Patients with pseudomyxoma peritonei associated with disseminated peritoneal adenomucinosis have a significantly more favorable prognosis than patients with peritoneal mucinous carcinomatosis

Abstract

BACKGROUND

Pseudomyxoma peritonei (PMP) is a poorly understood condition characterized by disseminated intraperitoneal mucinous tumors, often with mucinous ascites. The term PMP has been applied historically as a pathologic diagnostic term to both benign and malignant mucinous neoplasms that produce abundant extracellular mucin, resulting in a variable and poorly predictable prognosis. A recent study reported a pathologic classification that separated patients into prognostically distinct groups, but the follow-up was relatively short.

METHODS

Long-term follow-up data were analyzed for a previously reported series of 109 patients with PMP to examine the prognostic utility of a pathologic classification system that divided patients into three groups: disseminated peritoneal adenomucinosis (DPAM), peritoneal mucinous carcinomatosis (PMCA), and peritoneal mucinous carcinomatosis with intermediate or discordant features (PMCA-I/D). Patients whose tumors were classified 25 DPAM (n = 65 patients) had disease that was characterized by histologically bland to low-grade adenomatous mucinous epithelium associated with abundant extracellular mucin and fibrosis, often with an identifiable appendiceal mucinous adenoma that was the source of the peritoneal lesions. Patients whose tumors were classified 25 PMCA (n = 30 patients) had disease that was characterized by peritoneal lesions that displayed the cytologic and architectural features of mucinous carcinoma associated with extracellular mucin, often with an identifiable invasive mucinous adenocarcinoma of the gastrointestinal tract. Patients whose tumors were classified 25 PMCA-I (n = 11 patients) had peritoneal lesions that combined the features of DPAM and PMCA derived from well differentiated mucinous adenocarcinomas associated with adenomas. Patients whose tumors were classified 25 PMCA-D (n = 3 patients) had markedly atypical appendiceal adenomas associated with peritoneal lesions similar to PMCA.

RESULTS

Patients with DPAM had 5-year and 10-year survival rates of 75% and 68%, respectively (mean follow-up, 96 months; median follow-up, 104 months). Patients with PMCA and PMCA-I/D had a significantly worse prognosis, with 5-year and 10-year survival rates, respectively, of 50% and 21% for PMCA-I/D (mean follow-up, 58 months; median follow-up, 51 months) and 14% and 3% for PMCA (mean follow-up, 27 months; median follow-up, 16 months; P = 0.0001).

CONCLUSIONS

The term PMP should be used only as a clinical descriptor for patients who have the syndrome of mucinous ascites accompanied by a characteristic distribution of peritoneal mucinous tumors with the pathologic features of DPAM. DPAM should be used as a pathologic diagnostic term for patients with the bland peritoneal mucinous tumors associated with ruptured appendiceal mucinous adenomas and PMP. These patients should not be diagnosed with carcinoma, because they have disease that is distinct pathologically and prognostically from PMCA. Cancer 2001;92:85–91. © 2001 American Cancer Society.

The term pseudomyxoma peritonei (PMP) is applied most commonly to patients with a progressive disease characterized by disseminated, intraperitoneal, mucinous tumors; mucinous ascites; and a variable, poorly predictable prognosis. The association of the condition with mucinous tumors of the appendix, particularly mucinous adenomas, and with ovarian mucinous tumors in women is well described.1–3 PMP also is used as a pathologic diagnostic term that is applied to a diverse group of mucinous tumors derived from different sites that have in common only the presence of abundant extracellular mucin and may or may not contain epithelial cells. Thus, both tumors associated with mucinous adenomas of the appendix and mucin-producing gastrointestinal adenocarcinomas are included under the term PMP. In addition, there is considerable variability in the criteria and terminology used by different pathologists to diagnose patients with the same morphology. For example, some pathologists diagnose histologically benign appendiceal adenomas and similar peritoneal lesions as ruptured adenomas with PMP, whereas others diagnose them as well differentiated mucinous adenocarcinomas, because the tumor is disseminated throughout the peritoneal cavity. Consequently, survival analyses often are performed on a pathologically heterogeneous group of patients, because there is no uniformity in the pathologic reporting of these tumors. The lack of a specific pathologic definition for PMP has obscured the behavior of this condition.

We previously reported a clinicopathologic analysis of 109 patients with PMP characterized by multifocal peritoneal mucinous tumors who were treated uniformly by the same surgeon (P.H.S.).4 A classification system based solely on pathologic features was proposed that separated the patients into three groups: 1) patients with disseminated peritoneal adenomucinosis (DPAM), 2) patients with peritoneal mucinous carcinomatosis (PMCA), and 3) patients with peritoneal mucinous carcinomatosis with intermediate or discordant features (PMCA-I/D). In the original report, these groups demonstrated significantly different survival rates. The follow-up was sufficient to demonstrate the poor prognosis of patients with PMCA but not the long-term prognosis of patients with DPAM. We now report results of long-term follow-up that demonstrate significant differences in prognosis between patients with DPAM and patients with PMCA.

MATERIALS AND METHODS

All patients who were diagnosed with PMP or mucinous adenocarcinoma with multifocal peritoneal involvement were identified from a review of the patient files of one of the authors (P.H.S.) who surgically treated all of the patients during the period from 1983 to 1993. Patients with minimal sampling of the peritoneal lesions, minimal extracellular mucin in the peritoneal lesions, or lesions confined to the right lower quadrant were excluded from the analysis. The review included 109 patients in whom the peritoneal and serosal surfaces of the abdomen and pelvis were involved by lesions containing pools of dissecting mucin associated with fibrosis with or without epithelial cells. The surgical pathology reports and slides from all peritoneal lesions as well as those from any associated appendiceal or other gastrointestinal tumors or ovarian tumors removed prior to, at the time of, or shortly after the diagnosis of the peritoneal lesions were reviewed.

Patients were classified into three groups based on the pathologic features of their peritoneal lesions.4 The pathologic features are summarized below.

DPAM

The peritoneal lesions were comprised uniformly of scant, histologically bland to low-grade adenomatous, mucinous epithelium within abundant extracellular mucin associated with fibrosis (Fig. 1). The mucinous epithelium usually was nonstratified or demonstrated at most focal tufting and displayed minimal cytologic atypia and mitotic activity. Appendiceal adenomas were identified in 37 of 65 patients (57%), and there was frequent evidence of rupture. In a certain number of patients, however, a primary tumor was not identified, nor was a normal appendix found. In these patients, it was postulated that the appendix had ruptured and had been obliterated by the ensuing fibrosis. The extra-appendiceal lesions nearly always were confined to peritoneal surfaces, and only rarely were lesions identified within lymph nodes or invading the parenchyma of abdominal or pelvic organs. In addition, the peritoneal lesions demonstrated a characteristic anatomic distribution that has been described previously.5

Figure 1. Disseminated peritoneal adenomucinosis. (A,B) Peritoneal lesions consist of scant strips of histologically bland to low-grade adenomatous mucinous epithelium associated with abundant extracellular mucin and fibrosis. (C) A ruptured mucinous adenoma of the appendix is surrounded by pools of dissecting mucin that contain scant mucinous epithelium.

PMCA

The peritoneal lesions were comprised of more abundant, mucinous epithelium forming glands and/or signet ring cells and demonstrated sufficient cytologic atypia and architectural complexity to warrant a diagnosis of mucinous carcinoma (Fig. 2). Many of these lesions were associated with an identifiable primary mucinous adenocarcinoma of either the appendix or the colon; however, in some patients, a primary tumor was not identified. The appendiceal mucinous adenocarcinomas invariably demonstrated signet ring cell differentiation. Parenchymal organ invasion and/or lymph node metastases were often present.

Figure 2. Peritoneal mucinous carcinomatosis. A–C show three examples of peritoneal lesions derived from mucinous adenocarcinomas of the appendix. (A) The nuclei within a carcinomatous gland are hyperchromatic, atypical, and stratified. (B) Signet ring cell differentiation is evident in fragments of mucinous glandular epithelium. (C) Markedly atypical epithelial fragments are suspended in pools of extracellular mucin.

PMCA-I/D

Patients with tumors that were classified as PMCA-I included those in whom the peritoneal lesions demonstrated predominantly features of DPAM but also contained focal areas of well differentiated mucinous carcinoma (Fig. 3). For those with an identifiable primary tumor, the site of origin was a well differentiated mucinous adenocarcinoma of the appendix or, rarely, of the small intestine, and an associated mucinous adenoma was often present. Patients with disease that was classified as PMCA-D had lesions characterized by a markedly atypical adenoma of the appendix (high-grade dysplasia or intramucosal carcinoma) that was dilated without evidence of invasive carcinoma in the wall of the appendix. The peritoneal lesions surrounding the appendix and elsewhere were comprised uniformly of mucinous adenocarcinoma, often with signet ring cell differentiation and without areas of low-grade mucinous epithelium, as in the DPAM lesions or the patients with intermediate lesions. Follow-up information was available for 108 of the 109 patients.

Figure 3. Peritoneal mucinous carcinomatosis with intermediate features. (A) Well differentiated mucinous adenocarcinoma of the appendix infiltrates the appendiceal wall, giving rise to the peritoneal lesions illustrated in B and C. (B) An area that resembles disseminated peritoneal adenomucinosis contains strips of bland, mucinous epithelium associated with abundant extracellular mucin. (C) A focal area in which the mucinous epithelium displays nuclear stratification, hyperchromasia, loss of mucin, mitotic figures, and early cribriform growth is consistent with mucinous carcinoma.

Treatment

All patients were treated in the same fashion by the same surgeon (P.H.S.). These patients underwent a series of peritonectomy procedures and organ resections to maximally debulk (cytoreduce) the tumor. In the early postoperative period (postoperative Days 1–6), chemotherapeutic agents (mitomycin C and 5-fluorouracil) were instilled into the peritoneal cavity. Subsequently, three adjuvant cycles of systemic mitomycin C and intraperitoneal 5-fluorouracil were administered.6

Statistical Analysis

All analyses were performed using SAS software (version 6.12; SAS Institute Inc., Cary, NC). Ranges and frequency distributions of all continuous and categoric variables were examined. The analysis of variance test was used to compare the mean differences in age among the three groups. The chi-square test was used to compare the number of deaths in each group. Patient survival was analyzed according to the Kaplan–Meier method, with death due to disease as the endpoint. Statistical significance was tested by log-rank statistics (Fig. 4).

Figure 4. Kaplan–Meier survival curve illustrates the statistically significant differences in survival of patients with disseminated peritoneal adenomucinosis, peritoneal mucinous carcinomatosis with intermediate or discordant feature, and peritoneal mucinous carcinomatosis.

RESULTS

DPAM

This group (65 patients; 60%) was characterized by multifocal mucinous tumors adherent to serosal and peritoneal surfaces (Table 1). The condition was found more commonly in men than women (62% of men vs. 38% of women). In the original report, this group had a 5-year survival rate of 84% (mean follow-up, 55 months; median follow-up, 42 months). Based on the additional follow-up data, the 5-year and 10-year survival rates were 75% and 68%, respectively (mean follow-up, 96 months; median follow-up, 104 months). There were 20 deaths due to disease (31% of patients), with deaths occurring from 6 months to 144 months after diagnosis. The mean survival was 112 months. The median survival was not determined, because > 50% of patients were alive at the end of the follow-up period. Of the 40 patients who were alive, 34 patients were alive without evidence of disease for periods ranging from 95 months to 270 months. Six patients were alive with disease, and five patients had died of other causes.

PMCA

This group included 30 patients (28%). Follow-up information was available for 29 patients. In the original report, this group had a 5-year survival rate of 7% (mean follow-up, 20 months; median follow-up, 16 months). The long term follow-up confirmed the poor prognosis, as evidenced by 5-year and 10-year survival rates of 14% and 3%, respectively (mean follow-up, 27 months; median follow-up, 16 months). These survival rates were significantly worse than those for patients with DPAM (P = 0.0001). Twenty-eight of 29 patients died of disease (93%). The mean and median survival times were 27 and 16 months, respectively.

PMCA-I/D

This group (14 patients; 13%) included 11 patients with intermediate features and 3 patients with discordant features. In the original report, this group had a 5-year survival rate of 38% (mean follow-up, 45 months; median follow-up, 30 months). The additional follow-up demonstrated 5-year and 10-year survival rates of 50% and 21%, respectively (mean follow-up, 58 months; median follow-up, 51 months). These rates were significantly worse than those for patients with DPAM (P = 0.0001). Eleven of 14 patients died of disease (79%). These deaths included 9 of 11 patients with PMCA-I and 2 of 3 patients with PMCA-D. The mean and median survival times were 46 months and 51 months, respectively.

When these patients were grouped together with the PMCA patients (PMCA-I/D + PMCA), the 5-year and 10-year survival rates for all patients with carcinomatosis were 26% and 9% respectively. These rates were significantly worse than the rates for patients with DPAM (P = 0.0001). The mean and median survival times for all patients with carcinomatosis were 35 months and 22 months, respectively.

Table 1. Summary of Survival Analysis Comparing Patients with Disseminated Peritoneal Adenomucinosis and Patients with Peritoneal Mucinous Carcinomatosis

Characteristic	DPAM	PMCA-I/D	PMCA	P value
DPAM: disseminated peritoneal adenomucinosis; PMCA-I/D: peritoneal mucinous carcinomatosis with intermediate or discordant features; NED: no evidence of disease; AWD: alive with disease; DOC: dead of other causes. a There were 30 patients, 29 of whom were available for follow-up. b Results of an analysis of variance test. c Results of a chi-square test.
No. of patients	65	14	29a	—
Mean age (yrs)	49.1	47.5	47.1	0.74b
No. of deaths (%)	20 (30.8)	11 (78.6)	28 (93.3)	0.001c
Survival
Five-yr rate (%)	75	50	14	0.0001
Ten-yr rate (%)	68	21	3	0.0001
Mean (months)	112.4	46.2	26.5	—
Median (months)	—	50.5	16.0	—
Follow-up
Mean (months)	95.7	57.9	27.2	—
Median (months)	104.0	50.5	16.0	—
Status
NED	34	1	1	—
AWD	6	1	0	—
DOC	5	1	0	—

DISCUSSION

DPAM, as defined above, is a homogeneous pathologic entity characterized by histologically bland peritoneal mucinous tumors derived from ruptured appendiceal mucinous adenomas. When analyzed as a group, the patients with DPAM had a 10-year survival rate of nearly 70%. Morbidity and mortality are due to intra-abdominal mucus accumulation accompanied by fibrosis and complicated by small bowel obstruction. Some patients have long periods of disease-free survival after maximal cytoreduction by peritonectomy, whereas others develop multiple recurrences over months to years. Recurrences typically are confined to the peritoneum and display similarly bland histologic features. In contrast to DPAM, peritoneal lesions containing more abundant mucinous epithelium with the cytologic and architectural features of mucinous carcinoma should be classified as PMCA. They are associated with primary appendiceal or intestinal mucinous adenocarcinomas and are often characterized by parenchymal organ invasion and lymph node metastases rather than the diffuse, generally noninvasive peritoneal surface involvement seen in DPAM. The behavior of PMCA is significantly worse than that of DPAM. The patients with PMCA had a 10-year survival rate of < 5%. Morbidity and mortality are due to intra-abdominal carcinomatosis rather than the mucus accumulation, fibrosis, and bowel obstruction that occur in patients with DPAM.

Occasionally, patients are encountered in whom the peritoneal lesions display features intermediate between DPAM and PMCA. Other rare patients display a combination of an atypical adenoma and/or intramucosal carcinoma with PMCA (the discordant group). When analyzed as a group, patients with tumors that displayed intermediate or discordant features had significantly worse 5-year and 10-year survival rates compared with patients with DPAM. These survival rates reflect predominantly the behavior of the tumors with intermediate features, because they represented nearly 80% of tumors in the combined group. The intermediate tumors displayed pathologic features that were a hybrid of DPAM and PMCA in both the primary tumors and the peritoneal lesions. In particular, there was adenoma with invasive carcinoma in the primary tumors, and the peritoneal lesions reflected this by displaying a combination of bland mucinous tumor and well differentiated mucinous carcinoma. It is important to recognize this intermediate group, because the pathology closely resembles DPAM, but the behavior more closely resembles PMCA, particularly with longer follow-up. In the original classification, the discordant group had been separated from the other PMCA group. This was done because the discordant group had identifiable precursor lesions (adenoma with intramucosal carcinoma), often without an infiltrative component in the wall of the appendix, whereas the patients with PMCA who had an identified primary carcinoma had disease characterized by invasive carcinoma without an evident precursor lesion. In retrospect, it is probably best to consider the discordant tumors a variant of PMCA in which a precursor lesion happens to be evident because the peritoneal lesions were comprised uniformly of carcinoma, similar to the peritoneal lesions in patients with PMCA.

Some pathologists argue that the pathologic process we have labeled DPAM should be designated as well differentiated mucinous adenocarcinoma, because the mucinous tumor has spread beyond the appendix and recurs within the peritoneal cavity over time, sometimes leading to the death of the patient. However, the significant difference in the behavior of DPAM and PMCA with intermediate features argues against such an approach. The peritoneal lesions in PMCA with intermediate features demonstrate the cytologic and architectural features of well differentiated mucinous carcinoma (at least moderate nuclear atypia, mitotic figures, nuclear stratification, and cribriform growth) in combination with more bland areas, like what is seen in DPAM. When the primary tumor is identified, it is comprised of well differentiated adenocarcinoma that infiltrates the wall of the appendix. This histologic appearance contrasts with the bland appearance of the mucinous epithelium in DPAM. In addition, the appendiceal adenomas that are ruptured display similarly bland to low-grade adenomatous mucinous epithelium lining the lumen of the appendix, the rupture site, and/or the serosa without the infiltrative appearance of well differentiated adenocarcinoma. The wall adjacent to the rupture site is often attenuated, and dissecting mucin traverses this point with or without epithelium at the site of rupture. Thus, DPAM is a pathologic term that describes a specific clinicopathologic entity. Until now, this pathologic process and the clinicopathologically distinct cases of PMCA have been labeled with the nonspecific, poorly defined term PMP. It is therefore no surprise that the behavior of PMP was understood so poorly. To provide continuity with the historic use of the term PMP and the clinical syndrome of mucinous ascites, we advocate using the term DPAM as a pathologic diagnosis and including the phrase consistent with the clinical syndrome of PMP in the report.

Historically, the debate over site of origin of PMP has focused on women, because primary ovarian mucinous tumors exist and can rupture, and approximately half the women with PMP demonstrate ovarian involvement by mucinous tumor. Thus, whereas, in men, the gastrointestinal tract has always been considered the sole site of origin, in women, it was assumed that either the gastrointestinal tract or the ovaries were the sites of origin. However, recent morphologic, immunohistochemical, and molecular genetic studies have provided compelling evidence that all patients with DPAM have disease derived from an appendiceal mucinous tumor.1–3, 7–10 In our experience, we have not yet encountered a bona fide case of DPAM in which the ovary contained a mucinous tumor and an entirely normal appendix was found. In some patients, the appendix appears macroscopically normal; however, upon thorough histologic examination of the entire appendix, an adenoma invariably is found. It is noteworthy that none of the patients with PMCA was found with an unequivocally ovarian site of origin. The high frequency of bilateral ovarian involvement, smaller size of the ovarian tumors compared with primary ovarian mucinous tumors unassociated with PMP or carcinomatosis, and histologic appearance of the mucinous tumors are features that support the interpretation that ovarian involvement is secondary in DPAM and PMCA. When these features are present, it is important to examine the gastrointestinal tract for a primary tumor.

Based on studies of patients with PMP that included both benign and malignant tumors, it has been stated that patients with PMP in which the peritoneal mucinous lesions lack epithelial cells have a better prognosis than patients with lesions containing epithelial cells.2, 11 However, if the peritoneal lesions are sampled thoroughly, then some epithelial cells are detected in virtually all patients. Thus, the presence or absence of cells is not a prognostically significant feature in these patients. Prognosis is determined by the nature of the primary tumor associated with the condition and the histologic appearance of the peritoneal tumors. In addition, prognosis can be predicted from the histologic appearance of the peritoneal lesions alone for patients in whom a primary tumor cannot be identified. Thus, patients with histologically bland to low-grade adenomatous peritoneal tumors but who lack a demonstrable appendiceal adenoma can be diagnosed with DPAM. Such patients represent a clinicopathologic entity with an indolent clinical course relative to those with pathologic features of carcinoma. In addition, the morphology of the peritoneal lesions only rarely changes over time, as evidenced by the multiple recurrences in many of the patients who had disease that displayed the same histologic appearance each time. The tumors in patients with PMCA with intermediate features demonstrate the pathologic features of carcinoma and behave more like carcinomatosis than like DPAM; thus, they should be considered carcinomas.

The correlation of behavior with the pathologic features of the peritoneal lesions in this group of patients indicates an inherent biologic difference between DPAM and PMCA. It is important to emphasize, however, that the uniform treatment of patients with DPAM is certainly a factor that contributes to the more favorable prognosis. Maximal surgical intervention (peritonectomy) with complete cytoreduction at the outset seems to offer the best chance for disease control in patients with DPAM.5, 12 This is possible even in those patients who have a large tumor volume. Patients who have had multiple incomplete resections for diagnosis and disease recurrence prior to a definitive peritonectomy procedure experience difficulty with bowel obstruction due to fibrosis and incomplete resection of tumor. Thus, patients with DPAM in whom complete cytoreduction is not possible succumb to the disease. We have not been able to find any histopathologic factors that identify the subset of patients with DPAM who succumb to the disease earlier than most. These patients typically demonstrate greater involvement of small bowel than those who survive, suggesting that there is a critical point at which small bowel involvement is fatal. It is not clear, however, why some patients develop greater small bowel involvement. Such deaths are a result of bowel fibrosis and obstruction rather than the carcinomatosis that is seen in patients with PMCA.

In patients with PMCA, like patients with DPAM, the completeness of cytoreduction is a strong determinant of prognosis.13 However, in contrast to those with DPAM, the ability to achieve complete cytoreduction in patients with PMCA is correlated with tumor volume. The peritoneal tumors in PMCA are histologically malignant with evident mitotic activity and, thus, are likely susceptible to the cytotoxic effects of chemotherapeutic agents. Therefore, the pathologic features, aggressive behavior of PMCA, and inability to completely resect large-volume disease argue in favor of the addition of chemotherapeutic agents to the treatment regimen for patients with PMCA.14, 15 In the current study, all patients received intraperitoneal chemotherapy followed by systemic chemotherapy. The current management protocol for patients with PMCA includes the use of heated intraperitoneal chemotherapy that is initiated in the operating room using mitomycin C and is continued through postoperative Day 5 using 5-fluorouracil. Patients with DPAM receive only the intraoperative therapy.13 In light of the histologically bland peritoneal tumors in patients with DPAM, it is not clear how chemotherapy contributes to successful treatment. It is possible that the potential therapeutic effect of the intraperitoneal chemotherapy is due to the mechanical effects of the manual lavage procedure that accompanies the intraperitoneal instillation of the agents rather than to a cytotoxic effect of the drugs. Ideally, the role of chemotherapy in the treatment of patients with DPAM should be evaluated by a randomized clinical trial that compares maximal surgical treatment alone (including a thorough peritoneal lavage procedure) with the combination of surgery and chemotherapy.

In summary, the pathologic classification of the bland peritoneal mucinous tumors associated with the clinical syndrome of PMP as DPAM identifies a group of patients with a significantly more favorable prognosis than those with mucinous carcinomatosis, thus enabling clinicians to better predict prognosis for patients. In addition, therapeutic approaches based on homogeneous pathologic entities can now be considered rationally.