A case against the specificity of "cardiac" troponin-T

Table 1 Summary of laboratory findings

Values on Values 12 hours Reference rangeTest admission after admission (SI units)

WBC 18.9 29.0 4.8-10.8 x109/4Haemoglobin 10.1 9.5 14-17 g/dlPlatelets 238 175 150-450 x109/4Sodium 133 135 134-144 mmol/lAST 101 342 8-37 U/iALT 75 248 0-39 U/iPotassium 4.6 4.7 3.7-4.9 mmol/lCreatinine 105 106 53-106 smol/lBlood urea nitrogen 8.6 8.9 2.9-8.2 mmol/lTotal bilirubin 46.2 42.8 2-21 jsmol/lLDH 1118 1215 102-228 U/iCK 1121 660 0-183 U/iTroponin-T 5.5 5.4 0-0.1 ng/l

WBC = white blood cells; AST = aspartate aminotransferase; ALT = alanine aminotransferase;LDH = lactate dehydrogenase.

renal disease.3 4 Ours is the first report to indi-cate that such spurious rises may also be seenin the presence of intact renal function.As there were no clinical electrocardiograph

tracings in our patient suggestive of myocardialinfarction, we believe that the increasedconcentrations of troponin-T were an errone-ous indicator of myocardial infarction. Thereare two possible reasons for the rise introponin-T concentrations: firstly, as a result ofthe existence of minor myocardial injuriesundetectable in the ECG.' Lack of an increasein the CK-MB fraction could be explained bythe fact that cardiac troponin-T is known tohave a higher sensitivity and therefore a lowerdiagnostic efficiency than CK-MB in the diag-nosis of myocardial infarction.' A second,alternative explanation, and one that wefavour, is that the raised cardiac troponin-Tconcentration is a consequence of the multipledisorders, septicaemia and widespread malig-nancy present in our patient. There is unpub-lished information from retrospective reviewsof medical records to indicate that patientspositive for troponin-T and negative for

CK-MB have multiple illnesses. As the pres-ence of mnultiple illnesses is often a complica-tion of and not an uncommon clinical phe-nomenon in widespread malignant disease, webelieve that this anecdotal information may beof use. It is possible that follow up values afterthe acute period may have provided anadditional insight into the cause of theincreases observed in the present patient.However, this information was not available asthe patient died two days after admission tohospital. Permission to perform a necropsy wasnot granted and the cause of death waspresumed to be overwhelming sepsis.The CK-BB isoenzyme has been proposed

as a tumour marker as it indicates the presenceof solid organ malignancies.5 There is also anisolated case report of a patient with MDS withan elevated CK-BB fraction.6 Thus, theincreased CK-BB fraction observed in ourpatient may be attributable to either or all ofthe three malignancies.

Increases in cardiac troponin-T concentra-tions in the absence of diagnostic ECGchanges and an increase in the CK-MBfraction support the exclusion of myocardialinfarction as a clinical possibility and re-affirmthe use of CK-MB as a diagnostic tool in thiscondition.

1 Bhayana V, Henderson AR. Biochemical markers ofmyocardial damage. Clin Biochem 1995;28: 1-29.

2 Kobayashi S, Tanaka M, Tamura N, Hashimoto H, HiroseS. Serum cardiac troponin T in polymyositis/derma-tomyositis. Lancet 1992;340:726.

3 Wu AH, Valdes R Jr, Apple FS, Gornet T, Stone MA,Mayfield-Stokes S, et al. Cardiac troponin-T immunoassayfor diagnosis of acute myocardial infarction. Clin Chem1994;40:900-7.

4 Croitoru M, Taegtmeyer H. Spurious rises in troponin-T inend-stage renal disease. Lancet 1995;346:974.

5 Abbott B, Lott JA. Reactivation of serum creatine kinaseisoenzyme BB in patients with malignancies. Clin Chem1984;30:1861-3

6 Crook M, Williams A, Sankarlingam A, Tutt P. Raised con-centration of plasma creatine kinase BB isoenzyme inmyelodysplasia. J Clin Pathol 1994;47:552-3.

I Clin Pathol 1996;49:767-770

High temperature antigen retrieval and loss ofnuclear morphology: a comparison of microwaveand autoclave techniques

Immunocytochemistryand MolecularPathology Unit,Department ofPathology, UniversityHospital ofWales,Heath Park,Cardiff CF4 4XN

Correspondence to:Dr NCA Hunt,Clinical Lecturer,Nuffield Department ofPathology and Microbiology,John Radcliffe Hospital,Headington,Oxford OX3 9DU.

Accepted for publication4June 1996

N C A Hunt, R Attanoos, B Jasani

AbstractThe use of high temperature antigenretrieval methods has been of majorimportance in increasing the diagnosticutility of immunocytochemistry. How-ever, these techniques are not withouttheir problems and in this report attentionis drawn to a loss of nuclear morphologi-cal detail, including mitotic figures, fol-lowing microwave antigen retrieval. This

was not seen with an equivalent autoclavetechnique. This phenomenon was quanti-fied using image analysis in a group of Bcell lymphomas stained with the antibodyL26. Loss of nuclear morphological detailmay lead to difficulty in identifying cellsaccurately, which is important in thediagnostic setting-for example, whentrying to distinguish a malignant lym-phoid infiltrate within a mixed cell popu-

767

on January 22, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.49.9.767 on 1 S

eptember 1996. D

ownloaded from

http://jcp.bmj.com/

Hunt, Attanoos, J7asani

lation. In such cases it would clearly bewise to consider the use ofalternative hightemperature retrieval methods and accepttheir slightly lower staining enhancementcapability compared with the microwavetechnique.( Clin Pathol 1996;49:767-770)

Keywords: antigen retrieval, artefacts, immunocyto-chemistry.

The masking of tissue antigens in routinelyprocessed histological specimens has been amajor obstacle to the application of immuno-cytochemistry to archival material. A variety oftechniques have been successfully developed toreveal such antigens, including enzymatic andhigh temperature methods.' These techniqueshave been important in overcoming the obvi-ous limitations imposed by the dependence onfrozen tissue and their development haspermitted the diagnostic utility of immunocy-tochemistry to increase dramatically. However,these methods are not without disadvantages.Here, we present histomorphometric data froma study comparing our standard autoclave andmicrowave pretreatment protocols in a groupof B cell lymphomas stained with the L26 anti-body (anti-CD20 B cell lineage marker). Ourresults demonstrate a potentially significantdifference between these methods with impli-cations for diagnostic work.

MethodsSix cases of non-Hodgkin's B cell lymphomawere identified retrospectively from the ar-chives of the Department of Histopathology,University Hospital of Wales. All cases were ofdiffuse, centroblastic/centrocytic type. Sec-tions, 4-5 gm thick, were cut from eachtumour and were untreated or pretreated in anautoclave (0.01 M sodium citrate buffer, pH6.5; 15 minutes at 120°C; 15 psi),2 orpretreated in a microwave oven (0.01 Msodium citrate buffer, pH 6.5; 15 minutes at800 W (Matsui TC, Harrow, UK)).2

Following pretreament, all sections werestained with L26 (Dako, High Wycombe, UK)(diluted 1 in 40) and incubated at 40Covernight. Sections were then stained using asimple, indirect immunoperoxidase method.All sections then underwent histomorphomet-ric analysis of mean nuclear area and mitoticindex. All samples were examined blind.To assess the apparent difference in nuclear

size observed between the control, autoclaveand microwave treated sections, the cross-

Table 1 Comparison of mean nuclear area of centrocyticlcentroblastic cells and mitoticindex with and without pretreatment by autoclave or microwave antigen retrieval methods.Results expressed as mean (SD) for area

Control Autoclave Microwave

Case Area (pm2) Mitotic Area (,um2) Mitotic Area (pum2) Mitoticnumber count count count

1 19.03 (4.42) 11 20.87 (5.75) 18 32.05 (8.70) 42 18.25 (3.29) 8 17.22 (4.46) 8 36.17 (13.29) 33 20.56 (4.65) 30 23.24 (5.08) 44 31.81 (5.89) 74 17.29 (3.49) 37 16.43 (2.44) 51 30.55 (6.57) 175 16.43 (3.65) 33 18.22 (4.38) 46 31.97 (5.69) 176 17.36 (3.65) 17 20.72 (4.37) 15 30.12 (6.49) 2

sectional area of 300 randomly selected nucleiwas measured from each section using an IBASimage analyser (Kontron, Ecking, Germany),with a Videoplan software package, coupled toan Olympus (London, UK) BH2 microscope.The data were gathered by interactively defin-ing the nuclear circumference using computergenerated graphics superimposed over a live,colour television image. From these data themean cross-sectional nuclear area for each sec-tion was calculated. Mitotic indices were thendetermined for each section using a slightmodification of the methodology described byvan Diest et al.3 In each case 50 high powerfields were counted using an Olympus BH2microscope (x40 objective; field diameter 500gm).The histomorphometric data were analysed

by comparison of means using standard errors.The mitotic indices were analysed for signifi-cant differences using a Wilcoxon rank sumtest for non-parametric data.

ResultsThe results of the histomorphometric analysisof nuclear area are presented in table 1. Thereseems to be a constant difference between thesections pretreated with microwaves comparedwith the control and the autoclave pretreatedsections in terms of nuclear area, which wasconsistently higher in the former. This reachesstatistical significance in each case with the nullhypothesis being rejected at p < 0.001 (table1). Analysis of the mitotic indices for sectionsfrom each case showed a consistent differencebetween the sections pretreated with micro-waves compared with both control and auto-clave pretreated sections. In all cases themitotic index was lower in the microwavetreated sections.

In qualitative terms a number of differenceswere noted between the variously treatedsections. The microwave pretreated sectionsshowed a clear increase in intensity of stainingwhen compared with controls. The autoclavepretreated sections also showed an increase instaining intensity compared with the controlsections, although overall enhancement wasslightly less than in the microwave pretreatedgroup. However, there was also a clear, qualita-tive difference between the sections in terms ofcellular morphology with the control sectionshaving the best morphology and those pre-treated with microwaves the worst. Theautoclave pretreated group had lost much lessmorphological detail than the microwavepretreated group (fig 1) and compared favour-ably with the control group. Specifically, thecells in the microwave pretreated specimensappeared rather swollen with associated loss ofnuclear detail and often complete blurring ofmitotic figures. These observations are, webelieve, the qualitative correlates of the quanti-tative data presented above.

DiscussionImmunocytochemistry has become an invalu-able tool in routine diagnostic histopathologyas well as an important research technique.The development of antigen retrieval tech-

768

on January 22, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.49.9.767 on 1 S

eptember 1996. D

ownloaded from

http://jcp.bmj.com/

High temperature antigen retrieval and loss of nuclear morphology

Figure 1 Photomicrographs showing (A) an autoclave pretreated section and (B) amicrowave pretreated section oflymphoma (original magnification x200).

niques has been instrumental in this process asit has permitted the application of antibodies toroutinely processed tissue samples and has, inmany cases, abrogated the need for frozenmaterial. Such antigen retrieval can beachieved using both enzymatic and, morerecently, high temperature techniques.' Theformer techniques, whilst offering the potentialfor using formalin fixed, paraffin wax embed-ded material for immunocytochemistry, weredifficult to standardise and do not offer signifi-cant staining enhancement with many antibod-ies.4 The latter techniques involve superheatingtissue sections with either a microwave oven,5autoclave6 or pressure cooker.7 There are anumber of possible mechanisms likely to beresponsible for antigen retrieval using thesetechniques, including breakage of formalde-hyde induced protein cross linkages, denatura-tion of proteins to reveal previously maskedepitopes and unmasking of epitopes by re-moval of calcium ions.4 To date, mostexperience has been gained with microwavetechniques, perhaps because workers alreadyhave experience with this technology. How-ever, microwaves have certain drawbacks in-cluding the fact that they work best with a

small number of cases, problems with evapora-tion of buffer, inconsistent staining due to localsuperheating, and their relatively frequentmechanical failures due to overuse.7 In addi-tion, we have also noted in our own work on Bcell lymphomas that microwave pretreatmentseems to impair cellular morphology signifi-cantly. A diminution in morphology followingmicrowave pretreatment has also been notedby other workers,5 although they do not seemto have assessed this phenomenon in anydetail. Because of the implications of this indiagnostic work, especially when only limitedmaterial is available and enhancement isneeded, we decided to undertake a formalstudy of this phenomenon as presented in thisshort report.The cellular distortion with microwave

pretreatment of tissue sections is reflected inthe increase in mean nuclear area and theapparent decrease in the mitotic index. While itis possible to speculate on the underlyingchanges responsible for these observations, itwould be rather more difficult to delineatethese in exact detail. However, one possibilityis that microwaves cause some structural dam-age to intracellular macromolecules which, inturn, leads to an increase in the number ofosmotically active moieties within the nuclearcompartment, thus attracting water and caus-ing nuclear swelling. An intrinsic property ofmicrowaves is proposed as superheating alonecannot be the answer as evinced by the muchbetter morphological preservation using anautoclave. Such changes would certainly ex-plain the increase in mean nuclear areaobserved and furthermore, this swelling wouldhave the effect of blurring mitotic figures andmaking them less easy to distinguish accuratelyfrom apoptotic bodies or lymphocytes and thusless likely to be counted. Given that this studywas undertaken using the optimised protocolsused in our laboratory for diagnostic work,these observed changes may have importantconsequences for selection of retrieval methodsin the diagnostic setting when it becomesimportant to identify accurately the cells whichhave been stained. Both mitotic figures andmorphology may be of importance in achievingthis-for example, when trying to distinguish amalignant lymphoid infiltrate within a mixedcell population. In such cases it would clearlybe wise to consider the use of alternative hightemperature retrieval methods and accept theirslightly lower staining enhancement capabilitycompared with the microwave technique.

1 Cattoreti G, Pileri S, Parravicini C, Becker MHG, Poggi S,Bifulco C, Key G, et al. Antigen unmasking on formalin-fixed, paraffin-embedded tissue sections. J Pathol 1993;171:83-98.

2 Banfalvi A, Navabi H, Bier B, Bocker W, Jasani B, Schmid K.Wet autoclave pretreatment for antigen retrieval in diagnosticimmunohistochemistry. J Pathol 1994; 174:223-8.

3 van Diest PJ, Baak JPA, Matze-Cok P, Wisse-BrekelmansECM, van Galen CM, Kurver PHJ, et al. Reproducibilityof mitosis counting in 2469 breast cancer specimens:Results from the multicenter morphometric mammarycarcinoma project. Hum Pathol 1992;23:603-7.

4 Shi SR, Gu J, Kalra KL, Chen T, Cote RJ, Taylor CR. Anti-gen retrieval technique: A novel approach to immunocyto-chemistry on routinely processed tissue sections. CellVision 1995;2:6-22.

5 Shi S-R, Key ME, Kalra KL. Antigen retrieval in formalin-fixed, paraffin-embedded tissues: an enhancement method

769

on January 22, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.49.9.767 on 1 S

eptember 1996. D

ownloaded from

http://jcp.bmj.com/

Hunt, Attanoos, Jasani

for immunohistochemical staining based on microwaveoven heating of tissue sections. J Histochem Cytochem1991;39:741-8.

6 Navabi H, Douglas-Jones A, Bankfalvi A, Schmid K, SuzukiK, Katoh R, et al. Wet autoclave pretreatment: a reliablealternative to microwave technique for antigen retrieval[abstract]. J Pathol 1994;172:50A.

7 Norton AJ, Jordan S, Yeomans P. Brief, high-temperatureheat denaturation (pressure cooking): a simple andeffective method of antigen retrieval for routinely proc-essed tissues. J Pathol 1994;173:371-9.

8 Morgan JM, Navabi H, Schmid K, Jasani B. Possible role oftissue-bound calcium ions in citrate-mediated high-temperature antigen retrieval. J Pathol 1994;174:301-7.

JT Clin Pathol 1996;49:770-772

Focal rhabdomyosarcomatous differentiation inprimary liposarcoma

J H Shanks, S S Banerjee, B P Eyden

AbstractA unique case of primary myxoid liposar-coma of the thigh, in which focal pleomor-phic areas were present containingrhabdomyoblasts, is described. Focal rhab-domyosarcoma in liposarcoma has onlyrarely been reported previously and onlyin dedifferentiated liposarcomas of theretroperitoneum. All but one have beenrecurrences with rhabdomyoblasts beingabsent in the primary liposarcoma. Asrhabdomyoblasts were only focallypresent, the present case is regarded asliposarcoma with focal divergent rhab-domyoblastic differentiation rather thanmalignant mesenchymoma.(7 Clin Pathol 1996;49:770-772)

Keywords: liposarcoma, rhabdomyosarcoma, malignantmesenchymoma.

Liposarcoma may contain benign or malignantheterologous mesenchymal elements. Benigncartilage,' smooth muscle,2 or rarely bone' havebeen described in well differentiated lipo-sarcoma/atypical lipomatous tumours or inmyxoid liposarcoma. Until now, malignantheterologous elements have been found only indedifferentiated liposarcoma.' "Malignant myogenic elements found in

liposarcoma have included leiomyosarcoma,rhabdomyosarcoma, or both. Other heterolo-gous elements including angiosarcoma or oste-oid have also been described rarely.' Divergentrhabdomyosarcomatous differentiation hasbeen reported previously in only seven cases ofliposarcoma.' 3-5 All were dedifferentiated li-posarcomas arising in the retroperitoneum andin all but one case rhabdomyosarcoma wasfound only in tumour recurrences. There isonly a single previous case of rhabdomyosarco-matous differentiation in de novo liposarcoma,a retroperitoneal dedifferentiated tumour.5

Here, we describe a unique case of divergentrhabdomyosarcomatous differentiation in a denovo liposarcoma, which was of combinedmyxoid and pleomorphic subtypes arising inthe thigh. Immunohistochemical and ul-trastructural evidence is provided to supportour diagnosis.

Case reportA 75 year old man presented with a six weekhistory of a swelling in the right thigh. This wasassociated with oedema of the right leg, butthere were no systemic symptoms. There wasno evidence of any abdominal mass. On explo-ration, a mass was found to be displacing, butnot invading, the femoral vessels and nerves.The tumour was easily mobilised except on theposterior surface where it was adherent to thepectineus muscle.

Macroscopically, the tumour consisted of apartly encapsulated fatty mass measuring 12cm at its maximum dimension with a smallamount ofsurrounding soft tissue. It was partlynecrotic. Histologically, the tumour was aliposarcoma which was predominantly ofmyxoid type with a typical delicate branchingcapillary network and myxoid stroma (fig 1). Inplaces, the tumour had more cellular pleomor-phic areas. The latter were composed ofspindle cells with notable nuclear pleomor-phism and bizarre lipoblasts containing typicalscalloped nuclei and cytoplasmic lipid vacuoles(fig 1, left upper inset). Mitotic activity was fre-quent (21/10 high power fields) within thepleomorphic areas and large foci of necrosiswere seen. The tumour grade was 3 based onthe method by Trojani et al.6

Cells within the pleomorphic areas hadabundant eosinophilic cytoplasm and notice-ably hyperchromatic nuclei (fig 1, right lowerinset), many showing notable nuclear enlarge-ment and irregular nuclear contours. Pleomor-phic multinucleate tumour giant cells were alsoseen. No cross striations were visible by lightmicroscopy. Immunohistochemistry was per-formed using the streptavidin-peroxidase com-plex technique with diaminobenzidine chro-mogen substrate. The pleomorphic cellsshowed intense cytoplasmic positivity usingimmunostains for desmin (diluted 1 in 25;Dako, High Wycombe, UK). Myoglobin (di-luted 1 in 1000; Dako) and muscle specificactin (HHF 35) (diluted 1 in 40; Biomen, Fin-champstead, UK) were also positive in thesecells. A proportion of them were positive forfast myosin (diluted 1 in 600; Sigma, Poole,Dorset, UK) and sarcomeric actin (diluted 1 in100; Sigma). No tumour cells stained for

Department ofHistopathology,Christie Hospital NHSTrust, Wilmslow Road,Withington,Manchester M20 4BX

Correspondence to:Dr SS Banerjee.

Accepted for publication7 May 1996

770

on January 22, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.49.9.767 on 1 S

eptember 1996. D

ownloaded from

http://jcp.bmj.com/