Standardization of Negative Controls in Diagnostic Immunohistochemistry: Recommendations From the International Ad Hoc Expert Panel

Abstract

Standardization of controls, both positive and negative controls, is needed for diagnostic immunohistochemistry (dIHC). The use of IHC-negative controls, irrespective of type, although well established, is not standardized. As such, the relevance and applicability of negative controls continues to challenge both pathologists and laboratory budgets. Despite the clear theoretical notion that appropriate controls serve to demonstrate the sensitivity and specificity of the dIHC test, it remains unclear which types of positive and negative controls are applicable and/or useful in day-to-day clinical practice. There is a perceived need to provide “best practice recommendations” for the use of negative controls. This perception is driven not only by logistics and cost issues, but also by increased pressure for accurate IHC testing, especially when IHC is performed for predictive markers, the number of which is rising as personalized medicine continues to develop. Herein, an international ad hoc expert panel reviews classification of negative controls relevant to clinical practice, proposes standard terminology for negative controls, considers the total evidence of IHC specificity that is available to pathologists, and develops a set of recommendations for the use of negative controls in dIHC based on “fit-for-use” principles.

BACKGROUND

Great efforts have been made to standardize diagnostic immunohistochemistry (dIHC), especially with the establishment of external quality control assessment (EQA) services for immunohistochemistry. The UK National External Scheme for Immunocytochemistry and in situ Hybridization (UK NEQAS ICC & ISH) has been providing proficiency testing and education in dIHC for 25 years, and several other programs have subsequently developed [eg, Nordic immunohistochemical Quality Control (NordiQC), College of American Pathologists-External Quality Assurance/Proficiency Testing (CAP-EQA/PT), Canadian Immunohistochemistry Quality Control (CIQC), Royal College of Pathologists of Australasia Quality Assurance Program (RCPAQAP), and others]. However, it is only in the last decade that technical advances, including improved quality of antibodies (Abs), automated staining platforms, and highly sensitive detection systems have resulted in more reproducible staining methods, which have had a direct impact on interpretation of IHC results and patient care.^1–12 These standardization efforts have originated from practicing pathologists, national and international professional organizations, and industry. Although standardization of dIHC includes preanalytical, analytical, and post-analytical phases, this paper concentrates on the standardization of controls, specifically negative controls, rather than on standardization of IHC protocols overall. It is important to emphasize that the standardization of dIHC must include standardization of controls, which are essential elements of quality control (QC) (analytical phase), as well as interpretation of IHC results (post-analytical phase). The correct interpretation of dIHC is reliant on observation of the expected staining results achieved on known positive and negative controls. Daily positive and negative controls (as defined below) are an important link between analytical and postanalytical components of IHC testing.¹³ The focus of this paper is standardization of negative controls for dIHC.

DIAGNOSTIC IHC CONTROLS

Positive and negative controls serve the purpose of providing evidence that each IHC test (stain) is successfully performed and is giving the expected level of sensitivity and specificity as characterized during technical optimization and validation of the IHC test for diagnostic use. However, both positive and negative controls currently lack agreed upon criteria for standardization. In addition, the principles and approaches to the standardization of positive and negative controls are vastly different, as follows:

1. Positive tissue controls (PTCs) primarily monitor calibration of the system and protocol sensitivity, but also provide assurance that the right Ab is applied. Briefly, external positive tissue control (Ext-PTC) should be selected so as to have undergone fixation and processing in a manner as closely similar as possible to the test tissue. Antigens internal or intrinsic to the test tissue (patient sample) may also be used for this purpose, if present. This type of positive control is termed here as internal positive tissue control (Int-PTC). The optimal selection of Ext-PTC and interpretation of the results of the Ext-PTC and Int-PTC requires in-depth experience and knowledge of the biological processes, in the context of the intended use of each IHC test, to establish desirable and relevant calibration of the protocol. For example, Ext-PTC for CD20 should include tissue (preferably normal tissue with predictable levels of CD20 expression), representative of both low-expression and high-expression levels to demonstrate that the overall sensitivity of the protocol is sufficient to detect CD20 expression in lesions with low CD20 expression (such as small lymphocytic lymphoma or some diffuse large B-cell lymphomas), if the intended use of the CD20 IHC test is to be able to reproducibly detect CD20 expression in such lesions. A future issue of Applied Immunohistochemistry & Molecular Morphology will focus on selection and use of positive controls.

2. Negative controls are primarily used to evaluate the specificity of the IHC test to identify false-positive staining reactions. They include negative reagent controls (NRCs) and negative tissue controls (NTCs). Basic principles that govern the use of negative controls, including selection of the type of negative control, the frequency of use, and validity, tend to apply across all of diagnostic pathology, and are generally not subspecialty driven. Therefore, the use of negative controls can be standardized at this time, to be more relevant to current needs of pathology practice.

There is a perceived need to provide “best practice recommendations” for the use of negative controls. The perception of this need is driven not only by logistics and cost issues, but also by increased pressure for accurate IHC testing, especially in the context of predictive markers, which will continue to rise in number as personalized medicine continues to develop. The need for thorough discussion and standardization of negative controls was also highlighted by recently published recommendations regarding negative control use from the College of American Pathologists (CAP).¹⁴

EVIDENCE-BASED USE OF NEGATIVE CONTROLS IN DIAGNOSTIC IHC

Currently, there is no published evidence to support the diverse requirements emanating from different accreditation bodies throughout the world with respect to the selection and use of negative controls in dIHC. False-positive reactions may occur due to characteristics of the primary Ab or “nonspecific” (unwanted) binding of the components of the detection system, which may include a secondary Ab that lacks specificity. Although several countries have published guidelines and recommendations relating to quality assurance in IHC, there is little published evidence to support the use of particular types of negative controls in dIHC. Much of the published work relates primarily to biotin-based systems.^15–22

The recommendations in this paper are based on: (i) published literature; (ii) evidence generated in various EQA programs; and (iii) current expert opinion.

COST-EFFECTIVENESS OF NEGATIVE CONTROLS IN dIHC

Globally, there is great practice variation in the use of negative controls. Laboratory accreditation requirements may on occasion be the main driver for the use of certain negative controls. Additional slides serving as NRCs varied from 0 in some laboratories to ≥15% of the total IHC slide annual workload in other laboratories.²³ Even in countries with a long history of laboratory accreditation and EQA, there remains a non-standardized approach to the use of NRCs; UK NEQAS reported that 48% of dIHC laboratories used NRCs, whereas specifically in neuropathology is was even less at 36% (E.E.T., unpublished data, 2012).

The importance of negative controls has been emphasized for IHC protocols that use avidin-biotin detection systems, especially in combination with powerful antigen retrieval methods known to “unmask” endogenous biotin,^24,25 a procedure that exacerbates the undesirable binding of avidin-based detection systems. Although many laboratories no longer use avidin-biotin– based detection systems, the use of negative controls designed to detect such spurious binding continues unabated. This practice has resulted in the unnecessary processing of very large numbers of negative control slides, even for IHC protocols that utilize “polymer/multimer” detection systems with much less susceptibility for false-positive results.

One direct consequence of this “routine unthinking” approach is that pathologists often ignore the additional NRC slides, a practice justified on the basis that evaluation of a panel of IHC Abs, in the hands of an expert interpreter, provides substantial and sufficient evidence that the achieved reactions are specific. Therefore, it may be very cost-inefficient to comply with accreditation requirements that mandate one or more NRC slides on every case.

In contrast, many laboratories around the world may not run NRCs at all, either because these controls are not mandated, or their implementation would increase the cost of operations, or both.

In this context, there are expert opinions that NRCs now are largely or completely unnecessary. Recently, this approach was also partly accepted by the CAP in the United States for dIHC tests utilizing polymer-based/multimer-based detection systems. The CAP anatomic pathology checklist item concerning the use of negative controls (ANP.22570) was recently revised with the use of NRCs in dIHC now being at the “discretion of the laboratory director,” so long as the avidin-biotin detection systems are not used.¹⁴ However, the CAP checklist still requires NTCs (see Table 1 for definitions). The revised requirements from the CAP are highly appreciated because extensive experience with the use of polymer-based/multimer-based systems has shown that NRCs are not very useful, in particular if a panel of markers are used with the same detection system.

TABLE 1.

Terminology of Negative Controls in Diagnostic Immunohistochemistry

Type	Negative Control Preparation	Purpose	Application	Examples	Interpretation
Negative reagent control (NRC)
(a) NRC-primAb—replace primary Ab with “nonspecific” Ig	Application of nonimmune Ig; same isotype, incubation, concentration as primary monoclonal Ab. OR A dilution of Ig fractions nonimmune serum of same species as primary polyclonal Ab	Identification of false-positive staining reaction due to binding of primary Ab species. Important for initial validation and where performance characteristics of the primary Ab are unknown. Will also detect nonspecific staining due to detection system or other protocol steps (see NRC “b”)	Serial section of the patient block is incubated with nonimmune Ig (negative control slide), using same protocol as test section	For CD3, mAb clone PS1, Ig subtype IgG2a is used at a titre of 1:100 of a stock solution of 100 mg/mL. Nonimmune serum subtype IgG2a is applied on a separate slide in identical concentration and incubation time	Desired result is complete absence of staining. Positive reaction is due to nonimmune binding of Ab isotype, or other steps of detection system or protocol. Investigate by supplementary NRC. False positivity due to primary Ab contamination or other unexpected reactivity of the primary Ab will not be revealed
(b) NRC-detSys (supplementary negative controls)—omit primary Ab	Omit each reagent step of protocol in turn. First and most often primary Ab is omitted. Detection system components may be omitted subsequently	Identification of false-positive staining related to nonspecific reaction of the detection system	Apply to serial section(s) of patient block(s). Use when unexpected (unwanted) staining is seen in the NRC antibody negative control	Replace primary Ab with diluent instead of nonspecific Ig. Replace biotin-avidin detection system with polymer-based reagents. Replace HRP-DABchromogen, with alkaline phosphatase detection	Detects nonimmune binding of Ab isotype. Detects unwanted biotin-avidin binding Detects endogenous peroxidase
Negative tissue control (NTC)
(a) Int-NTC— using tissue elements in test section	Patient test slide(s)	Identification of false-positive staining reaction of structures known not to express the target antigen. False-positive staining related to primary Ab and/or detection system and protocol can be identified	The patient test slide is evaluated after application of the primary Ab, using the standard protocol	Antibody for CD117 is applied to the patient case and evaluated for “expected” and “unexpected” staining in internal nonlesional structures	Any aberrant positive staining is evaluated against other antibodies if a panel is employed, or against the NRC antibody control. Neither “internal” nor “external” control tissues can be guaranteed to contain all tissue elements that could potentially give unexpected staining
(b) Ext-NTC	Tissue stained in addition to the patient test tissue. May be placed on the separate slide, or “on-slide.” Expected negative tissue is included with expected positive		For optimal evaluation, the control tissue(s) is mounted on same slide as the patient material (on-slide), ensuring same exposure to all steps of IHC protocol	Antibody for CD117 is applied to the control tissue known to not to express CD117, and subjected to an identical IHC stain protocol, including retrieval	Relative expression level in test and control tissue may be unknown

IHC indicates immunohistochemistry.

THE EXTENT OF THE PROBLEM

The use of IHC-negative controls, irrespective of type, although well established, is not standardized and the relevance and applicability of negative controls continues to challenge both pathologists and laboratory medicine budgets. Despite the clear theoretical notion that appropriate controls serve to demonstrate the sensitivity and specificity of the dIHC test, it remains unclear for most dIHC tests that which types of positive and negative controls are applicable, and/or useful in day-to-day clinical practice.

This paper addresses the different types of negative controls and their recommended use in diagnostic pathology. There is a wide-spread tendency to use the general term “negative control” for all types of negative controls, including the “NRC,” which is a separate slide on which a parallel section of the patient’s test tissue sample is run omitting the specific primary Ab (Table 1). This lack of clear definition is misleading, because other types of negative controls exist, having different import.

It is critical to recognize that NRCs are not the only source of evidence of the specificity of the IHC test (Table 1) and that NRCs have limited potential to detect false-positive results due to unexpected cross-reactivity of the primary Ab. EQA experience has shown that false-positive results may unexpectedly appear with various Abs, even those that are well established and widely validated in the literature. Recent examples from CIQC experience include Ab clones 6F11 for estrogen receptor (ER) and SP4 for cyclin D1, as well as UK NEQAS ICC & ISH and NordiQC experience with the Ab products of clones SP2 and IE2 for progesterone receptor (PR).^26,27 False-positive results were observed that were not detected by NRC slides alone. A combination of experience, knowledge of the biodistribution of the target antigen, and employment of other types of negative controls is required in these instances; in particular this includes evaluation of the NTCs (Table 1). Where appropriate, absorption of the primary Ab with the target antigen (especially valuable for “polyclonal” antisera as opposed to monoclonal Abs) and the use of cell lines known not to express the antigen in question, by ELISA or mRNA comparative studies could be employed. These latter controls are beyond the scope of the “routine clinical laboratory” and usually reside within the domain of the Ab manufacturer as a part of characterization and validation; such data should be available to the user in package inserts or specification sheets.

This paper examines previously published recommendations for negative controls in dIHC and introduces the concept of “IHC Specificity Evidence” (IHC-SE), to reconcile theories around the use of negative controls with current diverse practices and observed deviation from recommendations. IHC-SE includes all evidence available to pathologists, by which they can determine the likelihood of whether IHC test results are specific, with respect to cell distribution and diagnostic utility.

STANDARD TERMINOLOGY AND DEFINITIONS

Definitions of the types of controls for dIHC are summarized in Table 1.

Positive Controls

The purpose of the positive control is to demonstrate that the IHC protocol is able to detect the antigen of interest (qualitative positive control) and is sensitive at the level of clinical relevance (calibrated positive controls; low and high expressers). Positive controls include samples that have been previously tested for the Ab in question, or samples which are known to express the Ab at levels within the desired range for the intended use following sample preparation (including fixation and retrieval).

Negative Controls

The purpose of the negative control is to demonstrate whether the IHC test is specific. Usually, “negative control” is a term applied to a separate slide with patient’s test sample on which no primary Ab is applied (NRC). However, rather than using results of the negative control(s) slides in isolation, pathologists evaluate the specificity of dIHC by using various approaches, including “internal controls” (Fig. 1), as well as other Abs applied in “panels.”

FIGURE 1.

“On-slide” external and internal negative tissue controls are illustrated. It is sound practice whenever possible to include cells (or tissue elements) that will serve as negative controls (expected to be nonreactive) when selecting tissue for the positive tissue control. Both internal and external negative on-slide tissues are so-called “specific” negative controls because all are exposed to the specific primary antibody. Separate slide negative controls are generally used for negative reagent controls, where the primary antibody is omitted or an irrelevant primary antibody is used. Note that reagent controls should have identical protocols to the specific immunohistochemistry test, including the same type of pretreatment, as far as is possible.

Even when using an optimized and validated primary Ab (polyclonal or monoclonal), it is still possible that nonspecific labeling, other than of the target epitope, may be found. In this context, “nonspecific” is defined as staining of normal or abnormal cells or tissues that are not part of the described intended use for the particular IHC test. This outcome may be due to contamination of the commercial product, including unpurified ascites mouse monoclonal Abs causing the so-called mouse as-cites Golgi reaction in human exocrine cells from individuals with blood type A, but may also be a consequence of low affinity binding to secondary target epitopes, or to cross-reactive (molecular) species that share the primary epitope.^28,29 These effects may be reduced (although not necessarily eliminated) through optimization/calibration of the entire protocol before clinical use. There are also newly developed tools to test the monospecificity of primary Abs, such as high-density protein microarray chips developed for Ab-specificity testing; for example, by using a chip spotted with >10,000 unique overexpressed proteins, it is possible to validate the specificity of an existing diagnostic Ab.³⁰ Also, it must be remembered that when elements of fixation/processing also are considered (antigen loss, masking, alteration, retrieval), then existing negative (and positive) controls may fall short,^30–35 as discussed below.

Immunohistochemistry-Specificity Evidence

IHC-SE consists of all components of IHC testing that provide evidence relating to the specificity of IHC tests. Evidence may be derived from on-slide controls (both internal and external), separate slide NRCs, evidence of specificity derived from an IHC “panel” used on the same tissue block with the same detection system, informed interpretation of differential intensity of staining in the internal and external positive controls, as well as subcellular signal localization/pattern, and informed interpretation based on understanding of biology (Fig. 1). This all-encompassing approach empowers pathologists to apply their expertise in interpretive science and allows for efficient and cost-effective practice.

Internal Tissue Controls

Internal tissue controls, also referred to as intrinsic or built-in controls, utilize the tissue elements in the patient’s test sample that is being evaluated for expression (internal tissue positive control) or absence [internal negative tissue control (Int-NTC)] of certain epitopes. Almost all tests in dIHC are selective and have reactivity with epitopes that are not universally present in all cells or all human tissue; all the elements in the patient’s test sample that are known to express the epitope of interest are designated as “internal positive controls” and all patient’s test tissue elements that are known to not express the epitope are designated as “internal negative controls.” For example, human colonic mucosa shows expression of CK20 in the normal epithelium (internal positive control for CK20), but all other tissues/cells in the colon should be negative (internal negative control for CK20) (Fig. 1).

Evaluation of dIHC results in the “tests section(s)” mandates careful assessment of results with internal controls. If unexpected patterns of positive staining are observed, then careful reexamination of all control materials is essential to look for evidence of features that may produce spurious staining, such as too high a concentration of Ab, prolonged incubation times, drying of sections, or inappropriate overly vigorous antigen retrieval.

It is wrongly assumed that internal controls are always better than external controls and that when internal control tissue is present, no external control is needed. Not every tested antigen has a normal internal counterpart that can be used as an internal control, and not every section will contain the tissue element in question and internal positive controls rarely can be used to provide information if the appropriate sensitivity of the protocol was achieved. Also, with newly developed Abs against new or recently discovered molecules (epitopes), it is often the case that the patterns of reactivity in normal and abnormal tissues are unknown, thus precluding appropriate internal control assessment.

Internal controls are also particularly valuable in the evaluation of sample preparation. Unexpected staining of negative control elements raises the possibility of nonspecific binding in areas of necrosis or overfixed/underfixed tissues. When the internal (positive) control produces a negative result and the tissue/tumor is also negative, a false-negative result (due to overfixation or technical failure) cannot be excluded and the test cannot be interpreted. An example of the value of an internal positive control is the observation of ER and PR receptors in normal breast epithelium to prevent false-negative interpretations in breast cancer, as recommended in American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) guidelines.³⁶ For other applications [eg, CK5 staining of basal cells in prostate biopsies, or mismatch repair (MMR) proteins MLH-1, MSH-2, PMS-2, and MSH-6 in colon], the internal positive and negative controls are generally universally present, and sufficient, and superior to external controls, as they also allow for assessment of factors related to tissue fixation and processing.

At the present time, internal controls provide only limited quantitative information regarding different levels of antigen expression, which may be important for determining whether a test is properly calibrated or not. However, proposals have been made for the development of a system of Quantitative Internal Reference Standards³⁷ that would go some way in providing a basis for calibration, whereas at the same time serving to “qualify” tissue as having undergone all steps of sample preparation in a manner permissible for satisfactory IHC staining.

External Tissue Controls

External controls designate the use of tissues, or cell lines, derived from sources other than the actual patient’s test sample (therefore they are termed “external”). It is important to note that because the tissue (or cell line) is not a part of the patient’s test sample used for IHC testing, preanalytical processing, despite best efforts, will not be identical (eg, different ischemic time, different fixation time). External controls may be tissue controls (single tissue sections or tissue microarrays), cell blocks prepared from cell lines (or body fluids such as pleural or peritoneal fluids), any combination of these, or even peptides deposited onto glass slides.^37–43 The tissues may originate from the diagnostic tissue archives or may be grown as xenografts, or faux tissues (artificial control tissue) such as “histoids” constructed from several cell lines.^44–47 The latter may be superior when antigens are not expressed in normal tissues, such as ALK-1, an important marker in anaplastic large-cell lymphomas or certain lung cancers.

External tissue controls are designated as “external positive tissue controls” (Ext-PTCs) if they express an epitope of interest and are designated as “external negative tissue controls” (Ext-NTCs) (Table 1) if they contain cells known not to express an epitope of interest. External controls may be run on the same slide as patient’s test sample (so-called “on-slide” controls) or on a separate slide. “On-slide” controls require the mounting of 2 cut paraffin sections on the same slide (the control section and the test section), which may be technically more demanding, but offers greater assurance of identical retrieval and staining conditions being applied for the external control and the patient’s test sample. External controls on separate slides may also be used as “batch” or “run controls,” when a large number of slides are stained with the same primary Ab/protocol. It is worth emphasizing that several EQA programs recommend that external controls should contain both expected positive and expected negative tissue elements (Fig. 1).^48,49

Specific and Nonspecific Negative Controls

In 2000, Taylor introduced a classification of negative controls as “specific” and “nonspecific.”¹⁰ “Specific” negative controls are tissue-based controls designed to test whether the reactivity with the tissue/staining is due to primary Ab-specific design for specific epitope so that the reaction will not be present if the specific primary Ab is replaced by nonimmune serum of the same immunoglobulin (Ig) subtype. “Nonspecific” negative controls are designed to test whether any part of the protocol or detection system (beyond the primary Ab) gives unexpected staining.

NRC (Negative Reagent Controls)

NRCs are used to confirm specific binding by the primary Ab and detection system. NRCs are always run on a separate slide using serial sections from the same paraffin block used for IHC testing of the patient’s test sample. It is important that the same antigen retrieval method be used for both, NRC slides and the patient’s test slides, because antigen retrieval may have profound effects on availability of epitopes that are detectable in IHC reactions; variation in antigen retrieval techniques may result in spurious staining.

There are 2 types of NRCs:

1. NRC for primary Ab (NRC-primAb) (Table 1):

   1. For monoclonal Abs, the primary Ab is replaced by a monoclonal Ab of the same species, isotype, and concentration, having no known reactivity against human tissues.
   2. For polyclonal Abs, the primary Ab is replaced by a dilution of whole normal/nonimmune serum of the same animal source, at the same Ig concentration.
2. NRC for detection system (NRC-detSys) (Table 1):

   1. The secondary or labeling Ab (or other component) is replaced by diluent or culture medium (McCoy tissue culture medium is commonly used), or buffer.

NTC (Negative Tissue Controls)

Specific negative controls are tissues that are known to not contain the antigen of interest and are evaluated using the specific primary Ab. More broadly, specific negative controls may also be created from cell lines, in which case they are most appropriately named negative cell line control.

There are 2 types of NTCs (depending on the source):

1. Int-NTC, if the control element is a portion of the patient’s test sample (internal, intrinsic or built-in negative control) (Table 1), or

2. Ext-NTC, if the control element is part of the external (usually on-slide) control, which contains both positive and negative cell types or tissues [“external on-slide negative control” (see above, Fig. 1 and Table 1)].

INTERPRETATION OF NEGATIVE CONTROL RESULTS

Pathologists should interpret the results of all positive and negative controls before examining the test/patient’s samples. Technologists also should interpret the results of NRCs and on-slide external (positive and negative) controls (Fig. 1). The results should be recorded as part of the routine laboratory QC.

If the NRC-primAb is positive, then this is a false-positive reaction, which may be due to characteristics of the primary Ab isotype or class (eg, IgG1) that result in nonspecific attachment of “nonimmune” Ig to the section at the specific concentration and incubation time employed (see below). Whether it is the primary Ab isotype or class or some components of detection system that are causing the false-positive staining may be further explored by using various types of NRC-detSys (see below and Tables 1 and 2).

TABLE 2.

Recommendations for Use of Negative Controls in Diagnostic Immunohistochemistry

Type of Control	CAP-ACP Clinical Use IHC Test Class I	CAP-ACP Clinical Use Class II Tests		Comments
	FDA IHC Device Class I	FDA IHC Device Class II	FDA IHC Device Class III
Negative reagent control (NRC)
NRC-primAb—replace primary Ab with “nonspecific” Ig	Recommended for initial antibody validation, and for use with avidin-biotin detection Not recommended for routine daily use of validated protocol using polymer-based detection Can be ordered by pathologist in specific situations (see text)	Recommended as per published guidelines When no guidelines exist, the NRC antibody control is recommended where results may dictate definitive treatment (ie, ER, PR), and are not confirmed by other aspects of pathology testing	Use negative reagent controls as per approved guidelines	When panels of several antibodies are used on serial sections, negative staining elements in the different sections serve as a negative reagent controls, obviating the need for a separate negative reagent control in most instances of class I testing Also, pathologists’ interpretation of IHC-SE determines if NRC-primAb is required May require multiple controls if several different retrieval methods are in use
–NRC-detSys (supplementary negative controls)		Use where unexpected staining is observed in the NRC antibody negative control slide (Table 1)		May require multiple controls for different components of detection system and if different retrieval methods are in use
Negative tissue control (NTC)
Internal NTC— evaluate tissue elements that should be negative in test section of the patient’s sample	Recommended	Recommended	Use negative and positive controls tissue as per approved guidelines	If test section does not include elements that serve as negative controls, then, external tissue control may be informative
External NTC— evaluate tissue elements in control tissue that should be negative	Recommended	Recommended		Control tissues may be derived from archived diagnostic tissue as single sections, or tissue microarrays. Cell lines prepared as cell blocks, if processed in the same way as patient samples can be also be used (see text)

If the NRC-detSys is positive, then the false-positive reaction is due to nonspecific binding by the secondary Ab or any other component of the detection system. If the detection system is based on avidin-biotin chemistry, then false-positive reactions are frequently caused by specific binding of avidin to endogenous biotin (Tables 1 and 2).

In practice it is usual only to run the NRC-primAb, reserving the performance of NRC-detSys for cases in which unexpected staining is observed in the NRC-primAb slide and is not explicable from experience and known mechanisms.

If the NRC is negative and the NTC is positive, the false-positive result is likely due to variables associated with the specific primary Ab, requiring further evaluation of the specificity and reaction patterns of the primary Ab. In the first place, a possibility that the wrong Ab was applied needs to be ruled out.

Occasionally, polyclonal Abs (also known as antisera) may be contaminated with other Abs due to impure antigen used to immunize the host animal, or due to high levels of Igs resulting from exposure of the animal to microbial or other antigens before its use for Ab generation. Similarly, new lots of the primary polyclonal Ab can show unexpected cross-reactivity. Unexpected cross-reactivity may be difficult to detect during validation of Ab by the manufacturer. Such false-positive reactions may or may not be detected by use of specific negative controls, as the tissues included in the patient sample or external control tissue(s) may or may not contain all of the epitopes that are relevant for proper interpretation of the results achieved with lesional tissue (as may be the case with the mouse ascites Golgi reaction).

Thorough understanding of underlying biology and the use of total IHC-SE approach is of value in determining whether the staining is correct or not.

A major cause of nonspecific “background” staining is nonimmunologic binding of the primary Ab reagent by hydrophobic, ionic, and electrostatic forces to certain sites within tissue sections. This form of background staining is usually uniform and is relatively common when very high primary Ab concentration is used. As already noted, another important culprit for nonspecific staining is the binding of avidin to endogenous biotin when avi-din-biotin–based detection systems are used; this type of binding is in fact chemically specific, but nonimmune based. Examples of false-positive IHC results due to endogenous biotin are illustrated in Figure 2. There may be also background or nonspecific staining when using polymer-based/multimer-based detection systems, especially with linking/amplification steps to increase sensitivity of the detection system. These linking Abs, which are usually polyclonal Abs, may bind nonspecifically to tissue components by cross-reactivity or nonimmunologic forces. In such cases addition of a casein block before the detection system may reduce this effect.

FIGURE 2.

Most frequently endogenous biotin is visualized as granular cytoplasmic reactivity; endogenous biotin in adrenal tissue (A) and kidney (B). On occasion, it also may be seen as nuclear reactivity shown here in endometrium (C). When strongly expressed, endogenous biotin cannot be blocked even with commercially available biotin-blocking reagents, shown here with CDX-2 IHC test (D), but when run with biotin-free detection systems false cytoplasmic positivity disappears (E). It is important to note that this is a specific reaction between avidin and biotin and therefore does not represent nonspecific background. It is also difficult to block.

RECOMMENDATIONS

Recommendations are summarized in Table 2. Negative controls for IHC on cell blocks in cytology should generally follow the same rules as employed for tissue blocks.

NRCs (NRC-primAb and NRC-detSys) (Nonspecific Negative Controls)

Separate slide NRCs should be processed in the same manner as the slides for specific IHC tests, including use of identical epitope retrieval procedures. This requirement is automatically achieved for internal and “on-slide” external negative controls, except of course that the latter will have experienced different preanalytic treatment (tissue processing).

The number of slides to be used as NRCs is also determined by the number of different pretreatment procedures: 1 NRC should be prepared for each methodological variation employed in a given clinical case. For example, if 3 different epitope retrieval procedures are used in a panel of several different Abs [eg, HIER (heat-induced epitope retrieval) in citrate buffer, HIER in EDTA, and protease digestion], 3 NRCs (1 processed with HIER in citrate buffer, 1 with HIER in EDTA, and 1 with protease digestion) should be prepared.

However, experience has shown that this approach is not necessary when polymer-based/multimer-based detection systems are used (after initial validation), although it is recommended if an avidin-biotin–based detection system is employed.

Interpretation

None of the NRCs are able to completely exclude the possibility of undesirable/unexpected cross-reactivity of the primary Ab with some epitopes. Specific negative controls (tissue controls) may detect such undesirable cross-reactivity, but detection depends on the presence of such epitopes in the internal or external control tissue, and also upon critical evaluation by the pathologist, to recognize that nonspecific staining is present. Therefore, negative results with NRCs do not ensure specificity of IHC tests in all cases, and when unexpected reactivity of the primary Ab is encountered, the potential for false-positive results still needs to be considered. Careful evaluation of NTCs (Table 1) is still required even if the NRCs are negative.

“Fit-for-Use” Principle

Pathologists use IHC tests for different purposes in diagnostic practice, and different weighting is given to the IHC findings, commensurate with intended use.

In the United States, the Food and Drug Administration (FDA) has included IHC tests under the category of diagnostic devices (Code of Federal Regulations. Title 21). The FDA classifies IHC test/devices in 3 classes (classes I to III) on the basis of risk to the patient.⁵⁰

1. Class I IHC test results are interpreted in the context of morphology and other relevant laboratory or clinical information. Class I IHC tests are used after the primary diagnosis of tumor has been made by conventional histopathology using nonimmunologic stains, such as hematoxylin and eosin. These IHC tests provide the pathologist with adjunctive diagnostic information that may be incorporated into the pathologist’s report, but is not ordinarily reported to the clinician as an independent finding.

2. Class II IHC tests are intended for the detection/measurement of target analytes to provide prognostic or predictive data that are not directly confirmed by routine histopathologic methods. Class II tests provide information that is ordinarily reported as independent diagnostic information to the ordering clinician. Claims associated with data from class II IHC tests are widely accepted and supported by valid scientific evidence.

3. Class IHC III tests are defined (by the FDA) by exclusion, as any IHC test not falling under classes I or II. They are of higher risk, complex, and are often used as standalone findings, not confirmed by other pathologic criteria. Class III devices are subject to Premarket Approval by the FDA, with defined guidelines for use, including controls, exemplified by the HercepTest, Dako as the prototype for the series.⁵⁰

It is important to recognize that the FDA classification relates to IHC devices and reagents and it is up to the end user (pathologist) to determine how and for what purpose various IHC tests are used (eg, ALK-1 in lung cancer is a “predictive” IHC test vs. ALK-1 for lymphoma, which is a diagnostic IHC test). On the basis of the “fit-for-use” concept, and the level of QA required in the clinical laboratory, the Canadian Association of Pathologists National Standards Committee has put forward recommendations that IHC tests/protocols (vs. IHC devices and reagents) be classified based on the “clinical use of the IHC tests” as follows:

1. Class I IHC tests—pathologists are the end users and the IHC results are used for diagnostic work up in the context of morphology, other laboratory data, and clinical information: usage generally is identical to class I IHC devices in the FDA classification, and

2. Class II IHC tests—the results of the IHC test are used by clinical physicians treating the patient for prognosis, treatment selection, or other clinical decisions, including selection for additional genetic testing/evaluation, etc.; usage encompasses both class II and class III IHC devices in the FDA classification.¹³

“Fit-for-use” principles should be used in selection of controls for these different types of IHC applications.

As noted above, recommendations for the use of NRCs also depend on the detection system.

• For biotin-based detection systems, NRCs are required and the number and type is defined by the different pretreatment methods used (1 NRC slide for each pretreatment method).

• For polymer/multimer detection systems, the recommendations differ and are amended as

  • Class I IHC tests—it is recommended that pathologists should order NRCs whenever a new Ab, detection system, retrieval process, or protocol is set up in the laboratory. Subsequently, and based upon experience, NRCs may be run when ordered by pathologists upon perceived need. Examples may include: (i) when the tissue has pigment that may interfere with interpretation of IHC staining; and (ii) when the tissue contains only tumor and there are no cells/tissues that serve as a suitable internal negative control. This latter option is particularly important when only 1 IHC test is ordered or when there is an insufficient panel of Abs and/or detection systems available for comparison (see below); (iii) when an NRC was not ordered initially, but there is suspicion of a nonspecific false-positive result. In such situations, the IHC test that produced the potential false-positive result should be repeated and a matching NRC should be ordered at that time.
  • When a panel of Abs is used on serial sections, the results of other tests in the panel may provide sufficient negative control information so that no additional NRCs are needed. This approach involves some proactive interpretation of IHC-SE (see above) by the pathologist who orders the panel. However, on occasion, IHC panels may not be sufficient or appropriate to serve as negative controls and, therefore, additional negative controls should be ordered by the pathologists when needed.
  • Class II IHC tests—if published guidelines address the type of negative controls to be used for a particular class II test, the guidelines should be followed. If the types of negative controls are not described by the published guidelines, principles suggested in this document should be followed. Additional slides with patient’s tissue shall be run as NRCs for each patient and each test in which results are used to direct patient treatment (eg, ER, PR, and HER2). The notable exception is when class II tests are run as panels, in which 1 test fulfils the requirements for reagent negative control for other tests, as is often the case with ER and PR. Results of negative controls for class II IHC test should be included in the pathology report. If separate pathology report is issued for class II tests, the findings still should be integrated with or clearly linked to the final pathology report (integrated reporting is recommended).
  • Screening class II tests: if used as screening tests, for example, ALK-1 in lung cancer, or IHC testing for MMR proteins may not require automatic NRC if the results are to confirmed by subsequent FISH (ALK-1), or other testing and evaluation (as is usual for MMR). Note that some tests, such as ALK-1 can be used as screening test in lung cancer, in which NRCs may not be necessary. However, if the same test is used as a definitive test for selection of patients for targeted therapy, then it is used to direct patient treatment, in which case it is recommended that separate slide NRC is used.

NTCs (SPECIFIC NEGATIVE CONTROLS)

Int-NTCs and Ext-NTCs are described below are always recommended.

Int-NTC

Int-NTCs should always be searched for and, if present, evaluated by the pathologist before interpretation of lesional tissue of interest. Not all tissue will have an appropriate internal negative control.

Ext-NTC

Ext-NTC elements may be present in the external on-slide positive controls: as noted, ideally positive control tissues should be selected so as also to include elements that are known to be negative (nonreactive) for the evaluated marker. In principle, this goal is best achieved by using multitissue positive control blocks (a combination of strongly positive, intermediate/weak, and negative tissue samples) preferably placed on the same slide as the patient sample. Such composite on-slide controls are recommended by several external quality assurance programs. In the absence of available multitissue blocks, single tissue blocks may be used, carefully selected so as to include tissue elements known not to contain the epitope in question.

Evaluation and “quality control” of the results of external positive and negative controls is normally performed by technologists/biomedical scientists before the slides are forwarded to pathologists for evaluation as part of the overall interpretation of the IHC results. Many EQA agencies require documentation of review and approval of such controls on a daily basis.