Immunohistochemical and Electron Microscopic Findings in Benign Fibroepithelial Vaginal Polyps J Clin Pathol: First Published As 10.1136/Jcp.43.3.224 on 1 March 1990

Home , Fibroma

224 J Clin Pathol 1990;43:224-229 Immunohistochemical and electron microscopic findings in benign fibroepithelial vaginal polyps J Clin Pathol: first published as 10.1136/jcp.43.3.224 on 1 March 1990. Downloaded from

T P Rollason, P Byrne, A Williams

Abstract LIGHT MICROSCOPY Eleven classic benign "fibroepithelial Sections were cut from routinely processed, polyps" of the vagina were examined paraffin wax embedded blocks at 4 gm and using a panel of immunocytochemical immunocytochemical techniques were agents. Two were also examined electron performed using a standard peroxidase- microscopically. In all cases the stellate antiperoxidase method.7 The antibodies used and multinucleate stromal cells were as follows: polyclonal rabbit characteristic of these lesions stained antimyoglobin (batch A324, Dako Ltd, High strongly for desmin, indicating muscle Wycombe, Buckinghamshire), monoclonal intermediate filament production. In anti-desmin (batch M724, Dako Ltd), mono- common with uterine fibroleiomyomata, clonal anti-epithelial membrane antigen (batch numerous mast cells were also often M613, Dako Ltd), monoclonal anti-vimentin seen. Myoglobin staining was negative. (batch M725, Dako Ltd), polyclonal rabbit Electron microscopical examination anti-cytokeratin (Bio-nuclear services, Read- confirmed that the stromal cells con- ing) and monoclonal anti cytokeratin NCL tained abundant thin filaments with focal 5D3 (batch M503, Bio-nuclear services). densities and also showed the ultrastruc- Mast cells were shown by a standard tural features usually associated with chloroacetate esterase method using pararo- myofibroblasts. saniline,8 which gave an intense red cyto- It is concluded that these tumours plasmic colouration, and by the routine would be better designated polypoid toluidine blue method. myofibroblastomas in view of the above An attempt was made to assess semiquan- findings. titatively the relative numbers of multinucleate and pleomorphic stromal cells and mast cells present by grading on a scale of 0- + + + . The

presence or absence of a more cellular http://jcp.bmj.com/ Benign vaginal polyps (fibroepithelial polyps, subepithelial dense zone was assessed both on pseudosarcoma botryoides, benign sarcoma- routine haematoxylin and eosin staining and tous vaginal polyps) can be misdiagnosed on sections stained with anti-desmin and as malignant embryonal rhabdomyosarcomas anti-vimentin. All cases were classified as (sarcoma botryoides) though the difference predominantly myxoid, predominantly colla- generally in age at which these two tumours genous, or mixed (when both patterns were occur must make this an uncommon occur- seen in roughly similar proportions). The on October 1, 2021 by guest. Protected copyright. rence.' They have received scant attention in number of mitotic figures per 10 high power the literature and their putative origin from fields was assessed by randomly counting 10 fibroblasts has been generally accepted."X fields on three occasions and averaging the They are usually small and asymptomatic but result. For this purpose a wide field microscope some may present with a mass or bleeding.' was used with a x 10 eyepiece, with a x 63 dry Department of Pathology, University This immunohistochemical and electron objective, at a final magnification of x 630. of Birmingham microscopic study was performed in an attempt Medical School, to elucidate further the nature of these uncom- Edgbaston lesions. ELECTRON MICROSCOPY T P Rollason mon Blocks were taken from the two most recent Department of cases after routine fixation in neutral buffered Pathology, formalin, postfixed in one per cent osmium Birmingham Maternity Hospital tetroxide for one hour at room temperature, A Williams Methods dehydrated in graded alcohols, embedded in A systematic review of all vaginal polyps in the Department of Araldite and ultrathin sectioning performed. Obstetrics and files of the Birmingham and Midland Hospital After staining in uranyl acetate and lead citrate Gynaecology, Dudley for Women was performed for the years from the grids were viewed in a JEOL 100S electron Road Hospital, 1970 to 1988 inclusive. This review yielded 11 microscope at 80Kv. Birmingham of the P Byrne "classic" fibroepithelial polyps vagina occurring in 10 women aged between 31 and Correspondence to: Dr Terence P Rollason, 64. Clinical details were obtained by review of Results Department of Pathology, the case notes and contacting the The clinical details are shown in table 1. No The Medical School, hospital by University of Birmingham, patients' general practitioners. During the clear association was seen with parity, hormone Egbaston, Birmingham B15 years included in the review about 72 000 treatment etc, though interestingly one patient 2TJ, England. cases were also had uterine and ileal leiomyomata. No Accepted for publication general gynaecological surgical 12 October 1989 examined. evidence of tumour recurrence or metastasis Immunohistochemical and electron microscopicfindings in benignfibroepithelial vaginalpolyps 225

Table 1 Clinical details of investigated cases

Follow up Case No Age Symptoms Associated pathology (months) Parity Site ofpolyp

1 48 Polyps catching clothes None 12 3 + 0 Right fornix and lower vagina J Clin Pathol: first published as 10.1136/jcp.43.3.224 on 1 March 1990. Downloaded from 2 41 Dyspareunia None 12 3 + 2 Posterior wall, near fourchette 3 54 Swelling and bleeding Leiomyomata of uterus 10 years previously 3 ? Mid-part, anterior wall 4 53 None Endometrial adenocarcinoma 120 1 + 0 Upper third, left side 5 52 None Leiomyomata of uterus and small bowel 132 2 + I Anterior wall, mid-third 6 51 None Leiomyomata of uterus 4 years previously 48 1 + 0 Anterior wall, lower third 7 56 None Breast carcinoma previously 24 1 + 0 ? 8 31 None None 120 0 + 0 Upper third, right side 9 57 Vaginal swelling None 6 0 + 0 Left fornix 10 64 Postmenopausal bleeding None 6 0 + 0 Left anterior vaginal wall

Table 2 Histologicalfindings

Size Mitotic rate Subepithelial Mast cell Pleomorphic Desmin Vimentin Case No Pattern (cm) (10 HPF) dense zone* infiltrate multinucleate cells staining staining 11 2 15 x 10 0 0 + ++ ++ ++ 12 1 20 x 15 0 0 ++ + + +++ ++ 2 1 10 x 05 0 2 + + ++ ++ 3 2 20 x 20 06 2 ++ + +++ +++ 4 1 14 x 08 0 2 ++ + +++ +++ 5 3 25 x 08 0 2 + +++ +++ +++ 6 1 15 x 08 0 1 ++ + ++ 7 1 16 x 07 0 2 +++ ++ ++ +++ 8 1 1-2max 0 2 +++ + +++ ++ 9 1 80 x 50 0 1 +++ ++ +++ +++ 10 2 20 x 15 0 1 +++ + +++ +

Pattern: 1 = myxoid *0 = absent 2 = collagenous 1 = present on PAP and haematoxylin and eosin staining 3 = mixed 2 = present on PAP staining alone

was seen in any patient during follow up. prominent than the "usual" stellate cells No patient later developed a second polyp. and was eosinophilic and slightly granular. The clinical symptoms were variable and Inflammation was not a striking feature in any inconsistent. of the cases, though this has been stated to be The overall light microscopical findings are seen in these tumours. Cross striations were not summarised in table 2. The background seen. stromal cells were spindled or stellate with fine The cellular pleomorphism classically des- collagen bundles between cells clearly defined cribed in the stromal cells in about 50% of cases against a pale background (figs 1 and 2). The was not, in its more pronounced form, apparent cytoplasm of the larger cells was more in any of our cases, and the pleomorphism that http://jcp.bmj.com/

X ° & A~~~~A Va i 40 on October 1, 2021 by guest. Protected copyright.

,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~......

~~~~~~~~~~~~~~~4, .~~~~~~~~..:~~~ ~~~~~~ ..... de i- S

0~~~~~.

.a...r

#1~~~~~~~~~~~~~~Y

-~~~~~~~~~~~~~~~~~- A~~~~~~~~~~~~~~~

W, "A-a

stroma.(Heaoyi anesi. q~~~-

(* 'jr

WAt FiueI acso Tpia per

Figure 3 Multinucleate cells showing overlapping 40 Desmin and vimentin staining were almost nuclei against a f' ..I always moderately to intensely positive in most background ofspindle v.. or all of the tumour stromal cells (fig 4), with cells. (Haematoxylin and only one exception where vimentin staining eosin.) was only weakly positive; desmin positivity in J Clin Pathol: first published as 10.1136/jcp.43.3.224 on 1 March 1990. Downloaded from this case was, however, intense. Cytokeratin, myoglobin, and epithelial membrane antigen (EMA) staining were invariably absent in stromal cells. The epithelium of the polyps was entirely squamous and generally showed only slight W"' os:s.\t@.. kO=AC. simple hyperplastic changes; no intraepithelial ... v .;9 neoplasia or other clinically important ts *. pathology was seen. Slight parakeratosis was S..: ; C... ' evident in a few cases. The stroma ofthe polyps * * ... was not .: notable only for the loose myxoid *i E ; ..... 9# appearance in most cases, but also for the ^:ki profusion of small vessels seen. Only two cases had a predominantly collagenous stroma. The mitotic rate in the tumours was extremely low; only one case showed any stromal cell mitoses. None of the cases showed the subepithelial "cambium" layer or perivascular cellular sheaths typically associated with embryonal rhabdomyosarcoma, but a feature seen in f several cases on simple haematoxylin and eosin staining (fig 5), and evident in all but one case on both vimentin and desmin staining, was the presence of a more compact subepithelial zone which seemed to be both more collagenous and was evident seemed to be mainly (due to the rather more cellular than the centres of the difficulty in differentiating in nnany cells polyps. The nature of this zone could not be between multinucleation and nuclhear atypia. further defined and it did not seem to be more The numbers of clearly multinucleate cells mitotically active than the centres of the present tended to parallel the niumbers of polyps. apparently pleomorphic cells. The nuclei ofthe A very striking feature and one not com- multinucleate cells were often poorly defined, mented on by other workers, was the invariable

overlapping, and irregular (fig 3). Only very presence of a stromal mast cell infiltrate which http://jcp.bmj.com/ occasional convincingly pleomorpihic mono- seemed to be diffuse in all cases and very nuclear stromal cells were evident. prominent in four.

Figure 4 Intense anti- ELECTRON MICROSCOPY (figs 6 and 7)

desmin positivity is seen in - In both cases the stromal cells showed clear Y '40. Al( all stromal cells (PAP t evidence of smooth muscle differentiation with

method, anti-desmin on October 1, 2021 by guest. Protected copyright. 'k numerous thin filaments present which showed monoclonal antibody.) ,, 4., focal densities. The stromal cell nuclei were .1 ... It folded and invaginated. Rough endoplasmic * ,4., 11@ reticulum and mitochondria were, however, *+ s often quite prominent and micropinocytotic vesicles were sparse. The cells were set in a very i . "loose" background with scattered bundles of

A csC mature collagen fibrils present, often closely applied to or within invaginations in the stromal cell bodies. The cells seemed to have - either no or a very incomplete external lamina. s # Subplasmalemmal densities associated with actin filaments were frequently very prominent, and some cells showed focal dilata- tion of rough endoplasmic reticulum, even &: J occasionally forming large intracellular cystic spaces. The Golgi apparatus was usually not a prominent feature. The above description indicates that the cells showed most -'..'j .. of the features typically associated with .....v, . .!. myofibroblasts.9 Rs0, Discussion The cases described in this study all fulfill the ;,( .... . and clinical criteria for the histological diag- Immunohistochemical and electron microscopicfindings in benign fibroepithelial vaginal polyps 227

previously reported case,2 although this occurred during pregnancy. On the basis ofthe immunocytochemical and electron microscopic findings presented

together with the tentative support obtained J Clin Pathol: first published as 10.1136/jcp.43.3.224 on 1 March 1990. Downloaded from

..... from the mast cell infiltrate, a feature typically seen in uterine leiomyomata, it seems clear that these lesions should no longer be designated fibroepithelial polyps, nor referred to by the nondescript term "vaginal polyp", or worse, the emotive label "pseudosarcoma botryoides". In our opinion as all or almost all of the stromal cells show the features of myofibroblasts, they should preferably be called polypoid myofibroblastomas. This opinion does not, however, concur with that expressed recently by another group6 who used electron microscopy and immunocyto- chemistry and consequently felt that vulval fibroepithelial polyps (which they thought were similar to the vaginal variant) were of fibroblastic origin and showed no desmin staining or electron microscopic smooth muscle .1~~~~~~~* features. They also indicated thar in thekJ I,.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~FfW,>- (r.y - II opinion the multinucleate cells present were X :,,DE"-.ksW histiocytic in origin. It is very difficult to explain this difference in findings other than to suggest that the vulval and vaginal polyps are of different pathogenesis: in view of their histological similarity this seems unlikely. Figure 5 Subepithelial dense zone is seen but no Interestingly, several other reactive and condensation ofstroma nosis of simple fibroepithelial polyps, and this tumorous proliferations previously thought to around vessels is evident. study confirms the entirely benign nature of be fibroblastic have been shown to contain (Haematoxylin and eosin.) these tumours. Occasional reported cases have myofibroblasts,9 and in particular they have recurred locally, but this was not seen in our been suggested to be the major cell type in series. All but one fell within the usual size another female genital tract benign neoplasm- range for these tumours previously reported namely, uterine plexiform tumour-by one elsewhere, and the unusually large polyp (case group. 10 The pleomorphic appearance noted in 9, fig 8) has been exceeded in size by one a proportion of vaginal "fibroepithelial" http://jcp.bmj.com/

Figure 6 Convoluted pattern of cell nucleus typical of stromal cells. No 4-, external cell lamina is visible. on October 1, 2021 by guest. Protected copyright.

A 228 Rollason, Byrne, Williams J Clin Pathol: first published as 10.1136/jcp.43.3.224 on 1 March 1990. Downloaded from --~~~~~~~~~~1

.2-- -.

if .1 :11, - .... .:", i. EW..11 ..

Figure 7 Focal densities onfinefilaments, rough endoplasmic reticulum, and a convoluted nucleus are evident in a cell showing no external lamina. Inset: subplasmalemmal densities are prominent. http://jcp.bmj.com/ polyps is also seen in other conditions, also mucosa in inflammatory and reactive condi- comprising myofibroblasts-for example, tions, though the electron microscopic nodular pseudosarcomatous fasciitis," and appearances favoured a histiocytic or fibro- these conditions also seem to be benign. In blastic origin for the pleomorphic cells. their original paper Norris and Taylor drew Some workers have commented on the attention to the histological similarity between similarity offibroepithelial polyps ofthe vagina

Figure 8 Case 9: this some vaginal "fibroepithelial" polyps and to the giant cell fibroma of oral mucosa,5 and on October 1, 2021 by guest. Protected copyright. unusually large polyp nodular fasciitis.' Multinucleate cells ofsimilar although they felt that the vaginal lesion (8 x 5 x 4 cm) showsa frond like surface pattern appearance have also been described in both was fibroblastic on electron microscopical and no stalk (base at upper respiratory tract213 and bladder415 sub- examination, the giant cell fibroma of oral bottom ofphotograph). mucosa ultrastructurally seems to be of myofibroblast origin.'6 We have previously seen intense desmin positivity in a single case of aggressive angiomyxoma ofthe female pelvis, a tumour which shows a predilection for a paravaginal and pelvic site of origin and shows many points of similarity with vaginal fibroepithelial polyps, though its clinical course is one of extensive local tissue infiltration. Elliot and Elliot have suggested that a subepithelial zone of looser tissue, with occasional pleomorphic cells similar to the appearance of a fibroepithelial polyp, exists normally in the cervix and vagina of a proportion ofwomen.'7 Clearly, if this is the case then it may provide the site of origin for fibroepithelial polyps. If the cells in this putative zone are also myofibroblasts then it is likely that, as at other sites, they have a contractile and possibly secretory role, maintaining the functional integrity of this organ, the Immunohistochemical and electron microscopic findings in benign fibroepithelial vaginal polyps 229

physiological functions of which require major 6 Ostor AG, Fortune DW, Riley CB. Fibroepithelial polyps with atypical stromal cells [Pseudosarcoma Botryoides] of expansile capability. vulva and vagina: A report of 13 cases. Int J GynecolPathol O'Quinn et al have emphasised that a 1988;7:351-60. 7 Sternberger LA, Hardy PHJ, Cuculis JJ, Meyer HG. The particular hormonal milieu is, in their opinion, unlabelled antibody enzyme method of immunocyto- J Clin Pathol: first published as 10.1136/jcp.43.3.224 on 1 March 1990. Downloaded from likely to be needed to form fibroepithelial chemistry. Preparation and properties of soluble antigen- antibody complex [horseradish peroxidase-antihorse- polyps, as they may be seen (rarely) at birth or radish peroxidase] and its use in the identification of may arise during pregnancy.'8 Expansion of spirochaetes. J Histochem Cytochem 1970;18:315-33. 8 Maloney WC, McPherson K, Fliegelman L. Esterase this hormone sensitive theory to the sub- activity in leucocytes demonstrated by the use of Napthol epithelial zone described AS-D Chloroacetate substrate. J Histochem Cytochem by Elliot and Elliot 1960;8:200-7. would clearly be of theoretical interest but is at 9 Ghadially FN. Diagnostic electron microscopy of tumours. present entirely speculative. Further studies of London: Butterworths, 1985:456. 10 Fisher ER, Paulson JD, Gregorio RM. The myofibroblastic the vaginal subepithelial zone are clearly nature ofuterine plexiform tumour. Arch Pathol Lab Med needed to clarify its structure and role. 1978;102:477-80. 11 Wirman JA. Nodular fasciitis, a lesion of myofibroblasts. Cancer 1976;38:2378-89. We thank Mr A Cooper for his excellent photographic assistance 12 KlenoffBH, Goodman ML. Mesenchymal cell atypicality in and Mrs S Smith for typing the manuscript. inflammatory polyps. J Otol Laryngol 1977;91:751-6. 13 Kindblom L-G, Angervall L. Nasal polyps with atypical stromal cells: a pseudosarcomatous lesion. Acta Pathol Lab Med 1984;92:65-72. 1 Norris HJ, Taylor HB. Polyps ofthe vagina. A benign lesion 14 Antonakopoulos GN, Newman J. Eosinophilic cystitis with resembling sarcoma botryoides. Cancer 1966;19:227-32. giant cells. A light microscopic and ultrastructural study. 2 Elliot GB, Reynolds HA, Fidler HK. Pseudo-sarcoma Arch Pathol Lab Med 1984;108:728-31. botryoides of cervix and vagina in pregnancy. J Obstet 15 Wells HG. Giant cells in cystitis. Arch Pathol Lab Med Gynaecol Br Commonwealth 1967;74:728-33. 1938;26:32-43. 3 Burt RL, Prichard RW, Kim BOKS. Fibroepithelial polyp 16 Weathers DR, Campbell WG. Ultrastructure of the giant ofthe vagina. A report offive cases. Obstet Gynaecol 1976; cell fibroma of the oral mucosa. Oral Surg Pathol Med 47:525-45. 1974;38:550-61. 4 Chirayil SJ, Tobon H. Polyps ofthe vagina: A clinicopatho- 17 Elliot GB, Elliot JDA. Superficial stromal reaction ef tL logical study of 18 cases. Cancer 1981;47:2904-7. lower female genital tract. Archives of Pathology 1973; 5 Miettinen M, Wahlstrom T, Vesterinen E, Saksela E. 95:100-1. Vaginal polyps with pseudosarcomatous features. A 18 O'Quinn AG, Edwards CL, Gallagher HS. Pseudosarcoma clinicopathologic study of seven cases. Cancer 1983; botryoides of the vagina in pregnancy. Gynecol Oncol 51:1 148-51. 1982;13:237-41. http://jcp.bmj.com/ on October 1, 2021 by guest. Protected copyright.