Document Kzb3Zq5w3rVjen08K4vB2Qv9Q

378 Diffuse Malignant Epithelial Mesotheliomas of the Peritoneum in Women A Clinicopathologic Study of 25 Patients Scott A. J. Kerrigan, m.d.1,2 Rael T. Turnnir, m.d.1,2 Philip B. Clement, m.d.1,2 Robert H. Young, m.d.3,4 Andrew Churg, m.d.1,2 1 Department of Pathology, Vancouver Hospital and Health Sciences Center, Vancouver, British Columbia, Canada. 2 Department of Pathology, University of British Columbia, Vancouver, British Columbia, Canada. 3 James Homer Wright Pathology Laboratories, Massachusetts General Hospital, Boston, Massa chusetts. 4 Department of Pathology, Harvard Medical School, Boston, Massachusetts. Address for reprints: Andrew Churg, M.D., Depart ment of Pathology, University of British Columbia, 2211 Wesbrook Mall, Vancouver, British Columbia, Canada V6T 2B5; Email: achurg@interchange.ubc.ca Received July 11,2001; revision received Septem ber 18, 2001; accepted September 21, 2001. 2002 American Cancer Society DOI 10.1002/cncr.10209 BACKGROUND. The behavior of diffuse peritoneal mesotheliomas in women and the possible relation between tumor morphology and outcome are uncertain. Reported survival has ranged from < 1 month to > 14 years, and a previous study found that tumor morphology could not be used reliably for predicting outcome. The authors examined the behavior of diffuse epithelial peritoneal mesotheliomas in women and the possible relation between pathologic features and outcome. METHODS. Twenty-five female patients with diffuse peritoneal epithelial malignant mesotheliomas were divided into two groups: those who survived for < 4 years (60%) and those who survived for > 4 years (40%). Both groups were compared in terms of age, presentation, treatment, survival, tumor architecture, mitotic rate, necrosis, nuclear grade, and immunohistochemical profile. RESULTS. Patients in the two groups were similar in terms of age at diagnosis (median ages, 50.7 years and 49.9 years), presentation, initial tumor burden, and treatment. In both groups, the most common initial clinical presenting features were ascites and abdominal pain. The tumors typically took the form of multiple nodules measuring < 1.5 cm in greatest dimension. Slightly less than 50% of patients in both groups received some form of chemotherapy or radiation therapy after undergoing tumor-reductive surgery or biopsy. Overall survival ranged from 1 month to 15 years. The median survival was 12 months in the group of women who survived for < 4 years and 7 years in the group of women who survived for > 4 years. Overall, 10 of 25 patients survived for a 5 years. One patient was alive with disease 15 years after diagnosis. Although there was a suggestion that the tumors in patients with short survival more often had solid architecture and high-grade nuclei, these findings were not significant statistically. The frequency of necrosis and the mitotic activity were the same in both groups. CONCLUSIONS. The spectrum of diffuse epithelial peritoneal mesotheliomas in women includes tumors that are highly aggressive and behave much like pleural mesotheliomas, although a sizeable number of tumors, unlike the pleural tumors, are relatively indolent. However, because there do not appear to be morphologic features that reliably identify favorable tumors versus unfavorable tumors, aggres sive therapy for all women with diffuse peritoneal mesotheliomas may be war ranted. Cancer 2002;94:378-85. 2002 American Cancer Society. KEYWORDS: peritoneum, female, mesothelioma, epithelial mesotheliomas. iffuse malignant mesotheliomas are rare tumors, with an inci Ddence of about 2500 per year in the United States.1 Peritoneal mesotheliomas account for 17-32% of mesotheliomas in females,2-5 and the majority of lesions are epithelial forms.3,6 Diffuse malignant mesotheliomas must be separated from well-differentiated papillary mesotheliomas, which usually are solitary and indicate an excellent Mesotheliomas of the Peritoneum in Women/Kerrigan et al. 379 prognosis.7-9 The prognosis for most patients with diffuse peritoneal epithelial mesotheliomas is less clear. Survival data reported in the literature have ranged from < 1 month to > 14 years, observations that contrast with the uniformly dismal survival re ported in patients with pleural mesotheliomas. Many of the studies that have evaluated prognos tic factors in patients with peritoneal mesotheliomas have focused on clinical factors;10-12 however, in a series of 19 women with peritoneal mesothelioma, Goldblum and Hart13 found that tumors with pure epithelial growth and particularly those with a pre dominately tubulopapillary pattern and low-grade nu clei tended to behave in a benign manner. However, there was considerable overlap in the histologic fea tures of the benign and malignant tumors, suggesting that histology was not a reliable prognostic factor. Indeed, those investigators found that the most reli able predictor of survival was the distribution of the tumor. Patients with solitary tumors had an excellent prognosis, whereas patients with diffuse tumors (with more than one site involved) had a poor prognosis. Our study of 25 women with diffuse peritoneal epithe lial mesothelioma compares the histologic features of short-term and long-term survivors in an attempt to elucidate useful prognostic features. MATERIALS AND METHODS All patients came from a group of 36 women from the authors' consultation files. Clinical information was obtained from the referring pathologists, clinicians, medical records, or a combination of these sources. Survival was calculated from the time of diagnosis to the time of death or last follow-up. The criteria for inclusion were 1) agreement among the authors that these were mesotheliomas, which required the pres ence of typical monotonous, cuboidal-to-flattened mesothelial cells arranged in a histologic pattern char acteristic of mesothelioma accompanied by the proper histochemical and immunohistochemical pro file (see below); 2) epithelial histology; 3) multifocal (two or more sites) peritoneal involvement; and 4) follow-up data extending either to the death of the patient or to at least 4 years after diagnosis. Twentyfive women satisfied our inclusion criteria. Two women (Patients 2 and 24) were reported previously in a series of nine patients with mesothelioma involving the ovary.14 The excluded patients included four pa tients with epithelial tumors without at least 4 years of follow-up data, two patients with solitary epithelial mesotheliomas, four patients with well-differentiated papillary mesotheliomas, and one patient with sarco matoid mesothelioma. No biphasic tumors were en countered. Well-differentiated papillary mesothelio mas consisted entirely of papillae lined by a layer of uniform cuboidal cells with bland cytologic features, no more than a rare mitotic figure (MF), and no evi dence of invasion. The number of sections from each tumor ranged from 1 section to 20 sections (mean, 10 sections). Nu clear grade was assessed using a previously reported grading scale,13 with Grade 1 indicating tumors with small nuclei, a uniform chromatin pattern, and incon spicuous nucleoli; Grade 2 indicating tumors with larger nuclei, some chromatin irregularity, and more promi nent nucleoli; and Grade 3 indicating tumors with large nuclei, an irregular chromatin pattern, and prominent macronucleoli. The number of MFs per 50 high-power microscopic fields (HPFs) was counted on a microscope for which the greatest dimension of the HPF was 0.35 mm. Sections stained by the periodic acid-Schiff (PAS) reaction before and after diastase and immunohistochemical methods were available for 21 tumors. Most of these tumors predated the introduction of immunohistochemical markers most useful for mesotheliomas, such as calretinin, thrombomodulin, and cytokeratins 5 and 6. Immunohistochemical staining employed the strepavidin-biotin method and antibodies directed against cytokeratin (Z622; Dako, Carpinteria, CA; 1:4000 dilution), epithelial membrane antigen (EMA) (M613; Dako; 1:400 dilution), carcinoembryonic antigen (CEA) (polyclonal;Dako;1:1000 dilution), Ber-EP4 (Dako;1:10 dilution), S100 (HSC;1:8000 dilution), placental alkaline phosphatase (PLAP) (Zymed Laboratory, Inc., San Fran cisco, CA;1:100 dilution), Leu-M1 (Becton Dickson, San Jose, CA;1:10 dilution), TAG-72 (Signet;1:100 dilution), and vimentin (BioGenex, San Ramon, CA;1:1000 dilu tion). Statistical analysis was conducted using Fisher ex act tests and chi-square tests. RESULTS Clinical Features There were 15 short-term survivors and 10 long-term survivors. The mean ages of the short-term and long term survivors were 50.7 years (range, 24-73 years) and 49.9 years (range, 21-85 years), respectively (Table 1). Details of clinical presentation were available for 10 of the short-term survivors, 6 of whom presented with as cites, 2 with abdominal or pelvic pain, 1 with fever and weight loss, and 1 with an incidental asymptomatic ad nexal mass. Details of clinical presentation were avail able for 9 of the long-term survivors, 4 of whom pre sented with abdominal pain (including 1 with melena and 1 other with an autoimmune hemolytic anemia), 3 with ascites, 1 with fever and lymphadenopathy, and 1 who presented during the course of treatment for endo metriosis. Twelve of 15 short-term survivors were known to have multiple foci of tumor at initial laparotomy; in 380 CANCER January 15, 2002 / Volume 94 / Number 2 TABLE 1 Clinical Features, Therapy, and Follow-Up of 25 Female Patients with Peritoneal Mesothelioma Patient Age (yrs) Clinical presentation Sites Therapy Survival Short-term survivors 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Long-term survivors 16 17 18 19 20 21 22 23 24 25 24 31 37 40 43 47 31 56 57 59 59 66 69 69 73 21 30 32 41 85 61 64 65 63 37 Ascites Abdo pain, ascites Abdo girth, ascites Abdo girth, ascites Abdo pain, ascites N/A Ascites N/A Asympt adnexal mass Abdo, pelvic pain N/A N/A Abdo pain, constipation Fever, weight loss N/A Abdo pain, melena Abdo pain Laparoscopy for endo Abdo pain Ascites Fever, lymphadenopathy Abdominal pain, AIHA N/A Ascites Ascites Omentum, liver, small bowel Bilateral ovaries, omentum Omentum, peritoneum Bowel, left adnexa Widespread implants N/A Bowel, peritoneum Widespread implants Widespread implants Widespread implants Widespread implants Widespread implants Omentum Widespread implants N/A Widespread implants Widespread implants Widespread implants Ladnexa, bladder, omentum Widespread implants Ovaries, peritoneum Ovaries, omentum Widespread implants Widespread implants Widespread implants N/A TAHBSO, oment N/A LSO, peritoneal BX TAH, LSO, chemo N/A BX, chemo N/A TAHBSO, oment, BX, rad TAHBSO, chemo, rad TAHBSO, oment BX, chemo TAHBSO, oment, tamox BX N/A TAHBSO, oment, chemo BX TAHBSO, oment, rad TAHBSO TAHBSO, oment, chemo Tamox, meg TAHBSO, oment TAHBSO TAHBSO, chemo BX, chemo BX, chemo 12 months, DOD 10 months, DOD 12 months, DOD 3 months, DOD 6 months, DOD 11 months, DOD 30 months, DOD 1 month, DOD 44 months, DOD 24 months, DOD 2 months, DOD 6 months, DOD 24 months, DOD 12 months, DOD 19 months, DOD 13 yrs, CFOD 12 yrs, CFOD 14 yrs, CFOD 13 yrs, CFOD 7 yrs, DWD 7 yrs, CFOD 5 yrs, CFOD 5 yrs, AWD 7 yrs, AWD 15 yrs, AWD Abdo: abdominal; AIHA: autoimmune hemolytic anemia; Asympt: asymptomatic; AWD: alive with disease; BX: biopsy; CFOD: clinically free of disease; chemo: chemotherapy; DOD: died of disease; endo: endometriosis; LSO: left salpingoophorectomy; meg: megesterol; N/A: not available; oment: omentectomy; rad: radiation therapy; RSO: right salpingoophorectomy; TAHBSO: total abdominal hysterectomybilateral salpingoophorectomy; tamox: tamoxifen. the remaining 3 patients, this information was not avail able. All of the long-term survivors had multifocal tu mors. Surgical treatment was extremely variable. Some patients underwent limited peritoneal biopsy, and others underwent more extensive surgery. Generally, it was not clear from the information provided whether patients who underwent more extensive sur gery, such as total abdominal hysterectomy and bilat eral salpingoophorectomy, also underwent aggressive debulking of peritoneal tumors. The use of chemo therapy and radiotherapy also varied, and information on the duration of treatment was available for only one patient. In the group of short-term survivors, five patients had some form of postoperative chemother apy, including two patients who also received abdom inal radiation therapy and one patient who also re ceived tamoxifen treatment. One patient was treated with intraperitoneal cisplatin and intravenous doxo rubicin. A second patient was treated with paclitaxel. A third patient was treated with intravenous melphalan and abdominal radiotherapy. A fourth patient was treated with cisplatin, doxorubicin, cytarabine, and dexamethasone. A fifth patient was treated with intra venous cisplatin, doxorubicin, and cytarabine and also was on adjuvant tamoxifen treatment for a prior car cinoma of the breast. In the group of long-term sur vivors, five patients received some form of postoper ative chemotherapy, including one patient who also received treatment with tamoxifen and megesterol. For one patient, we could not determine which che motherapeutic agents were used. A second patient was treated with five cycles of cisplatin, cytarabine, and melphalan and also was on tamoxifen and meges terol. A third patient was treated with intravenous bleomycin, etoposide, and doxorubicin. A fourth pa tient was treated with cytarabine and carboplatin. A fifth patient was treated with intravenous doxorubicin. Mesotheliomas of the Peritoneum in Women/Kerrigan et al. 381 A sixth patient received postoperative abdominal ra diotherapy without chemotherapy. Survival Data The overall median survival for both groups combined was 30 months (mean, 56 months) and ranged from 1 month to 15 years. The median survival for the group of short-term survivors was 12 months (mean, 16.2 months;range, 1-44 months). All of these patients died of disease. Ten patients (67%) survived for 1 year, but only two patients (13%) survived for > 2 years. The median survival for the group of long-term survivors was 7 years (mean, 10.3 years;range, 5-15 years). None of these patients died of disease, although one patient died with disease at 7 years. Two patients were alive with disease at 7 years and 15 years after diag nosis. The patient who was alive at 15 years had re current pleural effusions with positive cytology. At the time of last follow-up, seven patients were alive and clinically free of disease, including one patient at 14 years after diagnosis. Macroscopic Pathology There was limited information concerning the macro scopic appearance of these tumors, and some infor mation on these features was available for only half of the patients. Most tumors were described as white, tan, pink, or brown nodules varying in size from 0.5 cm to 1.5 cm. In one patient, the tumor nodules mea sured up to 5 cm. In another patient, the peritoneum was studded with minute foci of tumor resembling grains of sand. In two patients, the tumors were more plaque-like and encased portions of viscera. In the two patients with significant ovarian involvement, tumor enlarged the ovaries up to 9 cm in maximum dimen sion. Some of the tumors were papillary, and one tumor was cystic. The tumors most often were firm, but several were soft. None was necrotic macroscopically. Microscopic Findings Seven tumors in the group of short-term survivors and three tumors in the group of long-term survivors had a predominantly solid architecture. None of the tu mors had a deciduoid morphology (Table 2).9,15 The remaining tumors in both groups had tubulopapillary architecture (Figs. 1, 2). The difference in architecture between the two groups was not statistically signifi cant (P = 0.68). In the group of short-term survivors, one tumor was Grade 1, eight tumors were Grade 2, and six tumors were Grade 3. None of the tumors in the group of long-term survivors had Grade 1 nuclei, seven tumors were Grade 2 and three tumors were Grade 3. The nuclear grades of the tumors in the two TABLE 2 Microscopic Features of Tumors Microscopic feature Short-term survivors (%) Predominantly tubulopapillary Predominantly solid Necrosis Nuclear grade 1 2 3 8 of 15 (53) 7 of 15 (47) 1 of 15 1 of 15 8 of 15 6 of 15 Long-term survivors (%) 7 of 10 (70) 3 of 10 (30) 1 of 10 0 of 10 7 of 10 3 of 10 groups were not statistically different (chi-square value, 1.11;P = 0.57). The mean mitotic rates for the groups of short-term and long-term survivors were 1.47 MFs per 50 HPFs (range, 0-11 MFs per 50 HPFs) and 0.67 MFs per 50 HPFs (range, 0-5 MFs per 50 HPFs), respectively. The mitotic rate for short-term survivors was inflated by a single patient who had 11 MFs per 50 HPFs; if this patient is excluded, then the mean mitotic rate for this group becomes 0.79 MFs per 50 HPFs (range, 0-5 MFs per 50 HPFs). In nine tumors from the short-term survivors and in 7 tumors from the long-term survivors, no MFs were seen. The re maining two tumors in the latter group had mitotic counts of 1 MFs per 50 HPFs and 5 MFs per 50 HPFs, respectively (in one tumor that was not included in the mitotic rate calculation for this group because < 50 HPFs were available for assessment, no mitoses were seen). Only one tumor from each group showed necrosis. Multiple foci of necrosis were seen in the tumor of a patient who died of disease at 10 months, and a single microscopic focus was found in a patient who was clinically free of disease 7 years after diagno sis. Most tumors in both groups showed mild-to-moderate inflammatory responses. Almost all tumors showed lymphocytic infiltration, some with germinal center formation. However, the inflammatory infil trate in one patient in the group of short-term survi vors consisted predominantly of foamy macrophages, and, in one patient in the group of long-term survi vors, the infiltrate consisted predominantly of neutro phils. Special Stains and Immunohistochemistry PAS with (PAS/D) and without diastase staining and immunohistochemistry were available for 13 short term survivors and for 8 long-term survivors (Table 3). All tumors for which PAS and PAS/D staining was available were PAS/D negative. All tumors were strongly positive for keratin. Vimentin was at least focally positive in 12 short-term survivors and 4 long term survivors. S100 was weakly positive in four short- 382 CANCER January 15, 2002 / Volume 94 / Number 2 FIGURE 1. Tubulopapillary architecture. (a) Low-power view of tumor tissue from a short-term survivor (original magnification, X160). (b) High-power view of tumor tissue from a short-term survivor (original magnification, X320). (c) Low-power view of tumor tissue from a long-term survivor (original magnification, X80). (d) High-power view of tumor tissue from a long-term survivor (original magnification, X320). Note the extensive tumor invasion in both short-term and long-term survivors. term survivors and two long-term survivors. EMA was positive in nine and two of the patients, respectively, with no consistent differences in membranous or cy toplasmic distribution. All tumors were negative for Leu-M1, CEA, B72.3, and Ber-EP4. DISCUSSION This is the largest study of diffuse peritoneal epithelial mesotheliomas in women reported to date. An arbi trary cut-off value of 4 years was chosen to separate short-term and long-term survivors, because almost all patients with pleural mesotheliomas are dead at 4 years; thus, this cut-off time should show whether epithelial peritoneal mesotheliomas behave differ ently. Our results indicate that, unlike their pleural counterparts, a substantial proportion of diffuse peri toneal epithelial mesotheliomas in women behave in dolently. In our series, 40% of patients survived for at least 4 years, with one patient alive with disease 15 years after diagnosis and another patient alive and clinically free of disease at 14 years. The 15-year sur vival is the longest reported to date. In contrast, the 67% 1-year survival rate and the 13% 2-year survival rate in the group of short-term survivors are similar to the survival rates reported for patients with pleural mesothelioma: approximately 50% at 1 year with few patients surviving for 2 years.2,3 Could our results be an artifact of how patients were selected? One could argue that patient selection was compromised by a failure to match clinical fea tures, such as tumor bulk and treatment, that appear to be of some importance in the clinical literature.10-12 The most recent and largest series of peritoneal me sotheliomas reported on 33 patients, 10 of whom were Mesotheliomas of the Peritoneum in Women/Kerrigan et al. 383 FIGURE 2. Solid architecture with large sheets of tumor. (a) Low-power view of tumor tissue from a short-term survivor (original magnification, X160). (b) High-power view of tumor tissue from a short-term survivor (original magnification, X320). (c) Low-power view of tumor tissue from a long-term survivor (original magnification, X80). (d) High-power view of tumor tissue from a long-term survivor (original magnification, X320). women, and found that female gender, more limited prior pelvic or abdominal surgery, complete tumorreductive surgery, and the performance of a secondlook surgery were positive predictive factors for survival.10 The median survival was 31 months, but the relative survival rates for men and women were not specified. Antman et al.11 reported long-term disease free survival with up to 36 months follow-up in six patients (three women) who were treated aggressively with tumor-reductive surgery, postoperative chemo therapy, and whole abdominal radiation. Those au thors also found that age < 40 years conferred a much better prognosis. Eltabbahkh et al.12 described 15 women with peritoneal tumors and noted that tumorreductive surgery and chemotherapy both improved survival, but their median survival was still 12.5 months, and only 5 patients survived for > 1 year. The amount of clinical information available to us varied greatly between patients, which often is the situation with consultation patients. However, the two groups of patients appear to be very similar in terms of patient age at presentation, initial tumor burden, and treatment, and there is no evidence that these clinical features accounted for the observed differences in sur vival. Furthermore, because some of the studies men tioned above have shown that treatment confers mar ginal improvement in survival at best, it seems unlikely that any unknown differences in treatment that may exist, such as extent of tumor debulking or duration of chemotherapy, would account for the ob served survival difference between our two groups. The unintentional inclusion of tumors other than mesotheliomas is another important source of potential bias. Mesotheliomas must be distinguished from far more common lesions in the differential diagnosis, such as atypical mesothelial hyperplasia, primary and meta- 384 CANCER January 15, 2002 / Volume 94 / Number 2 TABLE 3 Immunohistochemical Staining Results Immunohistochemical Short-term survivors Keratin Vimentin LeuM1 S100 CEA EMA B72.3 Ber-EP4 13 of 13 12 of 13 0 of 13 4 of 13 0 of 13 9 of 13 0 of 13 0 of 13 CEA: carcinoembryonic antigen; EMA: epithelial membrane antigen. Long-term survivors 8 of 8 4 of 8 0 of 8 2 of 8 0 of 8 2 of 8 0 of 8 0 of 8 static serous carcinomas, ectopic decidua, and malig nant vascular tumors of the peritoneum.16 A recent pa per by the United States-Canadian Mesothelioma Reference Panel reviewed in detail the histologic features that are useful in distinguishing benign mesothelial re actions from malignant mesothelial neoplasms.17 Histo logic features that favor the diagnosis of mesothelioma over serous carcinoma include prominent tubulopapillary pattern, polygonal cells with eosinophilic cytoplasm, the absence of marked nuclear pleomorphism, the ab sence of high mitotic rate, and the presence of intracel lular acid (PAS negative) mucin rather than neutral (PAS positive) mucin.16 Several studies have reviewed the use of immunohistochemical markers in distinguishing me sotheliomas from serous adenocarcinomas.18-20 Most of the patients in our series predated specific immunohistochemical markers for mesothelioma and were consul tation patients from a large number of geographically dispersed sites. Therefore, it was not feasible to perform staining for newer markers, such as calretinin, thrombo modulin, and cytokeratins 5 and 6, which typically are positive in patients with mesotheliomas and negative in patients with serous carcinomas.18 Some authors have used ultrastructural features, as determined by electron microscopy, to distinguish mesotheliomas from carcino mas13,20; however, given the availability of immunohistochemical methods and a general lack of suitable ma terial on which to perform electron microscopy in our patients, this technique was not used in this study. De spite these considerations, we are confident that, among our patients, all tumors were mesotheliomas and not carcinomas based on their histologic appearance, lack of PAS staining, and the available immunohistochemistry. Another potential source of bias may have been the unintentional inclusion of well-differentiated pap illary mesotheliomas. Solitary, well-differentiated pap illary mesotheliomas are benign,13 whereas their be havior when multifocal is less certain, but at least some of these appear to be extremely indolent.9 We avoided this source of bias by excluding tumors that consisted entirely of papillae lined by a layer of uni form cuboidal cells with bland cytologic features, no more than a rare MF, and no evidence of invasion. The tumors in our long-term survivors were Grade 2 or 3, which excluded well-differentiated papillary mesothe liomas by definition. Therefore, we believe that all of the tumors in our study, including those in the group of long-term survivors, were diffuse epithelial malig nant mesotheliomas. Some of our findings are consistent with those of Goldblum and Hart.13 Their study of 19 patients with peritoneal mesothelioma included 17 purely epithelial tumors, 11 of which were diffuse. Those authors found that there was considerable overlap in the histologic features of the benign and malignant tumors but that tumors with a predominantly tubulopapillary pattern and low-grade nuclei tended to be less aggressive. Similarly, we found that the patients with short sur vival more often had tumors with solid architecture and high-grade nuclei, but there was considerable overlap between the two groups of patients, and the differences were not statistically significant. Necrosis, which correlates with an aggressive behavior in many types of tumors, was not useful, because only one tumor in each group showed any necrosis. Similarly, our mitotic counts, which were in keeping with others reported previously,13,14 failed to provide any correla tion with survival, an observation that has been made previously in patients with malignant mesothelioma in other sites.17 A major difference between our study and that of Goldblum and Hart13 was that, whereas they included solitary tumors, we limited ourselves to diffuse tumors. Perhaps not surprisingly, those au thors found that the most reliable predictor of survival was the distribution of the tumor: Patients with soli tary tumors had an excellent prognosis, and patients with diffuse tumors had a poor prognosis. By exclud ing patients with solitary tumors, we avoided this form of selection bias, because those patients would be expected to have longer survival than patients with diffuse tumors. Even so, survival in our group of long-term survivors was much better compared with the study by Goldblum and Hart.13 In part, this may reflect differences in patient selection, because two of their patients had tumors with sarcomatous com ponents, and sarcomatous tumors tend to be ag gressive. However, for the most part, we cannot account for the differences in patient survival be tween the two studies. In conclusion, we found that diffuse peritoneal epithelial mesotheliomas in women exhibit a varied spectrum with regard to outcome. Although many pa tients have a rapid course, a sizeable number have a Mesotheliomas of the Peritoneum in Women/Kerrigan et al. 385 relatively prolonged survival and do not die of their disease. No difference in tumor morphology was clearly evident between groups of patients with poor and favorable outcomes. To us, this implies that all women with peritoneal mesotheliomas should be treated for long-term cure and long-term survival. The proper treatment for these women remains controver sial: To date, there is no standardized therapy for patients with mesothelioma, although limited evi dence suggests that tumor-reduction surgery and che motherapy may improve survival.11,12 Accordingly, given our inability to predict outcome with routine histology, it may be appropriate to proceed with treat ment in all women with diffuse epithelial peritoneal mesotheliomas. REFERENCES 1. McDonald AD, McDonald JC. Epidemiology of malignant mesothelioma. In: Antman K, Aisner J, editors. Asbestosrelated malignancy. Orlando: Grune & Stratton, 1987:31-55. 2. Spitas R, Connelly RR, Tucker MA. Survival patterns of ma lignant mesothelioma: The SEER experience. Int J Cancer 1988;41:525-30. 3. Antman K, Shemin R, Ryan L, Klegar K, Osteen R, Herman T, et al. Malignant mesothelioma: prognostic variables in reg istry of 180 patients, the Dana-Farber Cancer Institute and Brigham and Women's Hospital experience over two de cades, 1965-1985. J Clin Oncol 1988;6:147-53. 4. Dawson A, Gibbs AR, Pooley FD. Malignant mesothelioma in women. Thorax 1993;48:269 -74. 5. Roggli VL, Oury TD, Moffatt EJ. Malignant mesothelioma in women. Anat Pathol 1997;2:147- 63. 6. Kannerstein M, Churg J. Peritoneal mesothelioma. Hum Pathol 1977;8:83-94. 7. Foyle A, Al-jabi M, McCaughey WET. Papillary peritoneal tumors in women. Am J Surg Pathol 1981;5:241-9. 8. Hoekman K, Tognon G, Risse EK, Bloemsma CA, Vermorken JB. Well-differentiated papillary mesothelioma of the peri toneum: a separate entity. Eur J Cancer 1996;32:255- 8. 9. Daya D, McCaughey WET. Well-differentiated papillary me sothelioma of the peritoneum. A clinicopathologic study of 22 cases. Cancer 1990;65:292-6. 10. Sebbag G, Yan H, Shmookler M, Sugarbaker PH. Results of treatment of 33 patients with peritoneal mesothelioma. Br J Surg 2000;87:1587-93. 11. Antman KH, Osteen RT, Klegar KL, Amato DA, Pomfert EA, Larson DA, et al. Early peritoneal mesothelioma: a treatable malignancy. Lancet 1985;8462:977-81. 12. Eltabbakh GH, Piver MS, Hempling RE, Recio FO, Intengen ME. Clinical picture, response to therapy, and survival of women with diffuse malignant peritoneal mesothelioma. J Surg Oncol 1999;70:6-12. 13. Goldblum J, Hart WR. Localized and diffuse mesotheliomas of the genital tract and peritoneum in women. A clinicopathological study of nineteen true mesothelial neoplasms other than adenomatoid tumors, multicystic mesotheliomas and localized fibrous tumors. Am J Surg Pathol 1995;19: 1124 -37. 14. Clement PB, Young RH, Scully RE. Malignant mesothelio mas presenting as ovarian masses. A report of nine cases, including two primary ovarian mesotheliomas. Am J Surg Pathol 1996;20:1067-80. 15. Covell LM, Disciullo AJ, Knapp RC. Decidual change in pelvic lymph nodes in the presence of cervical squamous cell carcinoma during pregnancy. Am J Obstet Gynecol 1997; 127:674-6. 16. Clement PB, Young RH. Atlas of gynecologic surgical pathol ogy. Philadelphia: W. B. Saunders Company, 2000:456. 17. Churg A, Colby TV, Cagle P, Corson J, Gibbs AR, Gilks B, et al. The separation of benign and malignant mesothelial proliferations. Am J Surg Pathol 2000;24:1183-200. 18. Ordonez NG. Role of immunohistochemistry in distinguish ing epithelial peritoneal mesotheliomas from peritoneal and ovarian serous carcinomas. Am J Surg Pathol 1998;22:120317. 19. Sheibani K, Esteban JM, Bailey A, Battifora H, Weiss LM. Immunopathologic and molecular studies as an aid to the diagnosis of malignant mesothelioma. Hum Pathol 1992;23: 107-16. 20. Bollinger DJ, Wick MR, Dehner LP, Mills SE, Swanson PE, Clarke RE. Peritoneal malignant mesothelioma versus serous papillary adenocarcinoma: a histochemical and immunohistochemical comparison. Am J Surg Pathol 1989;13:659-70.