Primary peritoneal serous borderline tumors.

From the Department of Pathology and Laboratory Services, Brooke Army Medical Center, Fort Sam Houston, Tex (CPT Hutton); and the Department of Pathology, Bassett Army Hospital, Fort Wainwright, Alaska (MAJ Dalton) Search for other papers by C. P. T. Robert L. Hutton, MC in From the Department of Pathology and Laboratory Services, Brooke Army Medical Center, Fort Sam Houston, Tex (CPT Hutton); and the Department of Pathology, Bassett Army Hospital, Fort Wainwright, Alaska (MAJ Dalton) Search for other papers by M. A. J. Scott R. Dalton, MC in Context.—Primary peritoneal serous borderline tumor is a rare epithelial proliferation that can present as an incidental finding at laparotomy and raises concern for a primary ovarian tumor with peritoneal implants. Objective.—To present a brief history of this condition and describe its distinctive histology and clinical presentation, as well as to review the chief differential diagnostic considerations, to include mesothelial proliferations, endosalpingiosis, endometriosis, high-grade primary peritoneal papillary serous carcinoma, and implants from primary ovarian serous neoplasms. Data Sources.—Relevant articles indexed in PubMed (National Library of Medicine) between 1966 and 2005, references thereof, and reference surgical pathology texts. Conclusions.—Primary peritoneal serous borderline tumor should be considered in the differential diagnosis of an epithelial proliferation with prominent psammoma bodies on the peritoneal surface of specimens submitted for nongynecologic complaints. Primary peritoneal serous borderline tumor (PPSBT), also known as serous micropapillomatosis of low malignant potential, is a rare proliferation of epithelial cells that can present as an incidental finding at laparotomy and raise concern for a primary ovarian tumor with peritoneal implants. Although some controversy remains regarding the histogenesis of this and related müllerian-type peritoneal proliferations, they are thought to arise from secondary müllerian system remnants either embedded within or closely approximating the peritoneum.1 Before this term was coined, similar lesions had been noted as “atypical endosalpingiosis” 23 or “primary papillary peritoneal neoplasia.” 4 In the last few decades, 2 series have been published discussing this rare proliferation.56 A recent case of a PPSBT, found incidentally after colectomy in a 57-year-old woman with diverticulosis, was diagnosed by the authors at our institution. The consideration of this diagnosis based on the histologic findings prompted an extensive review of the literature and consideration of the entities in the differential diagnosis, as discussed later in this article. GROSS AND MICROSCOPIC FEATURES At surgery, the gross characteristics of PPSBT have a variable appearance, with focal-to-diffuse lesions that may be miliary, granular, or nodular or may resemble adhesions. Blebs or cystic lesions have also been described. A dominant peritoneal nodule or cyst has been described in 4 cases.56 The microscopic features of these peritoneal/serosal foci have traditionally been divided into 2 categories: lesions with or without a prominent fibroblastic response surrounding the epithelial cell proliferation.7 Lesions with a prominent fibroblastic reaction appear “stuck on” to otherwise-normal serosa with an easily discerned demarcation between the reactive and normal serosal stroma. The epithelial component is entrapped within and often distorted by the surrounding stromal proliferation. If this distortion is severe, the lesion may contain individual cells and small solid clusters. The cells lack significant cytologic atypia and mitoses. Destructive stromal invasion is usually not identified. Lesions that lack a fibroblastic reaction consist of broad papillary groups of epithelial cells with mild-to-moderate nuclear atypia, either loosely attached to or detached from the peritoneal surface (Figure 1). The cells constituting the papillae may be flat, cuboidal, or columnar with one to several layers of lining cells. The lining cells may detach from the surrounding stroma, forming “clefts” (Figure 2, A and B). Psammoma bodies are quite often numerous and prominent among these lesions. REVIEW OF THE LITERATURE Bell and Scully5 described 25 cases originally diagnosed as PPSBT, atypical endosalpingiosis, or serous cystadenoma/cystadenofibroma with peritoneal implants. The patients were all women, aged 19 to 53 years, and had a variable clinical history, which included abdominal/pelvic pain (7 patients), infertility (6 patients), chronic pelvic inflammatory disease, small bowel obstructive symptoms, and amenorrhea. Eight of the patients' peritoneal lesions were incidental to laparotomy for other reasons (benign ovarian serous tumor, leiomyoma, ventral hernia, and cesarean section). Sixteen of the patients were treated with total abdominal hysterectomy with bilateral salpingo-oophorectomy. The remaining 9 underwent more conservative therapy (ie, various combinations of unilateral salpingo-oophorectomy with and without ovarian/peritoneal biopsies). After an average of 8 years of follow-up, no significant recurrences were seen in 21 of the 25 patients, including 8 who underwent conservative therapy. Of the 4 patients with recurrent disease, 2 were cured after repeat resection. One was diagnosed with low-grade peritoneal serous carcinoma after the repeat operation and was alive with extensive disease at publication. The fourth patient died of disease 10 years after the initial operation. Biscotti and Hart6 studied 17 cases of PPBST. The patients were all girls or women aged 16 to 67 years. In this group, only 8 (47%) of the patients had a clinical prodrome, as compared to 64% in Bell and Scully's series. The nature of the prodrome, however, was similar to those patients in the previous study (5 patients with chronic abdominal/pelvic pain and 1 each with small-bowel obstruction, adnexal mass, and clinical findings consistent with endometriosis). The other 53% of Biscotti and Hart's cases were incidental to surgery performed for other reasons, such as caesarean section, endometriosis, or adhesions. In this series, there were no disease-related deaths, and only 2 of 14 cases had a documented recurrence of disease (3 of the original 17 were lost to follow-up). Interestingly, both recurrences presented as small bowel obstruction, 1 with multiple episodes beginning 6 months after the primary operation, and the other 14.5 years after primary operation. Both patients were alive without morbidity 10.9 and 16.2 years after their initial operations, respectively. DIFFERENTIAL DIAGNOSIS When approaching a specimen similar to that described above, a few important diagnostic considerations besides PPSBT must be entertained. The pathological differential diagnosis includes endometriosis, endosalpingiosis, benign reactive mesothelial proliferations such as adenomatoid tumor or florid mesothelial hyperplasia, and borderline mesothelial proliferations with similar histological features, which include benign or well-differentiated papillary mesothelioma (Table 1). The most important neoplastic entities in the differential diagnosis are primary ovarian papillary serous borderline tumor with implants and high-grade primary peritoneal papillary serous carcinoma.8–12 Well-differentiated papillary mesothelioma (WDPM) may assume a combination of tubular and papillary forms loosely associated with the peritoneal surface, and both may have psammomatous inclusions. In addition, WDPM bears a clinical resemblance to PPSBT, often presenting with adhesions or a finely nodular peritoneum as an incidental finding (eg, during surgery for hernia or abdominal aortic aneurysm repair).13 One case of well-differentiated papillary mesothelioma in association with small bowel obstruction was also found in our review.14 The first helpful feature in the differential between these 2 entities is that although WDPM is far more common in women, it does occur in men as well. A microscopic feature favoring WDPM is the presence of scattered multinucleate giant cells in the stroma surrounding the epithelial tumor component. This feature is not described in PPSBT. Unfortunately, this feature is only present in 9% to 10% of WDPMs.13 The presence of a primary ovarian mass does not rule out WDPM. There have been descriptions in the literature of WDPM with synchronous ovarian teratoma and ovarian endometrioid adenocarcinoma.13 It is also important to note that despite its benign-sounding name and relatively bland cytology, WDPM is best categorized as a tumor of low malignant potential, as death from disseminated disease has been well described.15 It is interesting to note that residual/recurrent disease and death from disease disproportionately affected men1315; however, more study is needed before this conclusion can be reached with certainty. Florid mesothelial hyperplasia in association with ovarian tumors has also been described.16 The histologic findings of sheets, cords, and papillary structures of flat-to-cuboidal cells with hyperchromasia and mild nuclear atypia may mimic PPSBT; however, psammoma bodies generally are not prominent (Figure 3, A and B). In addition, the presence of binucleated or multinucleated cells in the epithelial component without the nuclear features of malignancy would favor a mesothelial origin. Coagulative necrosis may also be seen in association with florid mesothelial hyperplasia, a finding not generally seen in PPSBT. Artifactual clefting may be found surrounding florid mesothelial hyperplasia, making differentiation from PPSBT difficult on hematoxylin;cheosin;ndstained slides.17 In order to help distinguish mesothelial proliferations from PPSBT, an immunohistochemical panel is essential.18 Mesothelial cells are classically regarded as calretinin positive by immunohistochemical methods. However, 22% of ovarian-type epithelial neoplasms can also stain positive for calretinin.19 Lack of calretinin staining with positive epithelial markers (B72.3, Ber-EP4, and pancytokeratin) usually supports an epithelial origin (Figure 4, A and B).20 Endosalpingiosis is generally distinguished by the appearance of benign fallopian tube epithelium in peritoneal inclusions. A clinical history of a prior fallopian tube lesion (eg, tubal pregnancy, hydrosalpinx, chronic salpingitis, or salpingitis isthmica nodosum) is often (but not always) present. Psammoma bodies are frequently present, sometimes in large numbers, in the stroma underlying the inclusions (Figure 5). The glandular epithelium is columnar and commonly ciliated. Intercalated cells may be appreciated. Mitotic activity and nuclear atypia are absent, which in combination with the presence of cilia, generally separates this entity from PPSBT. When endometrial stroma and/or hemorrhage is present around the columnar epithelium, the diagnosis of endometriosis is warranted, though it should be noted that endometriosis and endosalpingiosis are not mutually exclusive diagnoses and are indeed commonly found adjacent to one another in peritoneal inclusions.8 In addition, reactive mesothelial hyperplasia with mild atypia, including characteristic “coffee-bean” nuclei with single prominent nucleoli, has been described in association with endometriosis.21 In those cases in which the ovaries have not been imaged or visualized during surgery, implants from a primary ovarian serous borderline tumor must also be considered. It is important to consider recent developments in this area in order to fully appreciate the difficulty and potential for pitfalls in this differential diagnosis. Recently, an attempt has been made to histologically reclassify ovarian serous borderline tumors based on prognostic data.910 To this end, some authors have subdivided serous borderline tumors into two groups: atypical proliferative serous tumors and micropapillary serous carcinomas (MPSCs).22–24 These entities demonstrate both noninvasive and invasive forms and may have associated noninvasive and invasive implants. Of the 2 subtypes, MPSCs have been shown to behave differently clinically and are considered by the authors of this subclassification to be low-grade serous carcinomas. Although survival rates for FIGO stage III and IV MPSCs with invasive implants approach those of advanced-stage serous carcinoma, MPSCs as a whole are usually indolent in behavior.2223 Historically, the diagnosis of invasive implants associated with ovarian serous borderline tumors required the presence of cytologically malignant cells irregularly infiltrating normal tissue.2425 The histologic features and definition of invasive implants associated with MPSCs has recently been reevaluated and characterized as (1) invasion of normal tissue, (2) a micropapillary architecture and (3) small nests of cells surrounded by a space or cleft, with the determination that implants demonstrating any one of these features should be classified as invasive.9 The histology of MPSC with invasive implants is important in the differential diagnosis of PPSBT because they can be virtually identical on hematoxylin-eosin–stained sections (Figure 6, A and B). Smith Sehdev et al10 hypothesized that even implants displaying a micropapillary architecture or small solid nests of epithelial cells surrounded by a space that do not show destructive stromal invasion represent metastatic MPSCs. However, they also entertained the possibility that these foci could represent independent primary peritoneal carcinomas.10 Clinical staging with computed tomographic scan and exploratory laparotomy is the only way to currently distinguish PPSBT from MPSC with invasive implants. Thus, the pathologist who is confronted with a possible PPSBT must insist on either imaging or direct observation of both ovaries before the diagnosis of PPSBT can be made with certainty. MOLECULAR EVENTS IN OVARIAN AND PERITONEAL SURFACE EPITHELIAL NEOPLASMS As with PPSBT and MPSC with invasive implants, a similar dilemma has been described in differentiating high-grade primary peritoneal neoplasms from primary metastatic ovarian serous papillary carcinoma.11 These lesions differ from both PPSBT and MPSC by the presence of high-grade nuclear features, which are characterized by markedly enlarged, pleomorphic, and vesicular nuclei displaying large prominent nucleoli (Figure 7). Halperin et al11 compared primary peritoneal and primary ovarian serous papillary carcinoma by analyzing the immunohistochemical status of estrogen receptor and progesterone receptor, the expression of cell proliferation marker Ki-67, and overexpression of HER-2/neu and p53 protein. Primary peritoneal neoplasms showed significantly less expression of estrogen receptor and progesterone receptor than ovarian primary tumors did. Conversely, primary peritoneal neoplasms demonstrated increased expression of Ki-67 and HER-2/neu when compared to primary ovarian tumors. The data suggested that different molecular events triggered these clinically distinct tumors. Others have proposed a hypothesis for the molecular events triggering primary ovarian tumors. De novo serous carcinoma is believed to arise from ovarian surface epithelium or inclusion cysts.2627 Singer et al27 proposed that mutations in K-ras and BRAF are present in MPSC (low-grade serous carcinoma), whereas p53 mutations occurred more often in high-grade serous carcinoma, suggesting dual pathways of serous carcinogenesis. This hypothesis was recently further supported by a molecular genetic study using a highly stringent nucleotide sequencing method.28 Based on the histologic similarities between MPSC with invasive implants and PPBST, we hypothesize that these entities may also possibly share similar genetic aberrations. CONCLUSION In summary, although PPSBT is a rare entity, it is nonetheless important because of its similarity to other peritoneal and müllerian proliferations of both benign and malignant biologic potential (Table 2). It should be considered in the differential diagnosis of epithelial proliferations with associated psammoma bodies on the peritoneal surfaces of viscera submitted for nongynecologic problems. Correlation with clinical presentation, radiographic findings, surgical staging, and histologic immunohistochemical analysis of the epithelial proliferation is essential in determining the correct diagnosis. Copyright: College of American Pathologists 2007 Figure 1. This “implant” demonstrates mild cytologic atypia with clefting and micropapillary architecture (hematoxylin-eosin, original magnification ×200). Figure 2. A, Serosal surface of a colon resection specimen for diverticular disease. Note the micropapillary architecture of epithelial cells surrounding prominent psammomatous calcifications (hematoxylin-eosin, original magnification ×40). B, High-power view, demonstrating the clefts surrounding the micropapillary groups (hematoxylin-eosin, original magnification ×400) Figure 3. A, Papillary mesothelial hyperplasia, low power. Note the lack of psammoma bodies (hematoxylin-eosin, original magnification ×40). B, Papillary mesothelial hyperplasia, high power. Occasional cells with cytologic atypia may focally mimic primary peritoneal serous borderline tumor (hematoxylin-eosin, original magnification ×400). Figure 4. A, Epithelial groups of primary peritoneal serous borderline tumor positive for Ber-EP4 (original magnification ×40). B, Primary peritoneal serous borderline tumor demonstrating a negative calretinin immunostain (original magnification ×40). Figure 5. Endosalpingiosis showing prominent psammoma bodies in the surrounding stroma (hematoxylin-eosin, original magnification ×100). Figure 6. A, Epithelial group of primary peritoneal serous borderline tumor with surrounding space/cleft (hematoxylin-eosin, original magnification ×400). B, Invasive micropapillary serous carcinoma with cleft surrounding the invasive micropapillary epithelial cells. Note the similarities between primary peritoneal serous borderline tumor and the invasive component of micropapillary serous carcinoma (hematoxylin-eosin, original magnification ×400). Figure 7. Primary peritoneal serous carcinoma with increased pleomorphism, hyperchromasia, and overall celluarity compared to primary peritoneal serous borderline tumor (hematoxylin-eosin, original magnification ×200). Contributor Notes Reprints: CPT Robert L. Hutton, Department of Pathology and Laboratory Services, ATTN: MCHE-PLA, Brooke Army Medical Center, 3856 Roger Brooke Dr, Fort Sam Houston, TX 78234 ([email protected])