Brushing cytology in cutaneous lesions of the head and neck

Background: Brushing cytology is a well established diagnostic procedure used by gynaecologists, physicians and surgeons to obtain representative samples from lesions. Our aim was to evaluate its reliability in ulcerative and tumour-like conditions arising in the skin of the head and neck.

Over 28 months, 86 patients with suspected cutaneous malignant lesions underwent a cytological examination with a cytobrush within the otolaryngology department.

Cytological analysis identified 63 out of 64 histologically documented malignant tumours (60 primary basal cell and squamous cell carcinomas and three metastatic adenocarcinomas), and 21 out of 22 benign lesions. There was one false positive and one false negative result.

Brushing cytology of suspected cutaneous malignant lesions is a rapid and reliable diagnostic method which helps the clinician to decide on appropriate planning and treatment. The technique can be performed as an out-patient procedure, and smear preparation can be done in the laboratory, even at a peripheral hospital. Key words: Cytology; Skin Neoplasms; Head and Neck

There are two absolute indications for pre-operative diagnosis of tumours and tumour-like conditions of the skin, subcutaneous tissue and soft tissues: a primary lesion, and clinical suspicion of a recurrence or metastasis. The ‘gold standard’ for diagnosis is open biopsy of the lesion with histological examin- ation of the excised tissue. Pre-operative cytological evaluation is not a well recognised modality in this field, yet it offers several advantages. 1,2 The aims of cytological evalu- ation are to establish an aetiological and/or morpho- logical diagnosis and thereby to establish a more accurate prognosis. Cytology is a useful tool for differentiating inflammatory and infectious lesions from those that are neoplastic. In many cases, cytology is also helpful in determining whether a tumour is malignant or benign. Cytology does however have its limitations, and these should be recognised. Problems may arise when an inflamma- tory response results in secondary dysplastic changes which can mimic those normally associated with neoplasia. It is also worth noting that, in poorly differentiated tumours, cytological examin- ation may not identify the tissue of origin. Cytology, therefore, should not be regarded as a substitute for histopathological examination of biopsy specimens. Histology is more likely to provide a definite diagnosis and, since biopsies preserve tissue architec- ture, grading and classification of the tumour is usually possible. Cytological examination of skin brushing material can be considered a rapid screening technique and may be used as an adjunct to biopsy. Patient anxiety can be relieved by providing an instant diag- nosis, followed by discussion of treatment options. Surgery can be avoided if the lesion proves to be non- neoplastic, or delayed for convenience if it is benign. A diagnosis of malignancy allows pre-operative staging and planning of the extent of surgery. In addition, some cases may be managed either by radiotherapy 3 or local (intralesional) interferon treatment.4,5 Furthermore, surgery in other cases may cause complications; for example, in older patients receiving systemic therapy, and in patients with multiple lesions for which restoration demands extensive skin allografting. The most common malignant tumours arising from a chronic disease background are squamous cell carcinoma (SCC) and basal cell carcinoma (BCC). Canti published, in 1979, a study of 1628 BCCs, which used scraping cytology and obtained very satisfy- ing diagnostic results. 6 In contrast, we used brushing cytology in our work, using a rapid staining method From the Departments of Cytopathology and *ENT, Chania General Hospital, Chania, Crete and the †ENT Department, General Hospital of Volos, Volos, Greece. Accepted for publication: 19 April 2006. The Journal of Laryngology & Otology (2007) , 121, 676 – 679. # 2007 JLO (1984) Limited doi:10.1017/S0022215107007104 Printed in the United Kingdom First published online 3 April 2007 676 https://doi.org/10.1017/S0022215107007104 Published online by Cambridge University Press (one minute duration) which was repeated when necessary, and we found that the sampling inade- quacy rate was significantly diminished. This techni- cal difference must be stressed; in the 1979 study, the sampling inadequacy rate was high due to inadequate exfoliation of cells. Correct smear prep- aration and proper fixation and staining techniques are also essential for optimal diagnostic results. 7 Our study investigated the diagnostic accuracy, specificity and sensitivity of brushing cytology in head and neck cutaneous lesions.

Eighty-six patients were examined in the department of otolaryngology, head and neck surgery of Chania General Hospital, over a 28-month period, and an equal number of lesions were documented in the files of the cytology department. In most cases, clinical examination found an ulcer or exophytic mass. In a small number of cases, there was a healing or a flat, red-grey lesion surrounded by a small halo. The lesions were localised on the preauricular area (27), the temporal area (26), the lateral aspects o ft h en e c k( 1 5 ) ,t h eo r b i t( fi v e ) ,t h en o s e( t h r e e ) , the jugal-gingival groove ( five) and the lips ( five). A gynaecological cytobrush was used to perform the examination (Figure 1). To test whether smears were satisfactory or not, a Giemsa quick-staining

(Hemacolor, Merck, Darmstadt, Germany) was applied, and if necessary the sampling was repeated. Subsequently, smears were stained with May–Gru¨ nwald–Giemsa and Papanicolaou stains.

Of the 86 cases undergoing cytological analysis, 63 were reported as malignant (60 primary BCCs and SCCs (Figures 2 and 3), and three metastatic adeno- carcinomas from the breast, kidney and gastrointesti- nal tract) and 21 were reported as benign (Table I). Cytological analysis resulted in one false negative result, a misdiagnosis of keratoacanthoma. Histologi- cal analysis documented a SCC arising on the jugal-gingival groove. Histological analysis also revealed one false positive result, concerning a case of endometriosis occurring in the nose, which was cytologically interpreted as a BCC (Figure 4). Our cytological method showed a high rate of diag- nostic accuracy, with 98.43 per cent sensitivity and 95.45 per cent specificity (95 per cent confidence interval).

Disorders in the head and neck are accessible to inspection and palpation. Therefore, biopsy is the established diagnostic approach, providing the clini- cian with accurate histological evaluation. Cytological diagnosis of common primary cutaneous tumours, both benign and malignant, such as SCC and BCC, is well documented in the literature. 7 When faced with the differential diagnosis of cutaneous BCC versus cutaneous SCC, a reasonable first approach would be immunostaining for epithelial membrane antigen (EMA) and epithelial specific antigen Ab-9 (clone Ber-Ep4). Negativity for the former and posi- tivity for the latter would favour a diagnosis of BCC. Cytology can almost always detect malignancy, so it plays an important role in the pre-operative investi- gation of primary skin tumours, as well as in the evalu- ation of possible metastasis from a previously documented neoplasm. This has been proven in our work too, with the exception of one case misdiag- nosed as keratoacanthoma, and another one misdiag- nosed as BCC. In cases of metastatic deposits, it is necessary to detect and manage the primary site. In our study, there were three cases of metastatic adenocarcinoma, primarily arising from the breast, kidney and gastrointestinal tract. Distinguishing keratoacanthoma from squamous cell carcinoma is a persistent issue in pathology practice. Putti et al. 8 have reported that analysis of telomerase activity, cyclooxygenase isoenzyme 2 (COX-2) and p53 expression provides evidence that keratoacanthoma and squamous cell carcinoma are indeed distinct lesions and also helps differentiate the two lesions, despite their similarity on convention- al morphology. Keratoacanthoma has recently been reclassified as squamous cell carcinoma- keratoacanthoma type to reflect the difficulty in histo- logical differentiation as well as the uncommon but potentially aggressive nature of keratoacanthoma. The term ‘squamous cell carcinoma-keratoacanthoma type’ has been introduced for otherwise classical ker- atoacanthomas that reveal a peripheral zone formed FIG.1 The cytobrush used to perform the examination. FIG.2 Basal cell carcinoma arising in the temporal area (May– Gru¨ nwald–Giemsa; /C2 400). BRUSHING CYTOLOGY IN HEAD AND NECK LESIONS 677 https://doi.org/10.1017/S0022215107007104 Published online by Cambridge University Press by squamous cells with atypical mitotic figures, hyper- chromatic nuclei and loss of polarity to some degree. These marginal cells may also penetrate into sur- rounding tissue in a more aggressive pattern. In the case of endometriosis misdiagnosed on cytological analysis as BCC, the smears contained fragments of palisading epithelial cells resembling those of BCC, with uniform nuclei and a bland chromatin pattern. No stromal cells were recognised. 7 In theory, cutaneous metastases may result from any neoplasm; this is borne out in practice. 9 To make a diagnosis of sweat gland carcinoma, one should first establish that the tumour shows sweat gland differentiation, as recognised by identification of extracellular ductal or intracytoplasmic lumen formations. This can be highlighted by their diastase-resistant periodic acid-Schiff, EMA, and carcinoembryonic antigen positivity. Demonstration of S-100 protein may be a useful pointer to sweat gland differentiation. Distinction of some types of sweat gland carcinoma from metastatic adenocarci- noma is not possible on morphological grounds. Immunohistochemical analysis does not allow dis- tinction between a primary adnexal tumour and a metastatic tumour, except in a few cases (i.e. prostate and thyroid). Malignant sweat gland tumours are often positive for oestrogen and progesterone recep- tors, and these markers are therefore of limited usage in the differential diagnosis. Finally, there are some benign lesions which can be cytologically diag- nosed. 10 – 13 Cytological analysis can rule out metas- tasis from a documented neoplasm in followed-up patients with newly arisen skin lesions. Cytological analysis can also diagnose uncommon cystic lesions (e.g. keratin cysts) and inflammatory processes. Significant indications for the brushing cytology

are as follows (Table II). Firstly, in cases in which rapid diagnosis is needed; a skin brushing can be stained and accurately interpreted in a few minutes, and this simple method requires no freezing, paraffin embedding or microtome. Secondly, in cases in which a biopsy is not possible: this may be due to patient refusal, lack of tissue-processing facilities, inaccessibility of the lesion or the danger of biopsy complications. Thirdly, in cases in which a biopsy is not needed; the physician need only to confirm a clini- cal diagnosis with a minimum of trauma or pain (e.g. a cutaneous lesion which is obviously benign and needs no biopsy but is of unknown aetiology, or a lesion for which the possibility of occult malignant change must be ruled out). Fourthly, in cases in which follow up of treated lesions is required, and in which repeated biopsies would be wasteful and /or poorly tolerated by the patient; cytology is usually highly sensitive in the diagnosis of recurrence when a good specimen is obtained. Fifthly, brushing cytology may be indicated as an adjunct to excisional biopsy frozen section, for example, in order to evaluate the persistence of tumour cells in the margins of an excision. Sixthly, brushing cytology may be indicated for its safety; biopsy of certain tumours (e.g. malignant melanoma) may release neoplastic cells into lymphatic and blood vessels, whereas gentle brushing for cytological analy- sis decreases the chances of such an occurrence. 14 . This study investigated the use of brushing cytology in cutaneous lesions of the head and neck . Cytology correctly identified 63 of 64 histologically documented malignant tumours of the skin of the head and neck . Brushing cytology was a rapid and reliable diagnostic method which could be performed in an out-patient setting TABLE II INDICATIONS FOR BRUSHING CYTOLOGY Rapid diagnosis required Biopsy not possible Biopsy not needed Follow up of treated lesions Adjunct to biopsy Safety /C3 /C3 Avoiding dissemination of disease. FIG.3 Squamous cell carcinoma arising on the upper lip (Papani- colaou; /C2 400). TABLE I

Method Total cases ( n) þve (n) 2ve (n) Cytology 86 63 21 Histology 86 63 23 True positives ¼ 63; true negatives ¼ 21; false positives ¼ 1; false negatives ¼ 1. þve ¼ positive; 2ve ¼ negative FIG.4 Endometriosis, misdiagnosed as basal cell carcinoma (Papani- colaou; /C2 400). D TAMIOLAKIS, E PROIMOS, G E PEROGAMVRAKIS et al.678 https://doi.org/10.1017/S0022215107007104 Published online by Cambridge University Press Brushing cytology has some limitations, including: the absence of a cell pool in which cells may accumu- late and remain moist; the difficulty of penetrating the superficial, horny, squamous layers to access diagnostic cells in deeper lesions; the comparative ease of obtaining punch biopsies; and cytologists’ unfamiliarity with this type of specimen.

Despite the exponential interest and growth in dermatopathology over the years, and the fact that the skin is the largest desquamating organ in the body, interest in cutaneous cytology has in the past been limited. Although not a substitute for standard histological analysis, in the hands of an experienced cytopathologist, brush smears can aid in establishing the clinical diagnosis with ease and rapidity and can serve as an adjunct to routine histological study. The technique is cheap, easy to perform and does not cause any discomfort to the patient. In remote areas where facilities for full histopathological examination are unavailable, brushing cytology may represent a simple and useful diagnostic adjuvant. Brushing cytology’s reliability, rapidity and easy performance without anaesthesia warrant serious consideration of its application within the field of head and neck surgery.

1 Orell S, Sterrett G, Walters M, Whitaker D. Supporting tissues. In: Orell S, Sterrett G, Walters M, Whitaker D, eds. Manual and Atlas of Fine Needle Aspiration Cytology , 2nd edn. New York: Churchill Livingstone, 1992;300–34 2 Linsk JA, Franzen S. Melanomas and skin nodules. In: Linsk JA, Franzen S, eds. Clinical Aspiration Cytology , 2nd edn. Philadelphia: JB Lippincott, 1989;319–36 3 Reisner K, Haase W. Electron beam therapy of primary tumors of skin. Radiol Med 1990;8:114–15 4 Dogan B, Harmanyeri Y, Balooglu H, Oztek L. Intrale- sional Alfa-2a interferon therapy for basal cell carcinoma. Cancer 1995;91:215–19 5 Greenway T, Cornell C, Tanner J, Peets E, Bordin M, Wagi C. Treatment of BCC with intralesional interferon. Acad Dermatol 1986;15:437–43 6 Canti G. Skin cytology and its value for rapid diagnosis. Acta Cytol 1979;23:516–17 7 Orell S, Sterrett G, Walters M, Whitaker D. The tech- niques of FNAC. In: Orell S, Sterrett G, Walters M, Whitaker D, eds. Manual and Atlas of Fine Needle Aspira- tion Cytology, 2nd edn. New York: Churchill Livingstone, 1992;8–23 8 Putti T, Teh M, Lee YS. Biological behavior of keratoa- canthoma and squamous cell carcinoma: telomerase activity and COX-2 as potential markers. Modern Pathol- ogy 2004;17:468–75 9 Reingold IM. Cutaneous metastases from internal carci- noma. Cancer 1966;19:162–9 10 Duperrat B, Badillet G. Cytological examination of Bullae and Vesicles [in French]. Rev Cytol Clin 1971;4:9–16 11 Dey P, Das A, Radhika S, Nijhawan R. Cytology of primary skin tumors. Acta Cytol 1996;40:708–13 12 Rege J, Shet T. Aspiration cytology in the diagnosis of primary tumors of skin adnexa. Acta Cytol 2001;45:715–22 13 Cheng L, Binder SW, Cajjar NA, Hirchowitz SL. Fine needle aspiration as an adjunct to the diagnosis of a rare adnexal tumor of hair follicle origin: trichoblastoma. Diagn Cytopathol 2003;29:225–8 14 Naib Z. The vulva and the skin. In: Naib Z, ed. Cytopathol- ogy, 4th edn. Boston: Little, Brown & Co, 1996;199–217 Address for correspondence: Dr Chariton E Papadakis, 1 Akrotiriou St, Chania, Crete 73133, Greece. Fax: þ30 2821055654 E-mail: [email protected] Dr C E Papadakis takes responsibility for the integrity of the content of the paper. Competing interests: None declared BRUSHING CYTOLOGY IN HEAD AND NECK LESIONS 679 https://doi.org/10.1017/S0022215107007104 Published online by Cambridge University Press