9 Clin Pathol 1992;45:1094-1098
Evaluation of three techniques for differentialdiagnosis of prostatic needle biopsy specimens
R F T McMahon, L J McWilliam, S Mosley
AbstractAims: To determine whether acidic mucinstaining, lectin histochemistry using Wis-teriafloribunda agglutinin, and immuno-histochemistry using the monoclonalantibody EAB 903 are of benefit in distin-guishing between hyperplastic and carci-nomatous prostatic glandular tissue inneedle biopsy specimens.Methods: Formalin fixed, paraffin waxembedded prostatic needle biopsy speci-mens ofbenign and malignant tissue wereexamined. Alcian blue-periodic acidSchiff staining was performed on 33benign and 34 malignant cases. Wisteriafloribunda agglutinin (WFA) binding siteswere demonstrated by the avidin-biotinperoxidase (ABC) technique with andwithout neuraminidase pretreatment on34 benign cases and 32 malignant cases.EAB903 anticytokeratin antibody bindingsites were demonstrated using both anindirect immunoperoxidase (IIP) tech-nique and an avidin-biotin peroxidasecomplex method on seven benign and 31malignant cases.Results: Acidic mucin staining was foundin 17 of 34 malignant glands and wasweakly positive in five of 33 benign glands.
Division ofHistopathology,Department ofPathological Sciences,University ofManchester, StopfordBuilding, OxfordRoad, ManchesterM13 9PTR FT McMahonS MosleyWithington Hospital,ManchesterL J McWilliamCorrespondence to:Dr R FT McMahonAccepted for publication8 July 1992
WFA positivity before neuraminidase pre-treatment was present in 29 of 32 malig-nant glands and in 19 of 34 benign glands.After neuraminidase all benign andmalignant cases showed positivity. EAB903 positivity was seen in 11 of 31 malig-nant glands using the IIP technique and intwo of 31 malignant glands using the ABCtechnique. In seven benign cases there waspositivity in all glands using the IIPmethod with predominant basal cell stain-ing in three and superficial cell staining infour. In benign cases using the ABCmethod two cases were negative.Conclusions: None of the three methodsstudied showed sufficient sensitivity andspecificity to allow their recommendationfor routine diagnostic use.
( Clin Pathol 1992;45:1094-1098)
Needle biopsy, by the transrectal or perinealroutes, is an increasingly popular method ofdiagnosis in prostatic practice. Relatively smallamounts of tissue are available from thistechnique in comparison with transurethralresections or total prostatectomies, thus mag-nifying the degree of diagnostic difficulty.
A
:4..
- SX , :;.,
.)...>....
Figure 1 EAB 903 immunoperoxidase staining of benign prostatic glands using the IIP method. Note negative glands(arrow), basal cell staining in some glands (arrowhead), and superficial cell staining focally (small arrowhead).
1094
group.bmj.com on July 31, 2011 - Published by jcp.bmj.comDownloaded from
Evaluation of three techniques for differential diagnosis of prostatic needle biopsy specimens
Figure 2 Focal positivity(arrow) of malignantprostate for EAB 903 usingthe IIP method.
Various special techniques have been proposedto aid in the differential diagnosis of benignfrom malignant lesions. Carcinomas of theprostate secrete acidic mucin in 40-70% ofcases, while normal and hyperplastic prostateglands produce only neutral mucin.' In apreliminary study of lectin histochemistry,Wisteria floribunda agglutinin (WFA) seemedto be expressed only by malignant glands,probably related to sialylation, as benignglands contained binding sites for WFA afterneuraminidase digestion.2 A monoclonal anti-cytokeratin antibody, EAB 903, has beenshown by several groups to be exhibited onlyby the basal cells of benign prostatic glands,3 6while these EAB 903 reactive cells are absentin malignant prostatic glands, with the excep-tion of two reported cases.7 This study wasundertaken to determine whether any or all ofthese techniques might be of benefit in distin-guishing between hyperplastic and carcinoma-tous prostatic glandular tissue in needle biopsyspecimens.
MethodsNeedle biopsy specimens of prostate glandwere obtained from the files of the Departmentof Surgical Histology, Manchester Royal Infir-mary, and of the Department of Histopathol-ogy,Withington Hospital: 67 cases (33 benign,34 malignant) were evaluated in the mucinstudies, 66 cases were examined with WFAlectin (34 benign, 32 malignant) and 38 caseswere studied with EAB 903 (seven benign, 31malignant). The usual diagnostic criteria fordifferentiating benign from malignant prostatictissue were applied.8 Of the 31 cases evaluatedwith the monoclonal antibody, in combinationwith mucin and lectin histochemical tech-niques, Gleason grading9 showed one grade 2,
one grade 2 + 3, 12 grade 3, four grade3 + 4, one grade 3 + 5, eight grade 4, twograde 4 + 5 and two grade 5 tumours. TheBrawn modification of the MD AndersonHospital system'0 disclosed five grade 1, 16grade 2, and 10 grade 3 carcinomas.
All of the tissue had been fixed in neutralbuffered 10% formalin and routinely pro-cessed, before embedding in paraffin wax. Thesections were dewaxed and rehydrated. Acombined alcian blue (pH 2-5) and periodicacid Schiff stain after diastase pretreatment(AB-DPAS) was performed. Wisteria floribundaagglutinin binding sites were demonstrated bythe avidin-biotin peroxidase technique using alectin concentration of 1O ug/ml; neuramini-dase pretreatment to remove sialic acid resi-dues was also performed. The antibody EAB903, which was purchased from Enzo Diag-nostics, is a high molecular weight (68, 58,56-5 and 56 kilodaltons) anticytokeratinmonoclonal antibody" and binding sites wereidentified by two methods. The first was astandard indirect immunoperoxidase tech-nique. Endogenous peroxidase activity wasblocked, sections were digested with trypsin,and the primary antibody was applied for 60minutes at a dilution of 1 in 100. Thesecondary rabbit anti-mouse antibody wasapplied for 30 minutes, the reaction productwas developed in diaminobenzidine (DAB),and the sections were counterstained withHarris's haematoxylin. The avidin-biotin per-oxidase complex method was also used. Againendogenous peroxidase activity was blockedand sections were digested with trypsin. Theprimary antibody was applied at a dilution of 1in 100 for 60 minutes, followed by the second-ary sheep anti-mouse biotinylated antibody ata dilution of 1 in 100 for 30 minutes. Sectionswere then incubated in freshly made up avidin-
1 095
group.bmj.com on July 31, 2011 - Published by jcp.bmj.comDownloaded from
McMahon, McWilliam, Mosley
Figure 3 Focal positivity(arrowheads) of malignantprostate for EAB 903 usingthe ABC method.
biotin-peroxidase complex for 60 minutes. Thereaction product was developed in DAB andsections were counterstained with methylgreen.
ResultsThe results of mucin staining are shown intable 1. The sections were initially assessed forthe presence of alcianophilia (blue staining)on an absent/definitely present/weakly positivebasis. The last two grades were then placed intoa "present" category. By this method, 50%(17/34) of the malignant glands containedacidic mucin while 15% (5/33) of the benignglands stained positively for alcian blue, albeitonly weakly so in all five cases.The results of Wisteria lectin studies are
shown in table 1. Again, a simplified absent/present grading was used. Before neuramini-dase pretreatment, 90% (29/32) of malignantprostates revealed binding sites forWFA while56% (19/34) of benign glands did so, although17 of these cases were only weakly positive.After neuraminidase, all glands were uniformlypositive, both benign and malignant.The findings with EAB 903 are shown in
table 2. In benign glands all seven casesassessed by the IIP method were positive.Reactivity was noted in basal cells predom-inantly in three instances and in superficial/luminal cells predominantly in four cases. Withthe ABC method, two of the seven cases werenegative, three were only weakly positive, andonly two showed definite basal cell reactivity.In malignant prostates 11 of the 31 (36%)cases were reactive for EAB 903 by the IIPmethod and two of 31 (6%) cases were positiveby the ABC method.
Table 3 shows the results of comparing thecombination of staining patterns with theGleason grade and simplified histological dif-
ferentiation in the 31 cases where all modalitieswere used. Gleason grading is extremely diffi-cult to perform on needle biopsy material dueto the paucity of malignant tissue available andan assessment of differentiation was easier toapply in this circumstance. There did not seemto be any significant differences in the distribu-tion of either Gleason grades or degree ofdifferentiation when compared with the use ofthese three techniques, including both mod-ifications of the monoclonal antibody method,as a panel.
DiscussionProstatic needle biopsy specimens are a chal-lenge to the diagnostic histopathologist in thatrelatively small amounts of tissue are providedfrom which clinically important decisions maybe made. To improve the accuracy of diagnosis,several methods have been used, includingmucin histochemistry,' 121' lectin histochem-istry,2 16 and immunohistochemistry.37 17 18Mucin histochemistry has been used for
many years and while acidic mucin can bedemonstrated in 40-70% of malignant cases, itis rarely seen in benign hyperplastic glands. Inthis study 50% of carcinomas contained alcia-nophilic mucin, an increase from the 38% ofpositive cases previously reported on transur-ethral resection material.1 In that series nobenign glands exhibiting acidic mucin were
Table 1 Prostatic needle biopsy specimens
Positive Negative Total
Alcian blue Benign 5 28 33Malignant 17 17 34
WFA-NA Benign 19 15 34Malignant 29 3 32
WFA +NA Benign 34 0 34Malignant 32 0 32
1096
group.bmj.com on July 31, 2011 - Published by jcp.bmj.comDownloaded from
Evaluation of three techniques for differential diagnosis of prostatic needle biopsy specimens
Table 2 Prostatic needle biopsy specimens: EAB 903 staining
Method Negative Positive B > S Positive S > B
Benign IIP 0 3 4ABC 2 2 3
Negative PositiveMalignant IIP 20 11
ABC 29 2
B > S: Basal greater than superficial stainingS > B: Superficial more than basal staining
Table 3 Comparison of staining patterns with mucin, monoclonal antibody, and lectinhistochemistry with Gleason grade and MDAH grading systems
Staining pattern No of cases Gleason' MDAH"'
AB +, IIP -,ABC ,WFA + 7 2 + 3,3,3,3 + 4,4,4,5 2,2,3,3,1,2,2AB +, IIP ,ABC +,WFA + 2 3,3 + 4 2,2AB +,IIP +,ABC ,WFA- 1 4 2AB +, IIP +, ABC ,WFA + 7 3,3,3,3,3 + 4,4,5 1,3,2,2,2,3,2AB ,IIP ,ABC ,WFA - 2 2,4 + 5 1,3AB I,IP , ABC ,WFA + 9 3,3,3,3,3,3 + 4,4,4,4 + 5 2,1,2,3,2,1,3,2,2AB ,IIP +, ABC ,WFA + 3 3 + 5,4,4 3,3,3
Staining:AB: Alcian blue positive (+) or negativeIIP: MA903 positive (+) or negative ( -) by indirect immunoperoxidase methodABC: MA903 positive (+) or negative ( -) by avidin-biotin peroxidase complex methodWFA: Wisteria floribunda agglutinin positive (+) or negative ( -) before neuraminidasedigestionMDAH (MD Anderson Hospital grading sytem)
identified; the present study highlighted fiveweakly positive cases out of 33 examined.Although the staining noted was quite weakand focal, the implication of this finding is thatalcian blue staining, which has a sensitivity ofbetween 30 and 70%, is not totally specific andthus cannot be recommended in isolation as amethod to differentiate benign from malignantprostatic tissue.
Lectins are proteins or glycoproteins of non-immune, mainly plant, origin which havehighly specific binding sites for mono-or oligo-saccharides.'9 Changes in the saccharide com-position of glycoproteins have been noted inneoplastic transformation of cells of varioustissues, both on cell surfaces and in cellcytoplasms.20 In a preliminary study of pros-tatic lectin binding Wlisteria floribunda aggluti-nin (WFA) binding sites were identified only inmalignant prostatic tissue, but pretreatmentwith neuraminidase, which removes terminalsialic acid residues, revealed reactivity forWFAin both benign and malignant tissue.2 Thecurrent study again showed 90% WFA pos-itivity in malignant tissue but also showedmainly weak reactivity for WFA in 56% ofbenign glands. While demonstrating excellentsensitivity, this technique lacks specificity and
alone cannot be recommended for diagnosticuse.EAB 903 is a monoclonal antibody which
recognises cytokeratins 1, 5, 10 and 1 1, corre-sponding to molecular weights 68, 58, 56-5and 56 kilodaltons. " A summary of the reportson its use in the prostate to date is shown intable 4. Brawer et al examined 31 malignantcases and found no positive results.3 Theydescribed its localisation in benign glands tobasal cells only, including cases of basal cellhyperplasia. Chastonay et al reported theirfindings in both frozen section and paraffinwax embedded material.7The patterns of basalcell staining altered with fixation, such thatfrozen section material exhibited a continuousbasal layer reactivity while formalin fixed tissuehad an interrupted pattern. They also des-cribed the only two previous cases of positivityin malignant glands with this antibody, both inwell differentiated adenocarcinomas. Nagle etal"7 and Brawer et al'8 examined the use of asimilar, though not identical, antibody KAI inboth frozen cell culture tissue and formalinfixed material; in neither instance were thereany positive findings in malignant glands.Bostwick and Brawer used EAB 903 to evalu-ate the concept of dysplasia/prostatic intra-epithelial neoplasia (PIN) and found that therewas increasing disruption of the basal layerrelated to increasing grades of dysplasia/PIN.4Hedrick and Epstein described three patternsof basal cell staining-continuous, continuouswith focal disruption, and disrupted and apply-ing EAB 903 to a wide variety of benign andmalignant tissue, found a high incidence ofdisrupted basal cell staining in benign condi-tions and a total lack of reactivity in carcino-mas.5 They cautioned against the absolute use
of an absence of a basal cell layer in thediagnosis of carcinoma because of the dis-rupted patterns which may be seen in benignprostates. O'Malley et al described usuallycontinuous basal cell staining in normal/hyper-plastic lesions and its absence in carcinoma.6The results of the present study show that in
non-neoplastic tissue the previously describedbasal cell only staining is not a constantfinding. The results varied with the methodsused and although small numbers of cases havebeen examined, this gives rise to concern as tothe specificity of the antibody. The dilutionrecommended by the manufacturers was 1 in8000 but by a process of trial and error, theworking dilution eventually chosen was 1 in100, as also used by Hedrick and Epstein.5 As
Table 4 Review ofprevious studies
Author Fixation Technique Dilution Before treatment Cases No positive
Brawer et al 1985 FFPE ABC 1 in 2000 Pronase 31 0Chastonay et al 1986 FFPE PAP 1 in 500 Protease 25 2
FS PepsinNagle et al 1987 (KAI) FS lIP NS NS 10 0Bostwick and Brawer 1987 FFPE ABC 1st 1 in 2000 Pronase 14 0
2nd 1 in 50Brawer et al 1989 (KAI) FFPE PAP NS Pronase 37 0Hedrick and Epstein 1989 FFPE ABC 1 in 100 Pronase 58 0O'Malley et al 1990 FFPE ABC NS Pepsin 21 0
Abbreviations:FFPE: Formalin fixed, paraffin wax cmbeddcd material; FS: Frozen section material; ABC: Avidin-biotin-peroxidase complexmethod; PAP: Peroxidase anti-peroxidase method; IIP: Indirect immunoperoxidase method; NS: Not stated
1 097
group.bmj.com on July 31, 2011 - Published by jcp.bmj.comDownloaded from
McMahon, McWilliam, Mosley
suggested for use on formalin fixed, paraffinwax embedded material, pretreatment with aprotease was performed, in our case with 0-1%crude trypsin. With the IIP method, thisproduced a very high carcinoma positivity rateof 36% (11/31), particularly in view of thefindings so far in all of the studies cited of onlytwo EAB 903 reactive carcinomas out of 200examined. We also used the ABC method asmost of the other groups have done, with theexception of the KA1 studies.7 17 18 Thisreduced the positivity rate for carcinoma con-siderably, and in the benign cases, unusualpatterns of basal cell staining were seen as wellas two totally negative results. In our hands,therefore, EAB 903 has not proved a usefuladjunct to the differential diagnosis of benignfrom malignant lesions and indeed, the usealone, particularly with the ABC method, wasmore confusing than helpful.
Overall, none of the three techniques eval-uated showed sufficient sensitivity and specific-ity to allow their recommendation for routinediagnostic use, even when evaluated as a panelof markers or by comparison with Gleason orother grading systems. Of the three methods,the simple mucin stain (AB-DPAS) providedthe most useful information, although thefinding of weakly positive alcianophilic mate-rial in benign glands suggests that it should notbe used to the exclusion of anything else.Wisteria floribunda lectin seems to be unsuit-able for further application in this context. Theanticytokeratin antibody EAB 903 proved dif-ficult in its practical use and problems aroserelated to the dilution and revealing methodused. The ABC method would appear to be themore likely technique to yield diagnostic infor-mation but extreme caution in its interpreta-tion is necessary in view of the "false" positiveand particularly the false negative results seenin this small study. The "gold standard" inneedle biopsy diagnosis still remains a goodhaematoxylin and eosin stained section exam-ined by an experienced pathologist.
I Pinder SE, McMahon RFT. Mucins in prostatic carcinoma.Histopathology 1990;16:43-6.
2 Rhodes H, McMahon RFT, McWilliam U, Stoddart RW.Lectin histochemistry ofhuman prostatic gland - normal,hyperplastic and carcinomatous. J Pathol 1990;161:359A.
3 Brawer MK, Peehl DM, StameyTA, Bostwick DG. Keratinimmunoraeactivity in the benign and neoplastic humanprostate. Cancer Res 1985;45:3663-7.
4 Bostwick DG, Brawer MK. Prostatic intra-epithelial neo-plasia and early invasion in prostate cancer. Cancer1987;59:788-94.
5 Hedrick L, Epstein JI. Use of keratin 903 as an adjunct inthe diagnosis of prostatic carcinoma. Am J Surg Pathol1989;13:389-96.
6 O'Malley FP, Grignon DJ, Shum DT. Usefulness ofimmunoperoxidase staining with high-molecular-weightcytokeratin in the differential diagnosis of small-acinarlesions of the prostate gland. Virchows Arch (Pathol Anat)1990;417:191-6.
7 Chastonay P, Hurlimann J, Gardiol D. Biological tissuemarkers in benign and malignant disease of the humanprostate. Virchows Arch (Pathol Anat) 1986;410:221-9.
8 Tannenbaum M, Romas N. The prostate gland. In: Silver-berg, SG, ed. Principles and practice of surgical pathology.Vol 2. NewYork: John Wiley and Sons, 1983;1 189-217.
9 Gleason DF. Histologic grading and clinical staging ofcarcinoma of the prostate. In: Tannenbaum M, ed.Urologic Pathology: The prostate. Philadelphia: Lea andFebiger, 1977:171-98.
10 Brawn PN, Ayala AG, von Eschenback AC, Hussey DH,Johnson DE. Histologic grading study of prostate adeno-carcinoma: The development of a new system andcomparison with other methods-A preliminary study.Cancer 1982;49:525-32.
11 Gown AM, Vogel AM. Monoclonal antibodies to humanintermediate filaments II. Distribution of filament pro-teins in normal human tissues. Am Y Pathol 1984;114:309-21.
12 Elbadawi A, Craig W, Linke CA, Cooper RA. Prostaticmucinous carcinoma. Urology 1979;13:658-66.
13 Franks LM, O'Shea JD, Thomson AER. Mucin in theprostate: a histochemical study in normal glands, latent,clinical and colloid cancers. Cancer 1964;17:983-91.
14 Hukill PB, Vidone RA. Histochemistry of mucus and otherpolysaccharides in tumors. II. Carcinoma of the prostate.Lab Invest 1967;16:395-406.
15 Ro JY, Grignon DJ, Troncoso P, Ayala AG. Mucin inprostatic carcinoma. Semin Diagnost Pathol 1988;5:273-83.
16 Abel PD, Keane P, Leathem A, Tebbutt S, Williams G.Change in glycoconjugate for the binding site of the lectinUlex europeus 1 following malignant transformation ofprostatic epithelium. BrJ Urol 1989;63: 183-5.
17 Nagle RB, Ahmann FR, McDaniel KM, Paquin ML, ClarkVA, Celniker A. Cytokeratin characterization of humanprostatic carcinoma and its derived cell lines. Cancer Res1987;47:281-6.
18 Brawer MK, Nagle RB, Pitts W, Freiha F, Gamble SL.Keratin immunoreactivity as an aid to the diagnosis ofpersistent adenocarcinomas in irradiated human pros-tates. Cancer 1989;63:454-60.
19 Damjanov I. Lectin cytochemistry and histochemistry. LabInvest 1987;57:5-20.
20 Cooper HS. Lectins as probes in histochemistry andimmunohistochemistry-the peanut (Arachis hypogaea)lectin. Hum Pathol 1984;15:904-6.
1098
group.bmj.com on July 31, 2011 - Published by jcp.bmj.comDownloaded from
doi: 10.1136/jcp.45.12.1094 1992 45: 1094-1098J Clin Pathol
R F McMahon, L J McWilliam and S Mosley biopsy specimens.differential diagnosis of prostatic needle Evaluation of three techniques for
http://jcp.bmj.com/content/45/12/1094Updated information and services can be found at:
These include:
serviceEmail alerting
the box at the top right corner of the online article.Receive free email alerts when new articles cite this article. Sign up in
Notes
http://group.bmj.com/group/rights-licensing/permissionsTo request permissions go to:
http://journals.bmj.com/cgi/reprintformTo order reprints go to:
http://group.bmj.com/subscribe/To subscribe to BMJ go to:
group.bmj.com on July 31, 2011 - Published by jcp.bmj.comDownloaded from
