A Cytohistologic Study of Pseudomyxoma Peritonei in 67 Patients Stephanie L
Total Page:16
File Type:pdf, Size:1020Kb
Gelatinous Ascites: A Cytohistologic Study of Pseudomyxoma Peritonei in 67 Patients Stephanie L. Jackson, M.D., Ronald A. Fleming, PhD., Brian W. Loggie, M.D., Kim R. Geisinger, M.D., Department of Pathology (SLJ, KRG), Wake Forest University Baptist Medical Center, Winston-Salem, North Carolina; Clinical Research Program (RAF), Glaxo Wellcome, Inc., Research Triangle Park, North Carolina; and University of Texas Southwestern Medical Center (BWL), Moncrief Cancer Center, Fort Worth, Texas KEY WORDS: Pseudomyxoma peritonei, Dissemi- Background: Pseudomyxoma peritonei (PMP) is a nated peritoneal adenomucinosis, Peritoneal muci- rare condition characterized by gelatinous ascites. nous carcinomatosis, Gelatinous ascites, Mucinous Although the histologic attributes of PMP have been ascites Peritoneal cytology. well studied, the cytologic features remain ill de- Mod Pathol 2001;14(7):664–671 fined. Methods: We reviewed the peritoneal wash- ings (PW) in 67 patients with PMP to identify cyto- Pseudomyxoma peritonei (PMP) is a condition morphologic features useful in classifying cases as characterized by the accumulation of mucinous either disseminated peritoneal adenomucinosis material in the peritoneal cavity (gelatinous as- (DPAM) or peritoneal mucinous carcinomatosis cites). Several theories have been proposed to ex- (PMCA). Histologic specimens were correlated with plain the pathogenesis of PMP including multifocal the cytologic diagnoses. Correlation between cyto- neoplasia of the peritoneum, ovaries, and appendix logic diagnosis and patient outcome was investi- (1), mucinous metaplasia of the peritoneum (2), gated. Results: Neoplastic epithelial cells were iden- and metastasis, rupture, or leakage from a primary tified in 63 of 67 PW (94%). Concordance with the mucinous neoplasm of the ovary or appendix (3–8). histologic diagnosis was obtained in 61 of 63 cases. Recent studies incorporating immunohistochemi- cal stains such as cytokeratins (CK) 7, 17, 18, and 20, Of these 36.5% were cytologically classified as carcinoembryonic antigen (CEA), and HAM-56 and DPAM with primary appendiceal neoplasms in 19 use of genetic analysis by polymerase chain reac- cases. Thirty-four of 63 cases (53.9%) were cytolog- tion techniques have added strong support to the ically diagnosed as PMCA based on PW cytology. interpretation that almost all cases of PMP originate Most were of appendiceal or colonic origin. Four from primary appendiceal neoplasms (9–14) Spo- cases displayed cytologic features of both DPAM radic cases associated with tumors of the breast and PMCA. Two discordant cases each with a cyto- (15), gallbladder and bile duct (16), pancreas (17), logic diagnosis of PMCA had an appendiceal ade- stomach and intestines (18, 19), lung (20), and uri- noma. Acellular mucin alone was identified in the nary bladder (18) have been reported. PW in four cases. Analysis of follow-up data re- In a pivotal study, Ronnett et al. (21) suggested vealed that cases diagnosed as DPAM had a better that the term PMP, to date, has been used inappro- prognosis than those diagnosed as PMCA. Conclu- priately to describe a clinical entity arising from a sions: Cytomorphologic features of epithelial cells in morphologically heterogeneous group of benign PW material can accurately categorize cases of PMP and malignant pathologic lesions. They proposed as either DPAM or PMCA. Furthermore, this cate- abandoning PMP as a histologic diagnosis with sep- gorization appears to have important prognostic aration of the clinical entity of GA into two histo- implications. logic categories, disseminated peritoneal adenomu- cinosis (DPAM) and peritoneal mucinous carcinomatosis (PMCA) based on the histologic ap- pearance of the neoplastic epithelium of the peri- Copyright © 2001 by The United States and Canadian Academy of Pathology, Inc. toneal lesions and the underlying primary tumor. VOL. 14, NO. 7, P. 664, 2001 Printed in the U.S.A. Date of acceptance: March 10, 2001. Cases diagnosed as DPAM demonstrated abundant Address reprint requests to: Kim R. Geisinger, M.D., Department of Pa- mucin containing scant strips of mucinous epithe- thology, Wake Forest University Baptist Medical Center, Medical Center Boulevard, Winston-Salem, NC 27157; e-mail: [email protected]; fax: lium with minimal to moderate cytologic atypia 336-716-7595. and rare mitoses. The glandular epithelium was 664 simple and nonstratified to focally proliferative. these patients has previously been reported (31). Cases diagnosed as PMCA had more abundant ep- PW fluid was collected intraoperatively and smears ithelium demonstrating marked cytologic atypia. of the gelatinous material were fixed in 95% ethanol This epithelium formed malignant glands or nests for Papanicolaou staining. In 56 of 67 cases, mate- or occurred as individual malignant cells including rial was also available for cell blocks; it was fixed in signet ring cells. An intermediate category was also 10% buffered formalin, embedded in paraffin, sec- utilized to indicate specimens that demonstrated tioned, and stained with hematoxylin and eosin. histologic features of both DPAM and PMCA as well Histopathologic specimens were fixed in 10% buff- as cases showing discordant features (21). These ered formalin, processed routinely, and stained cases were diagnosed as PMCA based on the find- with hematoxylin and eosin. ing of at least focal areas with more typical carci- All cytologic smears and corresponding cell nomatous epithelium. Most cases with dual histo- blocks were examined in a blinded fashion by one logic features were derived from well-differentiated of the authors (KRG). Based on cytologic features appendiceal or intestinal mucinous adenocarcino- alone, one of four diagnoses was made: 1) acellular mas. Clinically, cases of DPAM were associated with mucin only; 2) DPAM; 3) PMCA; or 4) features of superficial implants, rare parenchymal or nodal in- both DPAM and PMCA. Cytologic features that volvement and a more benign or indolent course were specifically evaluated included presence or characteristic of the clinical syndrome of PMP. absence of glandular epithelium, size and arrange- Cases of PMCA were associated with more exten- ment of cellular groups, variation in cell size and sive lymph node involvement, parenchymal inva- shape, nuclear size and shape, nuclear chromatin sion, and metastases outside the abdominal cavity. distribution, hyperchromasia, presence or absence The value of the categorization proposed by Ron- of nucleoli, nuclear-to-cytoplasmic ratio, presence nett was reflected in a statistically significant differ- or absence of mitoses and single neoplastic epithe- ence in survival between patients with DPAM, and lial cells and the presence of extracellular mucin. PMCA (21). The cytomorphologic features useful for diagnosing Other notable findings in this study included the DPAM were variably sized, cohesive clusters or following: 1) histologically identical cases of DPAM monolayered (honeycomb) sheets of cells with dis- were associated with a variety of appendiceal neo- crete cell borders, uniform small, round nuclei with plasms; 2) prognosis could be predicted from the smooth nuclear membranes and inconspicuous appearance of the peritoneal lesions alone in cases nucleoli. Mitotic figures and individual malignant with unknown primaries; 3) most recurrent lesions neoplastic cells were not identified. Features for the displayed the same histologic features seen in the cytologic diagnosis of PMCA included individual primary tumor (DPAM only rarely progressed to cells, small rounded three-dimensional clusters, or PMCA); 4) in all cases of DPAM and PMCA with occasional, irregular sheets of cells with enlarged, unknown primary tumors, the appendix was abnor- overlapping nuclei with irregular nuclear mem- mal; 5) no cases of unequivocal ovarian origin were branes, irregular chromatin distribution, variably identified; and 6) the clinical condition of GA was sized nucleoli, and malignant signet ring cells. Mi- almost always associated with a primary tumor of totic features were identified in some cases of the GI tract (21). PMCA (see Fig. 1). Although the histologic features of PMP have After the cytologic diagnosis was established, the been well studied, the cytologic features remain ill histologic sections were reviewed to determine the defined (22–30). We reviewed the cytologic features categorization of the peritoneal lesion (DPAM ver- of peritoneal washings (PW) collected intraopera- sus PMCA) using the criteria established by Ronnett tively and the corresponding histology in 67 cases et al. (21). The histologic specimens were also ex- of PMP to 1) identify cytomorphologic features use- amined to determine the site(s) of origin of PMCA ful in distinguishing between DPAM and PMCA and DPAM. In women in whom a primary tumor (21); 2) determine the sites of origin of DPAM and could not be identified and cases in which only an PMCA; and 3) evaluate the prognosis of DPAM and ovarian tumor was identified, immunohistochemi- PMCA. cal staining for CK7 and 20 was selectively applied to peritoneal implants and the ovarian tumor, when available, to determine the site of origin (9, 10). MATERIALS AND METHODS Positive staining for both CK7 and 20 or CK7 posi- tivity with negative staining for CK20 was consid- Cytologic smears from PW and corresponding ered consistent with origin from a primary ovarian histologic specimens from 67 patients clinically di- tumor. Positive staining for CK20 and negative agnosed with PMP were retrieved from