best practices in diagnostic immunohistochemistry - ipath-network

Arch Pathol Lab Med—Vol 131, October 2007 IHC in Pleomorphic Cutaneous Tumors—Folpe & Cooper 1517

CAP Laboratory Improvement Programs

Best Practices in Diagnostic ImmunohistochemistryPleomorphic Cutaneous Spindle Cell Tumors

Andrew L. Folpe, MD; Kumarasen Cooper, MBChB, DPhil

● Context.—Pleomorphic cutaneous spindle cell tumorscan be difficult to distinguish solely on histologic grounds.The use of ancillary immunohistochemical studies cangreatly assist in this differential diagnosis.

Objective.—To review histologic and immunohisto-chemical aspects of cutaneous spindle cell tumors and dis-cuss a basic panel of markers to assist in the differentialdiagnosis.

Data Sources.—English-language literature published be-tween 1981 and 2005.

Conclusions.—A basic immunohistochemistry panel forhigh-molecular-weight cytokeratin, melanocytic markers(S100 protein, HMB-45, Melan-A), smooth muscle actin, des-min, and endothelial markers (CD31, CD34) is effective indiagnosing most cutaneous spindle cell tumors.

(Arch Pathol Lab Med. 2007;131:1517–1524)

This article focuses on the use of immunohistochemistry(IHC) in the differential diagnosis of a relatively com-

mon diagnostic scenario, the ‘‘histologically undifferenti-ated, pleomorphic spindle cell tumor involving the skin.’’Applications of IHC in the diagnosis of monomorphicspindle cell tumors (eg, dermatofibrosarcoma protuber-ans) and histologically distinctive spindle cell tumors (eg,neurothekeoma) will be covered in subsequent articles inthis series.

DIFFERENTIAL DIAGNOSTIC CONSIDERATIONS

Most pleomorphic cutaneous spindle cell tumors occurin adults, and many involve sun-damaged skin. Thus, thetypical differential diagnosis for these lesions is actuallyquite limited. The most common tumors encountered inthis setting are sarcomatoid squamous cell carcinoma(SCC), sarcomatoid malignant melanoma, and atypical fi-broxanthoma (AFX)/superficial malignant fibrous histio-cytoma. Less common tumors that also need to be includ-ed in the differential diagnosis include leiomyosarcomaand spindle cell variants of angiosarcoma and very rarely

Accepted for publication June 6, 2007.From the Division of Anatomic Pathology, Mayo Clinic, Rochester,

Minn (Dr Folpe); and the Department of Pathology, University of Ver-mont, Burlington (Dr Cooper).

The authors have no relevant financial interest in the products orcompanies described in this article.

This article is provided for educational purposes only and is not in-tended to suggest either a practice standard or the only acceptablepathway for diagnostic evaluation. The views presented reflect the au-thors’ opinions. The application of these opinions to a particular med-ical situation must be guided by the informed medical judgment of theresponsible pathologist(s) based on the individual circumstances pre-sented by the patient. The College of American Pathologists has noresponsibility for the content or application of the views expressedherein.

Reprints: Andrew L. Folpe, MD, Division of Anatomic Pathology,Mayo Clinic, 200 1st St SW, Rochester, MN 55905 (e-mail: [email protected]).

unusual tumors such as rhabdomyosarcoma, pleomorphicliposarcoma, and so forth.

RECOMMENDED ANTIBODY PANEL

A panel of commonly evaluated antigens targeted to-ward the differential diagnosis is outlined in Table 1. Is-sues specific to each antigen and to each potential diag-nosis are discussed in the ‘‘Discussion of Antigens’’ and‘‘Issues Specific to Particular Tumor Types’’ sections, re-spectively. A logical approach would be to choose a ge-neric panel of antibodies in the first instance, followed bya secondary step composed of fewer, more specific anti-bodies based on the results of the initial panel. Such analgorithmic approach ensures maximum sensitivity andspecificity of the antibody panel.

BASIC PRINCIPLES OF DIAGNOSTIC IHC AS APPLIEDTO THE DIFFERENTIAL DIAGNOSIS OF CUTANEOUS

PLEOMORPHIC SPINDLE CELL TUMORS

Careful Attention to Positive and NegativeInternal Controls

Careful attention to positive and negative internal con-trols cannot be overemphasized. Every single IHC slidemust be evaluated for appropriate staining of normal tis-sues for the antigen in question. Internal positive and neg-ative controls completely ‘‘trump’’ external controls,whether those external controls are placed on the sameglass slide or not. This is because specimen fixation is thesingle greatest variable in IHC. Fortunately for the pa-thologist, the skin is rich in normal internal positive con-trols for the majority of antigens (Table 2). Evaluation ofnegative controls simply means making sure that tissuesthat do not express an antigen are negative. Inappropriatestaining of negative controls usually is the result of in-appropriate antibody concentration and/or excessive epi-tope retrieval.

1518 Arch Pathol Lab Med—Vol 131, October 2007 IHC in Pleomorphic Cutaneous Tumors—Folpe & Cooper

Table 1. Recommended Panel for the Workup of Common Pleomorphic Cutaneous Spindle Cell Tumors, With ExpectedImmunophenotypes*

Cytokeratins(Wide-Spectrum,

High MW, CK5/6) S100 ProteinMelanocytic Markers(HMB-45, Melan-A)

Smooth MuscleActin Desmin

EndothelialMarkers

(CD31, CD34)

Sarcomatoid SCC � � � � � �Melanoma �/� � �/� � �/� �AFX � � � �/� � �Leiomyosarcoma �/� �/� � � �/� �/�Angiosarcoma �/� � � � � �

* MW indicates molecular weight; SCC, squamous cell carcinoma; AFX, atypical fibroxanthoma; �, positive in more than 90% of cases; �,negative; �/�, positive in 50% to 75% of cases; and �/�, usually negative but anomalous expression may be seen in up to 25% of cases.

Table 2. Commonly Evaluated Antigens and TheirNormal Positive Internal Controls in the Skin

AntigenNormal Positive Internal

Control in Skin

Pan-cytokeratin and low-molec-ular-weight cytokeratins

Suprabasal keratinocytesand adnexae

High-molecular-weight cytokera-tins

Basal keratinocytes

S100 protein Melanocytes, Langerhanscells, nerves, myoepi-thelial cells

Melanocytic markers (eg, HMB-45, Melan-A, MiTF, tyrosi-nase)*

Dermal melanocytes (maybe HMB-45 negative)

Smooth muscle actins Pilar smooth muscleDesmin Pilar smooth muscleEndothelial markers (eg, CD31,

CD34, FLI-1 protein, von Wil-lebrand protein)

Endothelial cells

* MiTF indicates microphthalmia transcription factor.

Table 3. Cytokeratin Expression in Normal Skin

Cell Type Cytokeratins Expressed

Suprabasal keratinocytes 1, 10, (11)Basal keratinocytes 5 and 14Hair shaft 5, 6, (14), 15, 16, 17Eccrine ductular cells 7, 8, 17, 18, 19Eccrine myoepithelial cells 5, 6, 14

Careful Evaluation of All Pieces of Tissue on theEntire Slide

Careful evaluation of all pieces of tissue on the entireslide should be a routine exercise; however, many casesare simply missed because a small focus of cytokeratin-or HMB-45–positive cells was overlooked in a sarcomatoidcarcinoma or melanoma, respectively.

The Use of a Panel of Immunostains, Including BothExpected Positives and Negatives for All of the Entities in

the Histologic Differential DiagnosisA panel of immunostains, rather than single markers,

should always be performed. This approach greatly elim-inates the potential for misdiagnosis owing to anomalousexpression of antigens (eg, cytokeratin expression in mel-anoma or angiosarcoma). The routine use of a small, care-fully selected panel is also time and money effective.

DISCUSSION OF ANTIGENSCytokeratins

Background. The cytokeratins are a family of 20 in-termediate filament proteins, the expression of which islargely, but not entirely, restricted to epithelial cells. Thereare many different ways to approach the cytokeratins, asindividual cytokeratins (eg, CK7, CK20), as acidic and ba-sic cytokeratins pairs (eg, CK8/18), or as ‘‘low’’ and‘‘high’’ molecular weight cytokeratins. With regard to thelast category, it is important to realize that the division ofthe cytokeratins into low and high molecular weights isentirely arbitrary. It is more useful to think of the low-

molecular-weight cytokeratins as those of simple epithelia,such as simple ductules, and the high-molecular-weightcytokeratins as those of complex epithelia, such as urothe-lium or skin.

For practical purposes, it is simpler to consider cytoker-atins in terms of the antibodies used to identify them. Themost widely used pan-keratin antibody is the AE1/AE3cocktail. AE1 recognizes the acidic cytokeratins 10, 14, 15,16, and 19, whereas AE3 recognizes the basic cytokeratins1, 2, 3, 4, 5, 6, and 8. Both antibodies recognize a mixtureof high- and low-molecular-weight cytokeratins. So-calledwide-spectrum cytokeratin antibodies, such as the OSCARMAb and a variety of polyclonal antibodies, potentiallyhave somewhat broader cytokeratin coverage than doesthe AE1/AE3 cocktail, although this may be laboratorydependent. The most widely used low-molecular-weightcytokeratin antibodies are CAM 5.2 (CK8, CK18, CK19)and 35BH11 (CK8 and CK18). Almost all laboratories usedthe high-molecular-weight cytokeratin antibody 34BE12(CK1, CK5, CK10, CK14/15), also known as ‘‘cytokeratin903.’’ Many other antibodies exist and one should alwayscarefully read the product insert and any relevant litera-ture before using them in the laboratory.

Cytokeratin Expression in Normal Skin. The skinserves as a superb control tissue for the evaluation of cy-tokeratin expression, as each epithelial cell type presentwithin the skin expresses unique cytokeratins, as detailedlater (Table 3).1

Falsely Negative Basal Keratinocyte Staining WithAE1/AE3 and Other Wide-Spectrum Cytokeratin Anti-bodies. Although these antibodies should react withbasal keratinocytes (CK5 and CK14), these cells are fre-quently negative in many laboratories, for unknown rea-sons (Figure 1). This indicates that these high-molecular-weight keratins are not being identified. It is critical torealize that sarcomatoid SCCs may express almost exclu-sively high-molecular-weight cytokeratins such as CK5and CK14.2–4 Therefore, the failure to appreciate negativebasal cell staining with wide-spectrum cytokeratin anti-bodies may result in the misdiagnosis of sarcomatoidSCC.


Figure 1. Normal skin, immunostained with AE1/AE3. Note the negative basal cell staining, reflecting the sometimes variable sensitivity of thisantibody cocktail for high-molecular-weight cytokeratins. As sarcomatoid squamous cell carcinomas may express only high-molecular-weightcytokeratins, failure to appreciate this finding may result in a ‘‘false negative’’ study (original magnification �100).

Figure 2. A, Sarcomatoid squamous cell carcinoma, arising from a basaloid squamous cell carcinoma (hematoxylin-eosin, original magnification�100). B, The sarcomatoid component of this tumor, but not the basaloid component, shows strong vimentin expression (original magnification �200).

Figure 3. Malignant melanoma, negative for S100 protein. Note the positive internal control, Langerhans cells (original magnification �400).

Figure 4. Normal melanocytes, strongly positive with HMB-45. Positivity with this marker does not necessarily imply a malignant melanocyticlesion (original magnification �100).

Figure 5. Angiosarcoma, diffusely positive for CD31. There are also a large number of CD31-positive histiocytes, showing somewhat fainter,more granular staining. Intratumoral CD31-positive histiocytes may be misinterpreted as positive tumor cells, leading to the erroneous diagnosisof ‘‘angiosarcoma’’ (original magnification �200).


VimentinBackground. Vimentin is present in almost all embry-

onic cells, and most cells of any lineage re-express vimen-tin when grown in culture. This latter situation may beanalogized to spindled change in neoplasms, inasmuch asessentially any spindled cell will express vimentin.5–9

Diagnostic Uses. Vimentin is expressed by essentiallyall sarcomas, all melanomas, all spindled SCCs, all spin-dled angiosarcomas, and so forth (Figure 2, A and B). Itis also expressed in almost all mesotheliomas, most glio-mas, and a subset of nonspindled carcinomas. The diag-nostic uses of vimentin are almost exclusively restrictedto the subtyping of carcinomas, as certain carcinomas,such as endometrioid carcinomas, frequently coexpress vi-mentin.10

S100 ProteinBackground. S100 protein is a 20-kd acidic calcium-

binding protein, which gets its name from its solubility in100% ammonium sulfate. S100 protein is composed of 2subunits, with 3 isotypes—�� (expressed in muscle), ��(expressed in melanocytes, glia, chondrocytes, skin ad-nexae), and �� (expressed in Langerhans cells, Schwanncells).1 The commonly used polyclonal antibodies recog-nize all of these isotypes.

Diagnostic Uses. S100 protein is present in essentially100% of normal melanocytes and nevi and in approxi-mately 97% to 98% of melanomas.11–13 S100 protein is thusthe single best screening marker for melanocytic neo-plasms. It is important to realize that very rare melanomasare S100 protein negative (Figure 3). S100 protein is not,however, completely specific for melanocytic tumors andcan be expressed in smooth muscle tumors (weakly), somecarcinomas (but not SCC), myoepithelial tumors, periph-eral nerve sheath tumors, and others.

Melanocytic MarkersHMB-45. Background. HMB-45 is a monoclonal anti-

body that identifies a premelanosomal protein, gp100.14

HMB-45 is often negative in resting melanocytes and nevi,but may be expressed in ‘‘reactive’’ melanocytes (Figure4). HMB-45 is organelle specific but not lineage specific,meaning that nonmelanocytic tumors that contain pre-melanosomes (eg, Bednar tumor, angiomyolipoma) willbe HMB-45 positive.15–17 HMB-45 should not be used as ascreening marker for melanoma; S100 protein should beused instead.

Diagnostic Uses. Numerous studies have shown thesensitivity of HMB-45 in epithelioid melanomas to be ap-proximately 85%. Sarcomatoid melanomas are less oftenpositive (30%–50%), and true desmoplastic melanomas areessentially never HMB-45 positive in the spindled/des-moplastic zones.18

Melan-A. Background. Melan-A, the product of theMART-1 gene, is a 20- to 22-kd component of the preme-lanosomal membrane, of unknown function. Unlike HMB-45, Melan-A is expressed in resting melanocytes and nevi.As with HMB-45, Melan-A is organelle specific but notlineage specific.19–23 Melan-A should also not be used as ascreening marker for melanoma. The most widely usedantibody to Melan-A, A103, has a reproducible and di-agnostically useful cross-reactivity with a cytoplasmiccomponent of steroid-producing cells.22 This is helpful inthe diagnosis of adrenal cortical neoplasms but is of noconcern in the skin.

Diagnostic Uses. Melan-A has essentially the same sen-sitivity and specificity as HMB-45. It is present in someHMB-45–negative melanomas and vice versa.

Other Markers of Melanocytes. Other potential mark-ers for melanocytes include microphthalmia transcriptionfactor and tyrosinase. Microphthalmia transcription factorsuffers from relatively low specificity and is not recom-mended.24,25 The sensitivity and specificity of tyrosinaseare roughly equivalent to those of HMB-45 and Melan-A,and it is an acceptable alternative marker.19,26

Markers of Endothelial DifferentiationCD31 (Platelet Endothelial Cell Adhesion Molecule

1). Background. CD31 is a 130-kd transmembrane gly-coprotein that is expressed on all endothelial cells, includ-ing lymphatic endothelial cells. CD31 is also routinely ex-pressed by tissue macrophages and platelets.

Diagnostic Uses. CD31 is the single most sensitive andspecific marker of endothelial differentiation and is thesingle best endothelial marker to have in one’s laboratory,if one can have just one27,28 (Figure 5). CD31 is expressedby essentially all benign vascular tumors and by morethan 90% of hemangioendotheliomas and angiosarcomas.CD31 expression in carcinomas is extraordinarily unusual,with only case reports in breast and thyroid carcinomas.29

Nonendothelial sarcomas, in particular epithelioid sarco-ma, are CD31 negative. CD31-positive macrophages with-in nonendothelial tumors can be confused with positivetumor cells, leading to an erroneous diagnosis of an en-dothelial tumor.30

CD34. Background. CD34, a transmembrane glyco-protein of unknown function, is widely expressed on anumber of normal tissues, including hematopoietic stemcells, interstitial cells of Cajal, endothelial cells, and den-dritic interstitial cells in the skin and around nerves.31

CD34 is thus not a specific endothelial marker but is auseful marker in the appropriate histologic context.

Diagnostic Uses. CD34 is a highly sensitive marker ofbenign, borderline, and malignant endothelial tumors, be-ing expressed by more than 90% of cases.28 CD34 has alsogenerally been regarded as the most sensitive marker forKaposi sarcoma,32 although the Kaposi sarcoma–associat-ed herpesvirus latency-associated nuclear antigen proteinis probably the current marker of choice for Kaposi sar-coma.33 In the differential diagnosis of epithelioid angio-sarcomas, it is important to remember that CD34 is ex-pressed by up to 60% of epithelioid sarcomas but not bycarcinomas.34–36 It is also expressed by dermatofibrosar-coma protuberans, gastrointestinal stromal tumors, leio-myosarcomas, and solitary fibrous tumors.31

FLI-1 Protein. Background. FLI-1 protein is a memberof ETS family of nuclear transcription factors and is ex-pressed in all mature endothelia and during the earlieststages of endothelial differentiation.37 FLI-1 is best knownas the partner of EWS in the Ewing sarcoma–specific t(11;22)(q12;q24) (EWS-FLI-1) fusion gene. It is also routinelyexpressed by small lymphocytes (probably T cells). FLI-1is the only nuclear marker of endothelium.

Diagnostic Uses. FLI-1 is a highly sensitive marker ofendothelial neoplasms, including hemangiomas, heman-gioendotheliomas, angiosarcomas, and Kaposi sarco-ma.37,38 FLI-1 is not expressed by epithelioid sarcomas,melanomas, or carcinomas. It is important not to mistakeintratumoral FLI-1–positive endothelial cells and lympho-cytes for positive tumor cells.


von Willebrand Factor (Factor 8–Related Protein).Background. von Willebrand factor is a clotting factor thatis theoretically present only in the Weibel-Palade body ofendothelium and in platelets and that should be in theorythe most specific marker of endothelial differentiation.39

Unfortunately, von Willebrand factor is secreted into theserum, and the inevitable high background staining seenwith this marker greatly reduces its ‘‘real world’’ utility.von Willebrand factor staining in serum may be seen onlyaround the outer surface of cells and closely mimics truemembranous expression (Figure 6).

Diagnostic Uses. von Willebrand factor is the least sen-sitive endothelial marker, particularly in angiosarcomas.32

Given the technical issues discussed previously, there isarguably no real role for von Willebrand factor in the eraof CD31, CD34, and FLI-1.

Muscle MarkersSmooth Muscle Actin. Background. The actins are a

ubiquitously distributed family of intracellular proteinsthat may be broadly divided into muscle and nonmuscleisoforms. The muscle-specific isoforms can be dividedagain into smooth, skeletal, and cardiac isoforms.9,40–43

From a practical perspective, the single most useful actinantibody is MAb 1A4, which recognizes only smoothmuscle isoforms. In the skin, smooth muscle actin is rou-tinely expressed by pilar and vascular smooth muscle, re-active myofibroblasts, a subset of pericytes, and myoepi-thelial cells.41

Diagnostic Uses. Smooth muscle actin is a highly sen-sitive marker of smooth muscle and myofibroblastic tu-mors in the skin (Figure 7).

Desmin. Background. Desmin is an intermediate fil-ament protein that is expressed by muscle cells of alltypes, as well as submesothelial fibroblasts, a subset oflymph node dendritic cells, and endometrial stromalcells.44–48 Desmin expression is uncommon in myofibro-blasts.

Diagnostic Uses. Desmin is best used as a screeningmarker for tumors of skeletal muscle differentiation, whichare extraordinarily rare in the skin. Although desmin ex-pression is almost always present in pilar smooth muscletumors, it may be absent in smooth muscle tumors of vas-cular smooth muscle origin (vascular smooth muscle lesscommonly expresses desmin, for unknown reasons).Smooth muscle actin is a much better screening markerfor smooth muscle tumors. When the differential diagno-sis includes true smooth muscle and myofibroblastic tu-mors, strong desmin expression supports true smoothmuscle differentiation. Anomalous desmin expressionmay be seen in melanoma, schwannoma, tenosynovial gi-ant cell tumor, Ewing sarcoma, and angiomatoid (malig-nant) fibrous histiocytoma, among others.49–59

ISSUES SPECIFIC TO PARTICULAR TUMOR TYPESSquamous Cell Carcinoma

Extremely Focal or Absent Cytokeratin Expression inSarcomatoid SCC. Sarcomatoid SCCs may show onlyextremely focal cytokeratin expression, may entirely losecytokeratin expression in spindled areas, or may expressonly high-molecular-weight cytokeratins (Figure 8). Themost sensitive markers of sarcomatoid SCC are 34BE12and CK5/6.2–4 It is also important to recognize that essen-tially all sarcomatoid SCCs will express vimentin, as dis-cussed previously in ‘‘Discussion of Antigens.’’

Melanoma

Cytokeratin and/or Desmin Expression. Melanomasare notorious for anomalous expression of intermediatefilament proteins, with some series documenting anoma-lous expression in more than 30% of cases.60 Cytokeratinis the most common anomalously expressed intermediatefilament in melanoma, obviously raising significant poten-tial for misdiagnosis of cytokeratin-positive melanomas ascarcinoma. Use of a panel of immunostains, includingS100 protein, should greatly alleviate this problem. In gen-eral, anomalous intermediate filament expression tends tobe focal in nature, although it may rarely be diffuse. Wehave seen desmin-positive melanomas misdiagnosed asrhabdomyosarcoma and leiomyosarcoma; additional IHCwith markers such as HMB-45, myogenin, and muscle ac-tins is necessary for arriving at the correct diagnosis inthese unusual cases (Figure 9).

CD68 Expression. CD68 (KP1) is a relatively nonspe-cific marker of lysosomes, rather than a lineage-specifichistiocytic marker.61 CD68-positive melanomas may bemistaken for AFXs, if one is using this marker in an at-tempt to confirm ‘‘fibrohistiocytic differentiation.’’ CD68has a very limited role in the diagnosis of cutaneous spin-dle cell tumors, as it may also be expressed in angiosar-comas, carcinomas, and leiomyosarcomas, among oth-ers62–65 (Figure 10, A and B).

Entrapped Smooth Muscle Actin–Positive Myofibro-blasts. Myofibroblasts are frequently found intimatelyassociated with the neoplastic cells of desmoplastic mela-noma.18 We have seen such cases misdiagnosed as leio-myosarcoma or myofibroblastic lesions such as nodularfasciitis or fibromatosis, based on this staining. An S100immunostain is invaluable here to highlight the neoplasticmelanocytes.

Angiosarcoma

Cytokeratin Expression. Approximately 10% to 30%of angiosarcomas express cytokeratins (low molecularweight only).66–70 Failure to appreciate this may result inan erroneous diagnosis of carcinoma. Technically this isnot considered anomalous expression, as normal endothe-lial cells also express cytokeratins at low levels.

CD68 Expression. As in melanoma (see previous), an-giosarcomas may express CD68, particularly when theyhave granular cell features (so-called granular cell angio-sarcoma).

Leiomyosarcoma

Cytokeratin Expression. A minority of leiomyosarco-mas (�25%) will contain cytokeratin-positive cells, some-times numerous. Cytokeratin-positive smooth muscle tu-mors express only low-molecular-weight isoforms, unlikesarcomatoid SCCs.71–73 Application of a panel of immu-nostains, including those to smooth muscle actin and des-min, should allow for confident separation of these 2 en-tities.

AFX/Superficial Malignant Fibrous Histiocytoma

Diagnostic Criteria. Atypical fibroxanthoma typicallypresents as a rapidly growing mass in a sun-exposed re-gion of an older adult. Histologically, most AFXs are un-differentiated, pleomorphic spindle cells tumors, althoughrelatively monomorphic variants do exist. The diagnosisof AFX should be reserved for small (�1–1.5 cm) lesions


Figure 6. Conventional (clear cell) renal cell carcinoma, showing artifactual strong ‘‘membranous expression’’ of von Willebrand factor (vWF).This common artifact is because of the presence of circulating vWF in the serum and may be extremely difficult to distinguish from true vWFexpression (original magnification �100).

Figure 7. Apparent smooth muscle actin expression in desmoplastic/sarcomatoid malignant melanoma. Careful inspection shows the smoothmuscle actin–positive cells to represent entrapped myofibroblasts (smooth muscle actin, 1A4, original magnification �200).

Figure 8. Sarcomatoid squamous cell carcinoma, negative for cytokeratins (AE1/AE3, original magnification �100).

Figure 9. Anomalous desmin expression in malignant melanoma (original magnification �200).

Figure 10. A, Nodular fasciitis (hematoxylin-eosin, original magnification �100). B, Intense expression of CD68 (original magnification �100).Partially on the basis of this finding, this case was submitted in consultation with a suggested diagnosis of ‘‘malignant fibrous histiocytoma.’’ CD68expression is very common in a variety of neoplasms and does not imply histiocytic or ‘‘fibrohistiocytic’’ differentiation.


that are confined to the dermis and that are completelyvisualized.74

Immunohistochemistry. Atypical fibroxanthoma is ahistologic and immunohistochemical diagnosis of exclu-sion. There are no markers or combinations of markersthat establish the diagnosis of AFX/malignant fibrous his-tiocytoma. The lesional cells of AFX must be negative forcytokeratins and S100 protein; a small amount of actinexpression, indicative of myofibroblastic differentiation, isacceptable. Atypical fibroxanthoma commonly containsS100 protein–positive Langerhans cells, CD31-positive en-dothelial cells and macrophages, and factor XIIIa–positivedendritic cells, and these must be rigorously identified asnonlesional cells. CD68 expression does not support orexclude the diagnosis of AFX. CD99 (MIC2 glycoprotein,O13, HBA71) expression is not at all specific for the di-agnosis of AFX and should have no role in this diagno-sis.75

References1. Taylor CR. Immunomicroscopy: A Diagnostic Tool for the Surgical Pathol-

ogist. Philadelphia, Pa: Saunders; 1986:xviii, 452, [2] of plates.2. Reis-Filho JS, Simpson PT, Martins A, Preto A, Gartner F, Schmitt FC. Dis-

tribution of p63, cytokeratins 5/6 and cytokeratin 14 in 51 normal and 400 neo-plastic human tissue samples using TARP-4 multi-tumor tissue microarray. Vir-chows Arch. 2003;443:122–132.

3. Lin L, Skacel M, Sigel JE, et al. Epithelioid sarcoma: an immunohistochem-ical analysis evaluating the utility of cytokeratin 5/6 in distinguishing superficialepithelioid sarcoma from spindled squamous cell carcinoma. J Cutan Pathol.2003;30:114–117.

4. Sigel JE, Skacel M, Bergfeld WF, House NS, Rabkin MS, Goldblum JR. Theutility of cytokeratin 5/6 in the recognition of cutaneous spindle cell squamouscell carcinoma. J Cutan Pathol. 2001;28:520–524.

5. Akhtar M, Tulbah A, Kardar AH, Ali MA. Sarcomatoid renal cell carcinoma:the chromophobe connection. Am J Surg Pathol. 1997;21:1188–1195.

6. Eckert F, de Viragh PA, Schmid U. Coexpression of cytokeratin and vimentinintermediate filaments in benign and malignant sweat gland tumors. J CutanPathol. 1994;21:140–150.

7. Lopez-Beltran A, Escudero AL, Cavazzana AO, Spagnoli LG, Vicioso-RecioL. Sarcomatoid transitional cell carcinoma of the renal pelvis: a report of fivecases with clinical, pathological, immunohistochemical and DNA ploidy analysis.Pathol Res Pract. 1996;192:1218–1224.

8. Meis JM, Ordonez NG, Gallager HS. Sarcomatoid carcinoma of the breast:an immunohistochemical study of six cases. Virchows Arch A Pathol Anat His-topathol. 1987;410:415–421.

9. Skalli O, Gabbiani G, Babai F, Seemayer TA, Pizzolato G, Schurch W. In-termediate filament proteins and actin isoforms as markers for soft tissue tumordifferentiation and origin, II: rhabdomyosarcomas. Am J Pathol. 1988;130:515–531.

10. Azumi N, Battifora H. The distribution of vimentin and keratin in epithelialand nonepithelial neoplasms: a comprehensive immunohistochemical study onformalin- and alcohol-fixed tumors. Am J Clin Pathol. 1987;88:286–296.

11. Nakajima T, Watanabe S, Sato Y, Kameya T, Shimosato Y. Immunohisto-chemical demonstration of S100 protein in human malignant melanoma and pig-mented nevi. Gann. 1981;72:335–336.

12. Nakajima T, Watanabe S, Sato Y, Kameya T, Shimosato Y, Ishihara K. Im-munohistochemical demonstration of S100 protein in malignant melanoma andpigmented nevus, and its diagnostic application. Cancer. 1982;50:912–918.

13. Weiss SW, Langloss JM, Enzinger FM. Value of S-100 protein in the diag-nosis of soft tissue tumors with particular reference to benign and malignantSchwann cell tumors. Lab Invest. 1983;49:299–308.

14. Gown AM, Vogel AM, Hoak D, Gough F, McNutt MA. Monoclonal anti-bodies specific for melanocytic tumors distinguish subpopulations of melano-cytes. Am J Pathol. 1986;123:195–203.

15. Pea M, Martignoni G, Zamboni G, Bonetti F. Perivascular epithelioid cell[letter; comment]. Am J Surg Pathol. 1996;20:1149–1153.

16. Pea M, Martignoni G, Bonetti F, et al. Tumors characterized by the pres-ence of HMB45-positive perivascular epithelioid cell (PEC)—a novel entity insurgical pathology. Electron J Pathol Histol. 1997;3:28–40.

17. Folpe AL, Mentzel T, Lehr HA, Fisher C, Balzer BL, Weiss SW. Perivascularepithelioid cell neoplasms of soft tissue and gynecologic origin: a clinicopatho-logic study of 26 cases and review of the literature. Am J Surg Pathol. 2005;29:1558–1575.

18. Longacre TA, Egbert BM, Rouse RV. Desmoplastic and spindle-cell malig-nant melanoma: an immunohistochemical study. Am J Surg Pathol. 1996;20:1489–1500.

19. Kaufmann O, Koch S, Burghardt J, Audring H, Dietel M. Tyrosinase,melan-A, and KBA62 as markers for the immunohistochemical identification of

metastatic amelanotic melanomas on paraffin sections. Mod Pathol. 1998;11:740–746.

20. Busam KJ, Chen YT, Old LJ, et al. Expression of melan-A (MART1) in benignmelanocytic nevi and primary cutaneous malignant melanoma. Am J Surg Pathol.1998;22:976–982.

21. Busam KJ, Jungbluth AA. Melan-A, a new melanocytic differentiationmarker. Adv Anat Pathol. 1999;6:12–18.

22. Busam KJ, Iversen K, Coplan KA, et al. Immunoreactivity for A103, anantibody to melan-A (Mart-1), in adrenocortical and other steroid tumors [seecomments]. Am J Surg Pathol. 1998;22:57–63.

23. Jungbluth AA, Busam KJ, Gerald WL, et al. A103: an anti-melan-a mono-clonal antibody for the detection of malignant melanoma in paraffin-embeddedtissues [see comments]. Am J Surg Pathol. 1998;22:595–602.

24. Koch MB, Shih IM, Weiss SW, Folpe AL. Microphthalmia transcription fac-tor and melanoma cell adhesion molecule expression distinguish desmoplastic/spindle cell melanoma from morphologic mimics. Am J Surg Pathol. 2001;25:58–64.

25. Zavala-Pompa A, Folpe AL, Jimenez RE, et al. Immunohistochemical studyof microphthalmia transcription factor and tyrosinase in angiomyolipoma of thekidney, renal cell carcinoma, and renal and retroperitoneal sarcomas: compara-tive evaluation with traditional diagnostic markers. Am J Surg Pathol. 2001;25:65–70.

26. Hofbauer GF, Kamarashev J, Geertsen R, Boni R, Dummer R. Tyrosinaseimmunoreactivity in formalin-fixed, paraffin-embedded primary and metastaticmelanoma: frequency and distribution. J Cutan Pathol. 1998;25:204–209.

27. De Young BR, Frierson HF Jr, Ly MN, Smith D, Swanson PE. CD31 im-munoreactivity in carcinomas and mesotheliomas. Am J Clin Pathol. 1998;110:374–377.

28. Miettinen M, Lindenmayer AE, Chaubal A. Endothelial cell markers CD31,CD34, and BNH9 antibody to H- and Y-antigens—evaluation of their specificityand sensitivity in the diagnosis of vascular tumors and comparison with von Wil-lebrand factor. Mod Pathol. 1994;7:82–90.

29. Rizzo M, SivaSai KS, Smith MA, et al. Increased expression of inflamma-tory cytokines and adhesion molecules by alveolar macrophages of human lungallograft recipients with acute rejection: decline with resolution of rejection. JHeart Lung Transplant. 2000;19:858–865.

30. McKenney JK, Weiss SW, Folpe AL. CD31 expression in intratumoral mac-rophages: a potential diagnostic pitfall. Am J Surg Pathol. 2001;25:1167–1173.

31. van de Rijn M, Rouse R. CD34: a review. Appl Immunohistochem. 1994;2:71–80.

32. Folpe AL, Veikkola T, Valtola R, Weiss SW. Vascular endothelial growthfactor receptor-3 (VEGFR-3): a marker of vascular tumors with presumed lym-phatic differentiation, including Kaposi’s sarcoma, kaposiform and Dabska-typehemangioendotheliomas, and a subset of angiosarcomas. Mod Pathol. 2000;13:180–185.

33. Hammock L, Reisenauer A, Wang W, Cohen C, Birdsong G, Folpe AL.Latency-associated nuclear antigen expression and human herpesvirus-8 poly-merase chain reaction in the evaluation of Kaposi sarcoma and other vasculartumors in HIV-positive patients. Mod Pathol. 2005;18:463–468.

34. Miettinen M, Fanburg-Smith JC, Virolainen M, Shmookler BM, Fetsch JF.Epithelioid sarcoma: an immunohistochemical analysis of 112 classical and var-iant cases and a discussion of the differential diagnosis. Hum Pathol. 1999;30:934–942.

35. Sirgi KE, Wick MR, Swanson PE. B72.3 and CD34 immunoreactivity inmalignant epithelioid soft tissue tumors: adjuncts in the recognition of endothelialneoplasms. Am J Surg Pathol. 1993;17:179–185.

36. Laskin WB, Miettinen M. Epithelioid sarcoma: new insights based on anextended immunohistochemical analysis. Arch Pathol Lab Med. 2003;127:1161–1168.

37. Folpe AL, Chand EM, Goldblum JR, Weiss SW. Expression of Fli-1, a nu-clear transcription factor, distinguishes vascular neoplasms from potential mimics.Am J Surg Pathol. 2001;25:1061–1066.

38. Rossi S, Orvieto E, Furlanetto A, Laurino L, Ninfo V, Dei Tos AP. Utility ofthe immunohistochemical detection of FLI-1 expression in round cell and vas-cular neoplasm using a monoclonal antibody. Mod Pathol. 2004;17:547–552.

39. Burgdorf WH, Mukai K, Rosai J. Immunohistochemical identification offactor VIII-related antigen in endothelial cells of cutaneous lesions of allegedvascular nature. Am J Clin Pathol. 1981;75:167–171.

40. Lessard JL. Two monoclonal antibodies to actin: one muscle selective andone generally reactive. Cell Motil Cytoskeleton. 1988;10:349–362.

41. Skalli O, Ropraz P, Trzeciak A, Benzonana G, Gillessen D, Gabbiani G. Amonoclonal antibody against alpha-smooth muscle actin: a new probe for smoothmuscle differentiation. J Cell Biol. 1986;103:2787–2796.

42. Tsukada T, McNutt MA, Ross R, Gown AM. HHF35, a muscle actin-spe-cific monoclonal antibody, II: reactivity in normal, reactive, and neoplastic hu-man tissues. Am J Pathol. 1987;127:389–402.

43. Tsukada T, Tippens D, Gordon D, Ross R, Gown AM. HHF35, a muscle-actin-specific monoclonal antibody. I. Immunocytochemical and biochemicalcharacterization. Am J Pathol. 1987;126:51–60.

44. Kindblom LG, Seidal T, Karlsson K. Immuno-histochemical localization ofmyoglobin in human muscle tissue and embryonal and alveolar rhabdomyosar-coma. Acta Pathol Microbiol Immunol Scand A Pathol. 1982;90:167–174.

45. Andriko JW, Kaldjian EP, Tsokos M, Abbondanzo SL, Jaffe ES. Reticulumcell neoplasms of lymph nodes: a clinicopathologic study of 11 cases with rec-


ognition of a new subtype derived from fibroblastic reticular cells. Am J SurgPathol. 1998;22:1048–1058.

46. Cho J, Gong G, Choe G, Yu E, Lee I. Extrafollicular reticulum cells inpathologic lymph nodes. J Korean Med Sci. 1994;9:9–15.

47. Van Muijen GN, Ruiter DJ, Warnaar SO. Coexpression of intermediatefilament polypeptides in human fetal and adult tissues. Lab Invest. 1987;57:359–369.

48. Franquemont DW, Frierson HF Jr, Mills SE. An immunohistochemical studyof normal endometrial stroma and endometrial stromal neoplasms: evidence forsmooth muscle differentiation. Am J Surg Pathol. 1991;15:861–870.

49. Parham DM, Dias P, Kelly DR, Rutledge JC, Houghton P. Desmin positivityin primitive neuroectodermal tumors of childhood. Am J Surg Pathol. 1992;16:483–492.

50. Folpe AL, Goldblum JR, Rubin BP, et al. Morphologic and immunophe-notypic diversity in Ewing family tumors: a study of 66 genetically confirmedcases. Am J Surg Pathol. 2005;29:1025–1033.

51. Nikolaou I, Barbatis C, Laopodis V, Bekir S, Fletcher CD. Intra-abdominaldesmoplastic small-cell tumours with divergent differentiation: report of two casesand review of the literature. Pathol Res Pract. 1992;188:981–988.

52. Gerald WL, Miller HK, Battifora H, Miettinen M, Silva EG, Rosai J. Intra-abdominal desmoplastic small round-cell tumor: report of 19 cases of a distinctivetype of high-grade polyphenotypic malignancy affecting young individuals [seecomments]. Am J Surg Pathol. 1991;15:499–513.

53. Ordonez NG. Desmoplastic small round cell tumor, II: an ultrastructuraland immunohistochemical study with emphasis on new immunohistochemicalmarkers. Am J Surg Pathol. 1998;22:1314–1327.

54. Sugimoto T, Ueyama H, Hosoi H, et al. Alpha-smooth-muscle actin anddesmin expressions in human neuroblastoma cell lines. Int J Cancer. 1991;48:277–283.

55. Hurlimann J. Desmin and neural marker expression in mesothelial cellsand mesotheliomas. Hum Pathol. 1994;25:753–757.

56. Folpe AL, Weiss SW, Fletcher CDM, Gown AM. Tenosynovial giant celltumors: evidence for a desmin-positive dendritic cell subpopulation. Mod Pathol.1998;11:939–944.

57. Folpe AL, Weiss SW. Ossifying fibromyxoid tumor of soft parts: a clinico-pathologic study of 70 cases with emphasis on atypical and malignant variants.Am J Surg Pathol. 2003;27:421–431.

58. Fanburg-Smith JC, Miettinen M. Angiomatoid ‘‘malignant’’ fibrous histio-cytoma: a clinicopathologic study of 158 cases and further exploration of themyoid phenotype. Hum Pathol. 1999;30:1336–1343.

59. Hasegawa T, Seki K, Ono K, Hirohashi S. Angiomatoid (malignant) fibroushistiocytoma: a peculiar low-grade tumor showing immunophenotypic hetero-geneity and ultrastructural variations. Pathol Int. 2000;50:731–738.

60. Zarbo RJ, Gown AM, Nagle RB, Visscher DW, Crissman JD. Anomalouscytokeratin expression in malignant melanoma: one- and two-dimensional west-

ern blot analysis and immunohistochemical survey of 100 melanomas. ModPathol. 1990;3:494–501.

61. Holness CL, Simmons DL. Molecular cloning of CD68, a human macro-phage marker related to lysosomal glycoproteins. Blood. 1993;81:1607–1613.

62. Cassidy M, Loftus B, Whelan A, et al. KP-1: not a specific marker: stainingof 137 sarcomas, 48 lymphomas, 28 carcinomas, 7 malignant melanomas and 8cystosarcoma phyllodes. Virchows Arch. 1994;424:635–640.

63. Smith ME, Costa MJ, Weiss SW. Evaluation of CD68 and other histiocyticantigens in angiomatoid malignant fibrous histiocytoma [see comments]. Am JSurg Pathol. 1991;15:757–763.

64. Dei Tos AP, Doglioni C, Laurino L, Fletcher CD. KP1 (CD68) expression inbenign neural tumours: further evidence of its low specificity as a histiocytic/myeloid marker. Histopathology. 1993;23:185–187.

65. Gloghini A, Rizzo A, Zanette I, et al. KP1/CD68 expression in malignantneoplasms including lymphomas, sarcomas, and carcinomas. Am J Clin Pathol.1995;103:425–431.

66. Ben-Izhak O, Auslander L, Rabinson S, Lichtig C, Sternberg A. Epithelioidangiosarcoma of the adrenal gland with cytokeratin expression: report of a casewith accompanying mesenteric fibromatosis. Cancer. 1992;69:1808–1812.

67. Gray MH, Rosenberg AE, Dickersin GR, Bhan AK. Cytokeratin expressionin epithelioid vascular neoplasms. Hum Pathol. 1990;21:212–217.

68. O’Connell JX, Kattapuram SV, Mankin HJ, Bhan AK, Rosenberg AE. Epi-thelioid hemangioma of bone: a tumor often mistaken for low-grade angiosar-coma or malignant hemangioendothelioma [see comments]. Am J Surg Pathol.1993;17:610–617.

69. van Haelst UJ, Pruszczynski M, ten Cate LN, Mravunac M. Ultrastructuraland immunohistochemical study of epithelioid hemangioendothelioma of bone:coexpression of epithelial and endothelial markers. Ultrastruct Pathol. 1990;14:141–149.

70. Wenig BM, Abbondanzo SL, Heffess CS. Epithelioid angiosarcoma of theadrenal glands: a clinicopathologic study of nine cases with a discussion of theimplications of finding ‘‘epithelial-specific’’ markers. Am J Surg Pathol. 1994;18:62–73.

71. Miettinen M. Immunoreactivity for cytokeratin and epithelial membraneantigen in leiomyosarcoma. Arch Pathol Lab Med. 1988;112:637–640.

72. Ramaekers FC, Pruszczynski M, Smedts F. Cytokeratins in smooth musclecells and smooth muscle tumours. Histopathology. 1988;12:558–561.

73. Tauchi K, Tsutsumi Y, Yoshimura S, Watanabe K. Immunohistochemical andimmunoblotting detection of cytokeratin in smooth muscle tumors. Acta PatholJpn. 1990;40:574–580.

74. Weiss SW, Goldblum JR. Enzinger and Weiss’s Soft Tissue Tumors. 4th ed.St Louis, Mo: Mosby; 2001:xiv, 1622.

75. Stevenson A, Chatten J, Bertoni F, Miettinen M. CD99 (p30/32MIC2) Neu-roectodermal/Ewing’s sarcoma antigen as an immunohistochemical marker: re-view of more than 600 tumors and the literature experience. Appl Immunohis-tochem. 1994;2:231–240.

best practices in diagnostic immunohistochemistry - ipath-network

Documents