82–90

providers [16]. Rather than a statistical software application, our References

concept is a virtual version of the validation binders found in every
medical laboratory. Instead of requiring a specific software for calcu­ [1] Framework for Regulatory Oversight of Laboratory Developed Tests (LDTs). Draft
Guidance for Industry, Food and Drug Administration Staff, and Clinical
lations or file format, the goal would be to simplify file uploads and Laboratories. Center for Devices and Radiological Health. Rockville, MD. October,
review of variable formats, easy entry of demographics, criteria verifi­ 2014. Docket # FDA-2011-D-0360; [Link]
cation, and secure access for laboratory inspectors and anonymized re­ search-fda-guidance-documents/framework-regulatory-oversight-laboratory-
developed-tests-ldts, Accessed 11/20/22.
viewers. As with other measures proposed here, we support initially [2] J.R. Genzen, J.S. Mohlman, J.L. Lynch, et al., Laboratory-Developed Tests: A
offering incentives for voluntary compliance and then phasing in a Legislative and Regulatory Review, Clin Chem 63 (2017) 1575–1584, [Link]
requirement for external validation approval by accrediting agencies. org/10.1373/clinchem.2017.275164.
[3] J.R. Genzen, Regulation of Laboratory-Developed Tests, A Clinical Laboratory
3. Scarcity of value assigned patient samples for comparison Perspective, Am J Clin Pathol 152 (2019) 122–131, [Link]
aqz096.
testing
[4] Center for Medicare and Medicaid Services (CMS). State Operations Manual
QntLCMS-LDTs are often used because no other methods have Appendix C - Survey Procedures and Interpretive Guidelines for Laboratories and
adequate selectivity, sensitivity, and turn-around time for the clinical Laboratory Services, [Link] Guidance/Guidance/
Manuals/Downloads/som107ap_c_lab.pdf, Accessed 11/20/22.
purpose of the test [55,56]. The workload for such novel or esoteric
[5] What is CMS’ authority regarding Laboratory Developed Tests (LDTs) and how
markers is usually small in a single healthcare center, meaning any does it differ from FDA’s authority?, CLIA FAQs, [Link]
single hospital, clinic, or smaller reference laboratory typically has regulations-and-guidance/legislation/clia/downloads/ldt-and-clia_faqs.pdf,
limited access to value-assigned patient samples suitable for QntLCMS- Accessed 11/20/22.
[6] College of American Pathologists (CAP). Accreditation Checklists, Chemistry and
LDT method development and validation. Using a small set of native Toxicology Checklist, 2021, Northfield, IL.
matrix samples with minimal variability to perform development and [7] McConagha W, Chan AH. FDA Oversight of Laboratory-Developed Tests Continues
validation can lead to a less reliable method. to Evolve. Skadden, Arps, Slate, Meagher & Flom LLP and Affiliates, [Link]
[Link]/insights/publications/2020/09/quarterly-insights/fda-oversight-of-
Creating a Virtual Sample Bank: laboratory-developed-tests, Accessed 11/20/22.
[8] Khatami E. The Role of Lab-Developed Tests in the In Vitro Diagnostics Market,
Informal sharing of de-identified, archived patient specimen rem­ Pew Charitable Trust, October 2021, [Link]
nants for validations between QntLCMS-LDT providers is common analysis/reports/2021/10/the-role-of-lab-developed-tests-in-the-in-vitro-
practice. Is it feasible to develop a virtual registry of de-identified, diagnostics-market, Accessed 11/20/22.
[9] Robinson SA, Carter AR, Brindley DA. The changing regulatory landscape of
QntLCMS-LDT-value-assigned patient samples that medical labora­ laboratory developed tests. Regulatory Focus, 30 August 2021, [Link]
tories are willing to store onsite and share? Such a collaborative data­ org/news-and-articles/news-articles/2021/8/the-changing-regulatory-landscape-
base, engaging QntLCMS-LDT providers from a range of laboratories, for-laboratory-d., Accessed 11/20/22.
[10] Boiani JA. The Valid Act: Senate Action Brings FDA Regulation of LDTs Closer to
could increase access to a larger pool of validation samples at lower cost
Fruition, National Law Review, 5/20/22, [Link]
and result in improved laboratory harmonization [57]. However, valid-act-senate-action-brings-fda-regulation-ldts-closer-to-fruition, Accessed 11/
compliance with HIPAA and IRB requirements for de-identification and 20/22.
quality-assured use of sample remnants, creative solutions to incentivize [11] Borfitz D. Current Perspectives On The Valid Act, Diagnostics World, August 30,
2022, [Link]
participation and fair use, and operational funding are significant ob­ perspectives-on-the-valid-act/, Accessed 11/20/22.
stacles. Aside from medical laboratories willing to donate specimen al­ [12] Volk EE. Contextualizing the Valid Act, CAP Today, September 2022, https://
iquots to create such a resource, a sponsoring organization, a governing [Link]/from-the-presidents-desk-0922/, Accessed 11/20/22.
[13] O’Connor L. AMP applauds removal of VALID Act from appropriations bill,
body, and funding for IT development would be necessary at a [Link], 12/27/22, [Link]
minimum. and-regulation/article/15304887/amp-applauds-removal-of-valid-act-from-
appropriations-bill, Accessed 2/17/23.
[14] US CongressionalRecord_2022_02. U.S. Senate Congressional Record, S.1666 -
Summary VITAL Act of 2021, [Link]
1666, Accessed 11/20/22.
A collaboration of experts from the laboratory medicine community [15] A. Ferreira-Gonzalez, R. Emmadi, S.P. Day, et al., Revisiting Oversight and
Regulation of Molecular-Based Laboratory-Developed Tests, A position statement
is the strongest resource for improving the reliability of QntLCMS-LDTs. of the Association for Molecular Pathology, J Mol Diagn 16 (2014) 3–6, https://
Novel tests with increased utilization may migrate from QntLCMS-LDTs [Link]/10.1016/[Link].2013.10.003.
to more highly automated platforms with greater throughput when an [16] Association for Molecular Pathology. Proposal for Modernization of CLIA
Regulations for Laboratory Developed Testing Procedures (LDPs). 2015. https://
automated, FDA-approved test becomes available [56]. Thus, QntLCMS-
[Link]/AMP/assets/File/advocacy/
LDT validations of the next esoteric assay in demand can be a perpetual AMPCLIAmodernizationproposalFINAL8_14_15.pdf?pass=93, Accessed 11/20/22.
task, providing additional justification for innovation and support of a [17] M. Vogeser, J.A. Stone, A suggested standard for validation of LC-MS/MS based
stronger validation infrastructure. analytical series in diagnostic laboratories, Clin Mass Spectrom 16 (2020) 25–32,
[Link] 2020.02.002.
These ideas may seem obvious, unnecessary, or unrealistic from the [18] D.A. Herold, R.L. Fitzgerald, Immunoassays for Testosterone in Women: Better
reader’s perspective. Our suggestions are not of the same scope or than a Guess? Clin Chem 49 (2003) 1250–1251, [Link]
impact, but such views were surely prevalent when the first CAP 49.8.1250.
[19] D. French, J. Drees, J.A. Stone, D.T. Holmes, J.G. van der Gugten, Comparison of
External Quality Assessment program began in 1947–48 [58] and four clinically validated testosterone LC-MS/MS assays: Harmonization is an
certainly when CLIA’88 regulations were implemented. We encourage attainable goal, Clin Mass Spec 11 (2019) 12–20, [Link]
the QntLCMS-LDT community to be innovative and forward-thinking clinms.2018.11.005.
[20] M. Rehan, J.E. Raizman, E. Cavalier, A.C. Don-Wauchope, D.T. Holmes, Laboratory
rather than defensive about enhanced regulation and validation challenges in primary aldosteronism screening and diagnosis, Clin Biochem 48
criteria. An effective defense against inappropriate external regulation is (2015) 377–1287, [Link]
a demonstration of comprehensive, self-imposed quality measures. [21] M.M. Kushnir, A.L. Rockwood, W.L. Roberts, et al., Measurement of Thyroglobulin
by Liquid Chromatography -Tandem Mass Spectrometry in Serum and Plamsa in
the Presence of Antithyroglobulin Autoantibodies, Clin Chem 59 (2013) 982–990,
Declaration of Competing Interest [Link]
[22] A. Mattman, L.Y.C. Chen, G. van der Gugten, A. Chin, M. Carruthers, M.
L. DeMarco, D.T. Holmes, Comment on: IgG4-related disease presenting with raised
The authors declare that they have no known competing financial
serum IgG2—real timeline of IgG4-RD? Rheumatology 57 (2018) 1125–1126,
interests or personal relationships that could have appeared to influence [Link]
the work reported in this paper. [23] A.R. Parker, K.L. Sharp, E.L. Jenner, M.D. Coley, A.D. Kay, S.J. Harding,
Remediation of IgG4 Cross-Reactivity in IgG1 and IgG2 Subclass Assays, Clin Chem
64 (2018) 1253–1255, [Link]

89
J.A. Stone and J.G. van der Gugten Journal of Mass Spectrometry and Advances in the Clinical Lab 28 (2023) 82–90

[24] K.L. Kaul, L.M. Sabatini, G.J. Tsongalis, et al., The Case for Laboratory Developed HPLC− MS/MS, Anal Chem 73 (2003) 3019–3030, [Link]
Procedures: Quality and Positive Impact on Patient Care. Academic, Pathology. ac020361s.
(2017) 4, [Link] [42] Horowitz GL et al. Clinical and Laboratory Standards Institute (CLSI), Wayne PA;
[25] Chace DH et al. Clinical Laboratory Standards Institute (CLSI), Wayne, PA, Mass Defining, Establishing, and Verifying Reference Intervals in the Clinical
Spectrometry in the Clinical Laboratory: General Principles and Guidance, 1st Laboratory; Approved Guideline, Third Edition, C28. 2008.
Edition, C50. 2007. [43] A. Katayev, C. Balciza, D.W. Seccombe, Establishing Reference Intervals for
[26] Clarke W et al. Clinical Laboratory Standards Institute (CLSI), Wayne, PA, Liquid Clinical Laboratory Test Results. Is There a Better Way? Am J Clin Pathol 133
Chromatography-Mass Spectrometry Methods, Approved Guideline, C620-A. 2014. (2010) 180–186, [Link]
[27] A. Kruve, R. Rebane, K. Kipper, et al., Tutorial review on validation of liquid [44] K. Adeli, V. Higgins, K. Trajcevski, N.-A. Habeeb, The Canadian laboratory
chromatography mass spectrometry methods: Part 1, Analytica Chimica Acta 870 initiative on pediatric reference intervals: A CALIPER white paper, Critical Reviews
(2015) 29–44, [Link] in Clinical Laboratory Sciences 54 (6) (2017) 358–413, [Link]
[28] A. Kruve, R. Rebane, K. Kipper, et al., Tutorial review on validation of liquid 10408363.2017.1379945.
chromatography–mass spectrometry methods: Part II, Analytica Chimica Acta 870 [45] B.A. Rappold, Review of the use of liquid chromatography-tandem mass
(2015) 8–28, [Link] spectrometry in Clinical Laboratories: Part II-Operations, Annals of Lab Med 42
[29] B.A. Rappold, Review of the use of liquid chromatography-tandem mass (2022) 531–557, [Link]
spectrometry in Clinical Laboratories: Part I-Development, Annals of Lab Med 42 [46] J.A. Stone, R.L. Fitzgerald, Liquid Chromatography– Mass Spectrometry Education
(2022) 121–140, [Link] for Clinical Laboratory Scientists, Clin Lab Med 38 (2018) 527–537, [Link]
[30] S. Moon, H. Choi, S. Kim, et al., Recommendations for Liquid Chromatography- org/10.1016/[Link].2018.04.002.
Mass Spectrometry in the Clinical Laboratory: Part II, Method Validation, Lab Med [47] E.L. Williams, Automation and integration of LC-MS/MS services into the clinical
Online 10 (2020) 95–108, [Link] laboratory workflow, Clin Lab Int. (2020.).
[31] Taylor PJ. Matrix effects: The Achilles heel of quantitative high-performance liquid [48] G.L. Salvagno, E. Danese, G. Lippi, Mass spectrometry and total laboratory
chromatography–electrospray–tandem mass spectrometry. Clin Biochem 2005; 38: automation: opportunities and drawbacks, Clin Chem Lab Med 58 (2020)
328-334. [Link] 994–1001, [Link]
[32] [Link]. Ion, suppression in mass spectrometry, Clin Chem. 49 (2003) 1041–1044, [49] M. Vogeser, F. Kirchhoff, Progress in automation of LC-MS in laboratory medicine,
[Link] Clin Biochem 44 (2022) 4–13, [Link]
[33] S. Haymond, N.J. Clarke, R.E. Reitz, M.J. McPhaul, Z. Wu, Plasma Renin Activity: clinbiochem.2010.06.005.
The Importance of Correct Sample Type, Clin Chem 62 (2016) 408–409, https:// [50] S. Haymond, System Performance Monitoring in Clinical Liquid Chromatography-
[Link]/10.1373/clinchem.2015.251009. Tandem Mass Spectrometry (LC-MS/MS), Methods Mol Biol 2546 (2022) 13–25,
[34] R.Z. Shi, H.H. van Rossum, R.A.R. Bowen, Serum testosterone quantitation by [Link]
liquid chromatography-tandem mass spectrometry: Interference from blood [51] F.B. Vincente, D.C. Lin, S. Haymond, Automation of chromatographic peak review
collection tubes, Clin Biochem 45 (2012) 1706–1709, [Link] and order to result data transfer in a clinical mass spectrometry laboratory, Clin
clinbiochem.2012.08.008. Chimica Acta 498 (2019) 84–89, [Link]
[35] S.K. Grebe, R.J. Singh, LC-MS/MS in the Clinical Laboratory - Where to From Here? [52] J. Stone, High-Throughput Serum 25-Hydroxy Vitamin D Testing with Automated
Clin Biochem Rev. 32 (2011) 5–31. Sample Preparation, Methods Mol Biol 1378 (2016) 301–320, [Link]
[36] C. Seger, L. Salzmann, After another decade: LC–MS/MS became routine in clinical 10.1007/978-1-4939-3182-8_32.
diagnostics, Clin Biochem 82 (2020) 2–11, [Link] [53] P. Janetto, Eliminating the Disconnect Between People and Processes, Automating
clinbiochem.2020.03.004. Sample Preparation for Mass Spectrometry, Clin Lab News (2015).
[37] U.S. Department of Health and Human Services Food and Drug Administration [54] M. Vogeser, Mass Spectrometry in the Clinical Laboratory—Challenges for Quality
Center for Drug Evaluation and Research (CDER) Center for Veterinary Medicine Assurance, Spectroscopy (2015).
(CVM) Bioanalytical Method Validation Guidance for Industry May 2018. [55] P.J. Jannetto, R.L. Fitzgerald, Effective Use of Mass Spectrometry in the Clinical
[38] D. Hoffman, R. Kringle, A Total Error Approach for the Validation of Quantitative Laboratory, ClinChem 62 (2016) 92–98, [Link]
Analytical Methods, Pharm Res 24 (2007) 1157–1164, [Link] clinchem.2015.248146.
s11095-007-9242-3. [56] J.A. Stone, The Way Forward for Clinical Mass Spectrometry in Hospital
[39] Salt MI et al. Clinical and Laboratory Standards Institute (CLSI), Wayne PA; Laboratories, Clin Lab News (2019).
Metrological Traceability and Its Implementation. 1st Edition. EP32. 2006. [57] G. Miller, The standardization journey and the path ahead, J Lab Precis Med 3
[40] J.R. Budd, et al., Clinical and Laboratory Standards Institute (CLSI), in: Wayne PA; (2018) 87, [Link]
Measurement Procedure Comparison and Bias Estimation Using Patient Samples, [58] F.W. Sunderman Sr., The History of Proficiency Testing/Quality Control, Clin
Approved Guideline, Third Edition, 2013, p. EP9-A3. Chem 36 (1992) 1205–1209.
[41] B.K. Matuszewski, M.L. Constanzer, C.M. Chavez-Eng, Strategies for the
Assessment of Matrix Effect in Quantitative Bioanalytical Methods Based on

90