Abstract
The scrotum is an uncommon site for the presentation of extramammary Paget disease (EMPD). We describe a case of EMPD that was discovered in a patient who had been previously diagnosed and treated for squamous cell carcinoma in situ of the scrotum 3 years earlier. Pathologic examination of the current scrotectomy specimen revealed an erythematous patch with areas of pale induration. Microscopic examination revealed areas with the characteristic histology of Paget disease adjacent to areas characteristic of Bowen disease. Immunohistochemical findings demonstrated a strong expression of carcinoembryonic antigen, cytokeratin 7, and low-molecular-weight cytokeratins (CAM 5.2) in both of these areas, giving support to the overall diagnosis of EMPD. High-molecular-weight cytokeratins (34βE12) were uncharacteristically expressed in the cytoplasm of the Paget cells with equal or greater strength than in the surrounding keratinocytes, suggesting some degree of squamous differentiation. Very few publications have reported the coexistence of EMPD with squamous cell carcinoma in situ, occurring mostly in the vulva. To our knowledge, our case is the first report of scrotal EMPD with features of Bowen disease. Our findings support the theory that primary EMPD arises multifocally from multipotential epidermal cells.
The similarities between extramammary Paget disease (EMPD) and Bowen disease have been described in the literature since Bowen published his original article in 1912.1 In his discussion of atypical epithelial proliferations, Bowen commented on a histologic resemblance to Paget disease, with the presence of clear cells arising in both disorders. Jones et al2 stated that it might be impossible to distinguish EMPD and Bowen disease based on routine histologic examination alone. Currently, pagetoid Bowen disease is a well-described histologic variant of squamous cell carcinoma in situ that demonstrates nests of cells with pale staining cytoplasm within the epidermis, which lack the characteristic immunohistochemical features of Paget disease.3
One of the most elusive and frequently debated topics regarding EMPD pertains to its histogenesis. While it has been well accepted that mammary Paget disease (MPD) virtually always arises in association with an underlying breast carcinoma, EMPD has proven to be more heterogeneous in its presentation. The vast majority of EMPD cases arise in apocrine gland–bearing skin, most commonly the vulva followed by the perianal region, while other sites have been reported less commonly, including the male genitalia.4 Unlike MPD, most cases of EMPD do not arise in association with an underlying malignancy, and cases that are associated with underlying malignancy often possess different immunohistochemical staining properties, lending support to the classification of primary versus secondary disease.5 While ultrastructural and immunohistochemical studies have clearly defined the glandular features of EMPD, some findings have demonstrated a mixture of glandular and squamous features. We present a case of primary scrotal EMPD with microscopic findings that range from characteristic Paget cell morphology to full-thickness cytologic atypia resembling Bowen disease.
REPORT OF A CASE
A 60-year-old man presented to our institution with a diagnosis of recurrent squamous cell carcinoma in situ of the scrotal region. He had been seen by multiple dermatologists and urologists and had been treated with a combination of excisional biopsies and laser fulgurations during the course of 3 years. He presented with an erythematous, painful, eczematous lesion that involved the entire lower portion of the scrotum and extended to the proximal thighs bilaterally. He had no complaints of inability to void. He denied having any systemic symptoms. Past medical history was significant for coronary artery disease and peripheral vascular disease. Past surgical history included right carotid endarterectomy, coronary stent placement, and angioplasty of his left lower extremity. There was no history of malignancy.
Genitourinary examination revealed a normal circumcised phallus. Testicles were descended bilaterally and revealed no lesions. No inguinal lymphadenopathy was evident, and the patient refused prostate examination. Urinalysis findings were within normal limits. Results of a colonoscopy were unremarkable. The patient underwent wide resection, 1 to 2 cm beyond the extent of the entire scrotal lesion. The dissection was taken down to the tunica vaginalis, and the testicles were spared. Multiple frozen sections were sent intraoperatively, demonstrating the surgical margins to be free of tumor. The wound was packed with Kerlix dressing until the permanent sections revealed negative margins. The patient was then returned to the operating room 48 hours later for final grafting. His postoperative course was complicated by graft infection on postoperative day 10. He underwent debridement and recovered well, and is without recurrence 6 months later.
MATERIALS AND METHODS
The scrotectomy specimen was received fresh and measured 20.5 × 9.0 × 1.9 cm. At its center was an erythematous patch that measured 11.5 × 5.5 cm. The surface was irregular with areas of pale induration. Sections were taken from the lesion, fixed in 10% buffered formalin, processed in a standard fashion, and stained with hematoxylin-eosin. Histochemical staining was performed with periodic acid-Schiff–diastase (prediluted kit; Ventana Medical Systems, Tucson, Ariz). Immunohistochemistry on paraffin-embedded tissue was performed with the following antisera: carcinoembryonic antigen (CEA; prediluted; Ventana), CAM 5.2 (diluted 1:50; Becton Dickinson, Indianapolis, Ind), cytokeratin (CK)-34βE12 (diluted 1:50; Dako Corporation, Carpinteria, Calif), CK7 (prediluted; Cell Marque, Austin, Tex), CK20 (prediluted; Cell Marque), gross cystic disease fluid protein (GCDFP-15; diluted 1:500; Signet, Dedham, Mass), HMB-45 (prediluted; Ventana), S100 (diluted 1:100; Cell Marque), and prostate-specific antigen (prediluted; Ventana).
PATHOLOGIC FINDINGS
Microscopic examination of hematoxylin-eosin–stained sections revealed focal areas of epithelium containing nests and singly arranged large cells with pale staining cytoplasm and hyperchromatic nuclei containing occasional prominent nucleoli (Figure 1). The surrounding epithelium in these areas showed no evidence of atypia and demonstrated normal epidermal maturation. Other large segments of the epidermis revealed acanthosis with full-thickness cellular atypia, focal hyperkeratosis, and parakeratosis (Figure 2). Lymphocytes and occasional neutrophils heavily infiltrated the superficial dermis in a bandlike pattern (Figure 3), which was notably absent from the areas containing characteristic Paget cells. Atypical cells were also found tracking along hair shafts and within sweat gland ducts. We found no evidence of dermal invasion or involvement of an entire sweat gland with atypical cells.
Classic Paget cells, arranged singly and in clusters, appearing to arise from the basal layer (hematoxylin-eosin, original magnification ×200).Figure 2. Bowenoid features with full-thickness atypia and parakeratosis (hematoxylin-eosin, original magnification ×200).Figure 3. Acanthotic epidermis with full-thickness atypia and underlying bandlike dermal infiltrate (hematoxylin-eosin, original magnification ×40).Figure 4. Paget cells highlighted within the basal region of the epidermis (cytokeratin 7, original magnification ×200).Figure 5. Tumor island surrounded by normal epithelium (carcinoembryonic antigen, original magnification ×100).Figure 6. High-molecular-weight cytokeratin shown throughout the epidermis with focal cytoplasmic staining within Paget cells (HMB-45, original magnifications ×100 [left] and ×400 [right]).
Classic Paget cells, arranged singly and in clusters, appearing to arise from the basal layer (hematoxylin-eosin, original magnification ×200).Figure 2. Bowenoid features with full-thickness atypia and parakeratosis (hematoxylin-eosin, original magnification ×200).Figure 3. Acanthotic epidermis with full-thickness atypia and underlying bandlike dermal infiltrate (hematoxylin-eosin, original magnification ×40).Figure 4. Paget cells highlighted within the basal region of the epidermis (cytokeratin 7, original magnification ×200).Figure 5. Tumor island surrounded by normal epithelium (carcinoembryonic antigen, original magnification ×100).Figure 6. High-molecular-weight cytokeratin shown throughout the epidermis with focal cytoplasmic staining within Paget cells (HMB-45, original magnifications ×100 [left] and ×400 [right]).
Sections stained with diastase-resistant periodic acid–Schiff revealed focal cytoplasmic expression that was restricted to about 50% of the characteristic Paget cells. Immunohistochemical stains for CEA, CK7, and CAM 5.2 were strongly expressed in both characteristic Paget cells and in areas with full-thickness atypia. Staining for CK7 and CAM 5.2 was more prominent in the basal region of the epidermis (Figure 4). Expression of CEA was intense throughout the entire epidermis in all areas, with the exception of focally normal-appearing epithelium, revealing clearly demarcated tumor islands (Figure 5). High-molecular-weight cytokeratins also stained diffusely positive with an uncharacteristic cytoplasmic expression within Paget cells, which was equal to or greater than that in surrounding keratinocytes (Figure 6). Gross cystic disease fluid protein was focally expressed in a small number of the Paget cells. Stains for HMB-45, S100, and prostate-specific antigen were completely negative. Lastly, staining for CK20 was faint but diffusely positive.
COMMENT
Different methods have been used to study the histogenesis of EMPD. With the use of electron microscopy, Koss and Brockunier6 identified “secretory and non-secretory”–type cells, which were found to contain glycogen and mucin, respectively. They concluded from their ultrastructural findings that EMPD is derived from eccrine sweat glands. Other investigators have reported ultrastructural evidence suggesting that Paget cells may originate from keratinocytes.7 Fetherston and Friedrich8 proposed that EMPD comprises a group of lesions with similar light microscopic findings but different ultrastructural findings. Teixeira et al9 supported this opinion on a chromosomal basis when they studied tumors from the vulva and vagina, discovering that cases of EMPD contained cytogenetically unrelated clones, a finding that did not appear in any of the other tumors they looked at. They attributed this to the theory that primary EMPD arises multicentrically within the epidermis from pluripotent stem cells.9
A central feature to our understanding of EMPD is its association with apocrine gland–bearing skin and the immunohistochemical expression of apocrine cell markers. Nadji et al10 first reported that CEA is highly expressed in Paget cells, supporting a glandular origin, and they concluded that primary EMPD arises as an intraepidermal adenocarcinoma from apocrine or eccrine sweat gland ducts. Mazoujian et al11 then proposed more specific evidence for an apocrine gland origin, when they found antibodies to GCDFP-15 reacting with both Paget cells and apocrine gland cells, but not with cells of eccrine glands. They agreed with the theory, first put forward by Woodruff,12 that the squamous epithelium and pilar apparatus, including apocrine and eccrine sweat glands, are derived from the pluripotent embryonal germinativum, and that EMPD arises from the malignant transformation of a basal stem cell that expresses apocrine gland differentiation.
Other immunohistochemical stains that have demonstrated high specificity for EMPD include CK7 and low-molecular-weight cytokeratins (CAM 5.2).13 Bowen disease, on the other hand, has been reported to express high-molecular-weight cytokeratins (34βE12) but lacks the expression of low-molecular-weight keratins and other Paget cell markers, in a fashion similar to the normal epidermal keratinocytes.14 In our case, the strong expression of CK7, CEA, and CAM 5.2 found in both the Paget cells and the areas of full-thickness atypia suggests a common cell of origin and confirms the overall diagnosis of EMPD. The uncharacteristic expression of 34βE12 in the cytoplasm of the Paget cells may suggest some degree of keratinization that is normal to squamous epithelia. Evidence of heterogeneity includes the focal expression of GCDFP-15 and d–periodic acid-Schiff, which may relate to the secretor or nonsecretor status of a given cell.
Other authors have reported the association of squamous cell carcinoma in situ with EMPD, occurring mostly in the vulva. Orlandi et al15 described 2 cases of keratinizing vulvar intraepithelial neoplasia admixed with vulvar EMPD, in which atypical keratinocytes stained diffusely positive for 34βE12, whereas Paget cells lacked this expression. Peralta et al16 described a lesion from the right temple of a 71-year-old man that demonstrated histologic features of Bowen disease admixed with EMPD. Malignant cells involving the sweat glands and sweat ducts stained positive for CEA, while the atypical epidermal keratinocytes did not, suggesting the presence of a mixed carcinoma in situ.16 Lastly, Williamson et al described 2 cases of pagetoid Bowen disease in which neoplastic cells expressed CK7 but lacked other Paget cell markers.3 To the best of our knowledge, our case is the first to describe the association of EMPD with features of Bowen disease in the scrotum.
In conclusion, EMPD is a heterogeneous disease showing variability between and within individual cases, lending support to the multipotent stem cell theory. Ultrastructural, immunohistochemical, and cytogenetic reports have demonstrated the presence of mixed squamous and glandular features in some cases. The multifocal nature of EMPD may help explain the variable morphology seen within a given case. Additionally, it may be appropriate to use the descriptive term, bowenoid Paget disease or Paget disease with bowenoid features, for cases such as ours, in which immunohistochemistry supports the overall diagnosis of Paget disease, but histologic features are focally suggestive of Bowen disease. Finally, we hope awareness of this information will help prevent misdiagnosis on biopsy specimens that, owing to sampling error, lack the characteristic histology of EMPD.
References
Author notes
Reprints: Adam M. Quinn, DO, Department of Pathology, Loyola University Medical Center, 2160 S First Ave, Bldg 110, Maywood, IL 60153 ([email protected])