Quantification of Mucosal Mast Cells in the Gastrointestinal Tract: A Primer for Practicing Pathologists

Most articles that mention mast cells counts in one or more segments of the GI tract report data from patients with systemic mastocytosis, eosinophilic inflammatory conditions, IBS, or tumors. Although some of these publications include data from healthy controls, often in small numbers, most do not. Many of those that do fail to provide details on the criteria that led to the inclusion of an individual as a healthy control. The stain used to visualize mast cells is reported in most studies, but detailed and accurate information on how the counts were performed, such as the microscope lens used, the area of an hpf, the locations where mast cells were counted, and the type of cells considered as countable mast cells, is rarely provided. Most papers reported the number of mast cells (identified by tryptase, chymase, or CD117 immunohistochemical stains, or, less commonly, by the Giemsa, toluidine blue, or pinacyanol erythrosinate²⁶  histochemical stains) detected either in an hpf (usually of unspecified area) or in 1 mm² of mucosa. Some authors have counted only cells present in the lamina propria, whereas others have included those found within the epithelium. Several other methods have been used to measure the density of mucosal mast cells: For example, one group reported the number of cases in which more than 20 mast cells per hpf were counted²⁷ ; others measured the percentage of the area of the lamina propria occupied by tryptase-positive mast cells.^28,29 

To inform the reader of the ranges of mast cell density reported as normal, we have extracted data from English-language publications in which: (1) at least 10 reportedly healthy individuals (“controls”) were included. The criteria authors used to define “normal” or “healthy” were rarely detailed and could not be verified; therefore, readers should interpret these data with caution. (2) Adequate information on the stain used was provided; and (3) mast cell density was reported either as cells/hpf (even when the area of the hpf was not specified) or cells/mm². In the latter case, the number of cells/mm² was converted to cells/hpf (of 0.237 mm²) by dividing by 4.2 the counts per mm², an assumption that may not be correct in every case. Unless otherwise specified, throughout this paper a “high-power field” refers to an area visualized on a microscope using a ×40 lens and a 22-mm Ø ocular and corresponds to an area of 0.237 mm². We accepted counts reported as a mean ± SD or as a median followed by a range. The results of the data from the available control groups are summarized in the Table.

In a 2007 seminal systematic review, Dellon et al³⁰  evaluated the variability in diagnostic criteria for EoE and highlighted the fact that a large proportion of studies did not report the area of a “high-power field,” on which the then emerging criterion of “15 eosinophils/hpf” for the histopathologic diagnosis of EoE rested. Depending on the ocular and objective specifics in an observer’s microscope, the area of an hpf could vary between 0.120 and 0.440 mm², potentially causing dramatic discrepancies between counts that, when performed in the same hpf, would be identical. The authors’ exhortation urging researchers to report counts as cells/mm² was not widely heeded, and although it is accepted that the best practice is to at least define the size of these fields in scientific publications, journals have not enforced this requirement.³¹  For the past decade, most microscopes used for quantifying cells in diagnostic and research work have used a ×40 objective with a ×10 22-mm Ø ocular and an hpf measuring ∼0.237 mm². In the past few years, however, wider and more luminous oculars have increased the areas evaluated to 0.280 mm² or larger. This phenomenon translates to increasingly used digital pathology systems, where square, rectangular, or circular ×400-equivalent fields are created with area values set by the user. In light of the lack of a standard area, wide variability in these digital hpfs is likely. We therefore recommend that all counts be reported in numbers of cells per mm². Pathologists should be responsible for the conversion from the area of their hpf to square millimeters. In keeping with this recommendation, all counts provided in this primer are expressed in mast cells/mm². An easy-to-use conversion table, adapted from the 2019 World Health Organization Classification of Tumours of the Breast,³²  is available in the 2021 publication by Cree et al.³¹ 

For decades after Ehrlich’s discovery, toluidine blue, a stain that exploited their metachromasia, was used to highlight mast cells in tissues. Other stains, such as pinacyanol erythrosinate,²⁶  Bismarck brown,³³  methyl green–pyronin,³⁴  and ruthenium red,³⁵  were occasionally employed for experimental studies, but they never became widely used. Although it is still necessary for the detection of mast cells in animal tissues, toluidine blue has been almost completely abandoned for human use, after it became clear that tryptase (alone, or associated with chymase) was a universal component of mast cells, and immunohistochemical stains were developed to specifically stain tryptase.³⁶  Chymase, used in some studies, is characteristic of mast cells, but not all mast cells contain it. Specifically, human GI mast cells are either devoid of, or contain only small amounts of, chymase, which makes an anti-chymase immunohistochemical stain unsuitable for the detection of gut mast cells.³⁷  CD117 (c-Kit) stains human mast cells, but it also stains the interstitial cells of Cajal, as well as endothelial cells, hematopoietic progenitor cells, melanocytes, embryonic and fetal brain, endothelium, gonads, breast epithelium, and germ cells.³⁸  Several of these may be found in the GI mucosa and are therefore potential confounders when counting mast cells.

Figure 1 compares CD117 and tryptase staining. Figure 1, A, shows that staining with CD117 results in cleaner sections, almost completely free of the background stain frequently found when anti–mast cell tryptase is used (Figure 1, B). However, the lack of specificity of CD117 may be a hindrance, particularly in the GI tract where interstitial cells of Cajal may be abundant. Therefore, most researchers prefer using tryptase, which is unique to mast cells, and learn to ignore the imperfections of the stain. In addition, background staining with tryptase may not always be artifactual (per Alain). Studies relating the background staining to serum tryptase levels have not been published, and therefore the “unclean” tryptase stains may be a consequence of extraintestinal pathology. Until we know if biopsies showing background have elevated serum tryptase levels, tolerating the less-than-clean staining is prudent.

Although a very experienced pathologist may have learned to separate mast cells from lymphocytes in the esophageal squamous epithelium (Figure 2), this is essentially impossible in the remainder of the GI tract, where the columnar mucosa contains a population of resident inflammatory cells that cannot be distinguished from mast cells. Thus, we can confidently state that the evaluation of mast cell infiltrates requires a stain that specifically highlights these cells, and, for reasons detailed earlier in this review, we suggest the use of anti–mast cell tryptase.³⁶ 

The most common reason for a pathologist to perform an unusual diagnostic activity is a request by the clinician who submits the specimen. The ubiquitous “rule out…” may trigger a search for unexpected conditions or the performance of a special stain or a molecular test, but to make a pathologist count cells, a direct request is usually necessary. Such requests are becoming increasingly common for eosinophils when a diagnosis of EoE is suspected, but they remain relatively rare for mast cells. Occasionally, when a young or middle-aged woman with a diagnosis of IBS fails to respond to the usual management, well-intentioned clinicians suspect mastocytic enterocolitis and request mast cell counts in colonic, ileal, and sometimes duodenal biopsies.^8,27–29,39  Although several studies have reported increased mast cell density in the intestinal mucosa in a subset of patients with refractory IBS, other studies have yielded conflicting findings.³⁹  Furthermore, in the absence of accepted standards for normal, the denotation “increased” has little contextual meaning. The best a pathologist can do in these circumstances is to report the peak or the average counts of a few hpfs (usually 3 or 5) for each biopsy site and appraise their clinical audience of the potentially wide variation of normal mast cell numbers throughout the GI tract, based on currently available literature.⁸  Other than when requested to do so, there are currently no circumstances when a practicing pathologist should spontaneously report the counts of mucosal mast cells in GI biopsies. In contrast, as the possible pathogenic role of mast cells in EoE and EGIDs is attracting interest, studies requiring the quantification of mast cell infiltrates are slowly becoming increasingly common.^15,16,40,41 

The purpose of counting is to express the density of mast cells in one area of the mucosa in a measurable manner that can be readily compared to measurements performed in other areas and in other specimens. Thus, our first concern must be how representative a section is. A suboptimally oriented mucosal section may yield skewed counts because the plane of section might include only areas with greater or lower densities of cells than one would be able to see in a section where both the surface epithelium and the muscularis mucosae are included. Examples of ideal specimens for the esophagus, stomach, small intestine, and colon are depicted in Figure 3.

The general approach to a mucosal biopsy examined to assess the density of mast cells is the same, irrespective of the segment of the GI tract. Each biopsy specimen should first be examined at low magnification (×40 and ×100) to evaluate for proper orientation (crucial to generate reproducible mast cell counts) and for the presence of lesions or conditions (eg, features associated with EoE,^12,14,42 Heliobacter pylori gastritis,^43,44  celiac disease,⁴⁵  IBS^27,46,47) that could affect the density of mast cells or make their quantitation less relevant. If no other pathologic features are present, each section should then be examined at medium magnification (×200) to detect areas with unusual mast cell density or clustering. In contrast to eosinophils, which may be extremely patchy even within the same biopsy fragment, mast cells tend to be distributed more uniformly. Therefore, unless clusters are seen, one can count a set of adjacent, nonoverlapping hpfs (×400) and expect them to contain similar numbers of mast cells.

A special situation exists in the esophagus. Even when absent from the squamous epithelium, mast cells are frequently present within the lamina propria of the papillae (Figure 4), and they should not be included in the counts. In all segments of the GI tract mucosal orientation is of foremost importance because tangential sections can show artificially high concentrations of mast cells. Likewise, pinched, crushed, or twisted sections (Figure 5) as well as heavily degranulated areas should be avoided. Also, it is crucial to focus on the correct compartment (ie, the mucosa) and not count cells in the lamina propria of the esophagus or the submucosa of the stomach and intestine.

Mast cells can have many appearances, sizes, and shapes, but most studies dealing with mast cell counting do not detail the criteria they used for defining a “countable” cell. A few studies have suggested only counting nucleated cells; however, because the most commonly used stains stain the granules and not the nucleus, what is and what is not a nucleus can be difficult to ascertain. One does not “see” the nucleus on a tryptase stain; rather, one looks for the unstained space where the nucleus is. Eosinophils are counted on hematoxylin-eosin–stained slides, where their granules are usually distinct. Conversely, mast cells are counted using a stain that, although specific in the sense that the only cells it stains are mast cells, also highlights fragments of cells, granules, possibly other debris, and often leaves a light brownish hue in the lamina propria, increasing the difficulty and subjectivity.

Figure 6 shows examples of mast cells with different shapes, clusters, and degranulation. Recognizing unequivocally countable mast cells is not difficult once an observer learns to appreciate the variations of shape, from round (Figure 6, A and C) to fusiform (Figure 6, B). Normal mast cells in aggregates or clusters (Figure 6, D) tend to maintain their roundish shape, even when closely adjacent to one another. Loose granules (resulting from either physiologic degranulation or the mechanical trauma of the biopsy forceps) are easily identified and must not be counted. The main difficulty comes in deciding whether tryptase-positive corpuscles smaller than intact mast cells but larger than granules (Figure 6, F) should be counted. Such corpuscles may result from partially sectioned mast cells, fragmented cells, aggregates of granules, and nonspecific staining. Because visual observation, even at high magnification, rarely if ever yields provable answers, we suggest not to include these uncertain particles in the counts.

In the collective experience of the authors of this primer, individuals with elevated mast cell counts tend to have aggregates of cells that cluster together (Figures 6, D, and 7, B), morphologically analogous to “eosinophilic microabscesses” and without satisfying diagnostic criteria for systemic mastocytosis, as summarized in the following section. We have agreed to arbitrarily define these clusters as “aggregates of at least 3 mast cells touching one another.” Although their significance remains elusive at this time, we suggest that clusters be reported in the hope of creating a sufficient base of knowledge to allow for mechanistic interpretations.

When evaluating mast cell numbers and aggregates in gut biopsies, one must take care to differentiate clusters of normal mast cells (as defined above) from involvement of the gut by systemic mastocytosis, a clonal neoplastic proliferation of mast cells. The current 4th edition of WHO Classification of Tumors of Haematopoietic and Lymphoid Tissues (2017)⁴⁸  defines systemic mastocytosis as abnormal infiltrates of mast cells accumulating in one or more organ systems, involving at least one extracutaneous organ.^49,50  The diagnosis may be made either when 1 major criterion and at least 1 minor criterion are satisfied, or when at least 3 minor criteria are present in the absence of the major criterion. Major criterion is the presence of multifocal, cohesive, dense infiltrates of mast cells, with 15 or more mast cells in the aggregate, in bone marrow and/or an extracutaneous organ (Figure 8, A and D). The minor criteria include: (1) more than a quarter of mast cells in the infiltrate with spindled or atypical morphology in tissue sections; (2) the detection of activating point mutation D816V at codon 816 of the KIT gene on a bone marrow, blood, or extracutaneous sample; (3) the aberrant expression of CD25 (a more sensitive marker; Figure 8, D), with or without CD2, by the mast cells in addition to normal mast cell markers; and (4) a persistent elevation of serum total tryptase greater than 20 ng/mL (this latter parameter is not valid if there is an associated myeloid neoplasm). If the above criteria are satisfied and the diagnosis of gut involvement by systemic mastocytosis can be made or suggested, the mast cell counts are not needed beyond appreciating clusters composed of at least 15 cells. Conversely, nonneoplastic mast cells should not be significantly clustered, have significantly aberrant morphology, aberrantly express CD25, or harbor the characteristic KIT D816V mutation.⁴⁹  As seen in Figure 8, B, expression of mast cell tryptase is more variable; in one study, of 24 involved biopsies stained with tryptase, 19 of 24 (79%) showed positive staining in only a subset of neoplastic mast cells.⁴⁹ 

In the workup of idiopathic MCAS, particularly with GI symptoms, it is unclear whether counting mast cells is beneficial. Mast cell counting identifies increased numbers of mast cells, yet MCAS can show normal numbers of mast cells that are overactive. Previous studies demonstrated that morphologically normal mast cells were present in appropriate numbers with a normal distribution within the mucosa in MCAS.^49,51  Increased mast cell numbers are also not specific for MCAS. Duodenal mast cells can be increased in patients with functional GI diseases, for which many of the symptoms overlap with MCAS.¹⁷  On the other hand, to the extent that a rationale for specific therapy, such as high-dose H1 and H2 blockers or oral cromolyn, is being considered, counting mast cells may support the treatment decision.

In some biopsy specimens there are tryptase-stained granules of relatively uniform size (approximately 1–2 μm in diameter) seemingly loose in the tissue between intact or fragmented mast cells (Figure 5, E). The mechanisms of this degranulation, which is known to occur in response to mast cell activation,⁵²  cannot be determined with certainty by the pathologist who observes a static image: it could be activation that occurred before the biopsy was taken or the result of the mechanical forces exerted by the biopsy forceps. A simple semiquantitative degranulation score based on the percentage of mast cells exhibiting degranulation has been recently proposed by 2 of the authors (R.M.G. and K.O.T.) and is currently awaiting validation.⁵³  Although we suspect that in most cases finding substantial mast cell degranulation in a GI biopsy may indicate a process that antedates the biopsy procedure and has biologic significance, we recommend that degranulation be reported without speculating about its likely origin. Just like in the case of clusters, the creation of a large body of data may eventually help reach a better understanding of this phenomenon.

We suggest that pathologists develop a systematic method of counting and reporting mast cells. Although various methods can be devised, we feel that becoming comfortable with one of them will help reduce barriers and increase consistency and speed. Once a field is selected to count, an orderly approach should be adopted to ensure that the entire area is evaluated. This can be accomplished by counting by quadrant or using a “lawnmower” sweeping pattern. Care should be taken to not overlook cells that are squashed between glands or deep in the lamina propria in the subcryptal area. Although in the esophagus only mast cells within the squamous epithelium are counted, different observers have different approaches on whether intraepithelial mast cells in the columnar mucosae of the stomach and intestine should be included in the counts. The exact location of mast cells within the mucosa is not always easily detected, especially if the counterstain is light; furthermore, mast cells within the columnar epithelium are not particularly common. Therefore, we suggest that all mast cells above the muscularis mucosae should be counted, irrespective of their location (ie, in both lamina propria and epithelium).

Digitized images from tryptase-stained sections show mast cells as clearly as in glass slides; even if the colors are somewhat altered, as often happens with scanned images, it is always a clear “brown on pale blue” contrast, and no additional difficulties are encountered. As such, artificial intelligence (AI)–assisted automatic counting is possible for mast cells because the immunohistochemical stain commonly used to detect them (mast cell tryptase) has good specificity, and the generally pale background makes their detection easy. Commercially available algorithms are available and have been used in several studies, including for assessment of mast cells in esophageal biopsies in EoE.^54–58  The general steps in this process are as follows. First, the area of interest for counting is selected. This selection should use the same principles outlined above, so in the esophageal epithelium for example, lamina propria should be excluded from the selection. Different AI programs will have different ways to do this selection, some manual and some automatic. Next, the AI algorithm is run. The exact processes will depend on the algorithm. Some will count specific cells (which often requires training the algorithm), others will count staining intensity of individual pixels, and units can be converted to cell counts. It is also possible to train an algorithm to count granules in intercellular areas. The final step is converting the data generated in the algorithm, which are typically in units of density per unit area, into cell counts. These types of algorithms have primarily been used for research purposes and have several pitfalls, including the need for training, uniform and good quality staining, and selection of appropriate areas for either general peak or overall cell counts. In addition, nonspecific background staining may hinder the detection of stained cells (Figure 1, B), and the presence of aggregates of granules (Figure 6, E) and fragments of possible mast cells (Figure 6, F) may require the design of complex algorithms before AI-assisted systems can be confidently used.

In our experience, the distribution of mast cells in the GI tract is generally less patchy than that of eosinophils, making peak counts reliable indicators of the density of mast cells. In clinical practice, a pathologist could determine a peak count, then estimate if more, or less, than half of the tissue contains the peak count, and use that estimate to report diffuse or focal mast cell infiltration, along with the peak count expressed as mast cells per mm². In clinical research, the method for reporting mast cell density is at the discretion of the investigators. In the experience of some of the authors, some protocols have requested the peak counts in 5 nonoverlapping hpfs per biopsy fragment, a task that some pathologists might find too onerous even for research purposes. Although consensus meetings on this topic do not appear to be on the horizon, evidence-based recommendations guiding both clinical practice and research protocols will be indispensable to establish a uniform approach to the evaluation of mast cells.

To our knowledge, neither intraobserver nor interobserver variability in counting mast cells has been formally assessed. Pathologists working together on the same project (as, for example, 2 pairs of the authors, R.M.G.-K.O.T. and J.B.W.-N.C.A.) tend to train together by first establishing rules on what, how, and where to count, then by spending time at a double-headed microscope (or remotely, using digitized images) discussing discrepancies and problem cases. Such joint training has excellent chances of resulting in a highly satisfactory degree of uniformity among observers.

However, not all pathologists can take advantage of such one-to-one training. Therefore, this review is primarily aimed at helping pathologists who work alone and have no immediate access to colleagues experienced in the evaluation of mast cells. We hope that following the guidelines and suggestions outlined in this primer can help them become confident in reporting the intensity of mast cell infiltrates with accuracy and consistency.