Title: Pseudomyxoma peritonei: a French multicentric study of 301 patients treated
with cytoreductive surgery and intraperitoneal chemotherapy
Authors: D. Elias, F. Gilly, F. Quenet, JM. Bereder, L. Sideris, B. Mansvelt, G.
Lorimier, O. Glehen
Reference: YEJSO 2935
To appear in:
European Journal of Surgical Oncology
Received Date: 11 November 2009
Accepted Date: 18 January 2010
Please cite this article as: Elias D, Gilly F, Quenet F, Bereder JM, Sideris L, Mansvelt B, Lorimier
G, Glehen O. Pseudomyxoma peritonei: a French multicentric study of 301 patients treated with
cytoreductive surgery and intraperitoneal chemotherapy, European Journal of Surgical Oncology (2010),
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
Pseudomyxoma peritonei: a French multicentric study of 301 patients
treated with cytoreductive surgery and intraperitoneal chemotherapy.
D. Elias1, MD, PhD,
F. Gilly2, MD, PHD,
F. Quenet3, MD,
JM. Bereder4, MD
L. Sidéris5, MD,
B. Mansvelt6, MD
G. Lorimier7, MD
O. Glehen2, MD, PhD,
And the Association Française de Chirurgie
1Institut Gustave Roussy, Villejuif.
2Center Hospitalo-Universitaire Lyon-sud
3Center Val d’Aurel, Montpellier
4Center Hospitalo-Universitaire, Nice
5Hopitâl Maisonneuve, Montréal
6Hopital Jolimont, Haine St Paul
7Center Paul Papin, Angers
Corresponding author : Dominique ELIAS
Chief - Department of Oncologic Surgery
39 Rue Camille Desmoulins
94805, Villejuif, Cédex,
with complete cytoreductive surgery were considered, the extent of peritoneal seeding
The results obtained in 301 patients with pseudomyxoma peritonei treated by
cytoreductive surgery combined with perioperative intraperitoneal chemotherapy in 18
French-speaking centers confirms good results and that these procedures should be
recommended in specialized centers.
To analyze a large series of patients with pseudomyxoma peritonei (PMP) treated with
cytoreductive surgery associated with perioperative intraperitoneal chemotherapy
(PIC) in 18 French-speaking centers.
Patients and methods:
From March 1993 to December 2007, 301 patients with diffuse PMP were treated by
cytoreductive surgery with PIC. Complete cytoreductive surgery was achieved in 219
patients (73%), and hyperthermic intraperitoneal chemotherapy (HIPEC) was performed
in 255 (85%), mainly during the latter period of the study.
Postoperative mortality and morbidity were 4.4% and 40%, respectively. The mean
follow-up was 88 months. The 5-year overall and disease-free survival rates were 73%%
and 56%, respectively. The multivariate analysis identified 5 prognostic factors : the
extent of peritoneal seeding (p=0.004), the center (p=0.0004), the pathologic grade
(p=0.03), gender (p=0.02), and the use of HIPEC (p=0.04). When only the 206 patients
was the only significant prognostic factor (p=0.004).
This large multicentric retrospective study confirms that cytoreductive surgery
combined with PIC (with the use of hyperthermia) should be considered as the gold
standard treatment of PMP and should be performed in specialized centers. It
underlines the prognostic impact of the extent of peritoneal seeding, especially in
patients treated by complete cytoreductive surgery. This prognostic impact appears to
be greater than that of the pathologic grade.
Key-words: peritoneal pseudomyxoma, cytoreductive surgery, intraperitoneal
chemotherapy, hyperthermia, prognosis;
Pseudomyxoma peritonei (PMP) is a clinico-pathologic entity characterized by the
production of mucinous ascites (1,2) and mostly originates from appendiceal epithelial
neoplasms. The tumor grows, obstructing the narrow lumen of the appendix and
subsequently leading to appendiceal perforation (1,2). Then neoplastic cells
progressively and slowly colonize the peritoneal cavity and copious mucin production
builds up in the peritoneal cavity. Recently, pathologic enteric bacteria were shown to
be present in all specimens and their density was correlated with the prognosis (3).
Appendix tumors causing PMP range from a benign pathologic appearance
(disseminated peritoneal adenomucinosis or DPAM), to malignant pathologic findings
(peritoneal mucinous carcinomatosis or PMCA), with some intermediate pathologic
grades (4). Clinically, this syndrome ranges from early PMP fortuitously discovered on
imaging or during a laparotomy performed for another reason, to advanced cases with
a distended abdomen, occlusive syndrome and starvation.
The conventional treatment of PMP is surgical debulking repeated, as necessary to
alleviate pressure effects (5,6). However, this treatment is palliative, leaving visible or
occult disease in the peritoneal cavity. Repeated debulking surgeries become ever
more difficult due to progressively thickened intra-abdominal adhesions (7).
A combination of cytoreductive surgery (CRS) with perioperative intraperitoneal
chemotherapy (PIC) was proposed by Sugarbaker (8). The rationale behind this strategy
was to resect visible disease and any occult residual disease would be treated with a
high concentration of chemotherapy and hyperthermia which mutually potentiate one
another’s effects (9,10). Fifteen years and 600 publications later (7), a systematic
review of this novel treatment concluded that this combined approach was very
probably superior to conventional therapy, even in the absence of controlled
procedures which included a total anterior and lateral parietal peritonectomy,
randomized trials, and even with evidence from only five series of more than 100 cases
(11-16). Each series reported the experience of only one team with various
experiences, and there was no multicentric series reflecting “reality”.
This study aimed to collect patients with PMP treated with this combined approach in
18 different French-speaking centers, to analyze morbidity, mortality and survival
results, and to identify the main prognostic factors for better patient selection.
Patients and methods
The AFC (Association Française de Chirurgie) conducted a multicentric retrospective
study in French (n=20), Belgian (n=2), Canadian (Quebec, n=2) and Swiss (n=1) centers.
Patients were treated between March 1993 and December 2007. Inclusion criteria were
1) pathologic confirmation of PMP 2) patients treated with the intent of achieving an R-
0 resection with cytoreductive surgery combined with hyperthermic intraperitoneal
chemotherapy (HIPEC) or early postoperative intraperitoneal chemotherapy (EPIC).
Exclusion criteria were tumors originating from the colon, small bowel and appendiceal
adenocarcinoma without a predominant mucinous component.
Resection of PMP obeyed principles described elsewhere (17). In short, peritonectomy
omentectomy and splenectomy, right and left sub-phrenic peritonectomy, pelvic
peritonectomy and lesser omentectomy and cholecystectomy with stripping of the
omental bursa were performed exclusively in the macroscopically pathologic areas.
More or less extended resections of the colon, rectum, small-bowel and stomach were
associated, if invaded by PMP. Intraperitoneal chemotherapy was delivered
immediately after cytoreductive surgery. EPIC was performed from day 1 to day 5 after
surgery, and was used during the early period of this retrospective study. HIPEC was
performed with an opened abdominal cavity (Coliseum technique), or a closed one and
was used during the major second period of the study.
A standard data form was created to retrieve information on the primary colorectal
tumor, on the status of the patient prior to the combined procedure, and on previous
treatment with systemic chemotherapy. The extent of PC is assessed through
intraoperative exploration using the Peritoneal Cancer Index (PCI) which scores from 0
to 3 each of the 13 defined areas of the abdominal cavity, and therefore ranges from 1
to 39. Four PCI subgroups were arbitrarily defined for the analysis : 1 to 6, 7 to 12, 13
to 19, and > 19. Information recorded about the combined procedure included the
completeness of cytoreductive surgery, any simultaneous resection of the primary
tumor and liver metastases, the presence or absence of lymph node metastases, the
type of perioperative intraperitoneal chemotherapy (HIPEC or EPIC) and its modalities,
and treatment with adjuvant systemic chemotherapy. Assessment of the completeness
of the cancer resection (CCR) with cytoreductive surgery was done by the surgeon at
the end of the procedure and classified into three categories: CCR-0 signified no
macroscopic residual cancer, CCR-1, no residual nodules greater than 2.5 mm, and
CCR-2, that the diameter of the residual nodules exceeded 2.5 mm. Information was
obtained regarding the postoperative course, including postoperative deaths (within 30
postoperative days) and the cause of death. Major complications (grade 3 and 4
complications according to the National Cancer Institute’s Common Toxicity Criteria)
were considered and the reasons for re-operation were detailed.
PMP was classified into 3 pathological grades according to the Ronnett classification
system (4): grade 1 corresponded to disseminated peritoneal adenomucinosis (DPAM),
grade 2 to peritoneal carcinomatosis with intermediate features, and grade 3 to
peritoneal mucinous carcinomatosis (PMCA).
Data recorded included the most recent follow-up information, the status of the
patient (alive with disease, alive without disease, died with disease, died without
disease), and the site of the initial recurrence.
The descriptive analyses of the patients included in the registry, as well as the survival
analyses were performed on a per-patient basis (one record per patient). However, all
procedures were considered for the analyses of postoperative morbi-mortality events
(death or major complications occurring within 30 days after surgery).
Categorical variables were described in terms of frequency and percentages. The
distributions of continuous variables were described with the mean, standard error,
median, 1st and 3rd quartiles. Patient, disease and treatment characteristics were
correlated with postoperative morbi-mortality events using univariate and multivariate
logistic regression models, adjusted on centers, because of their strong prognostic
The analysis of long-term mortality censored information after the cut-off date of
December 31st 2007 since an active enquiry was performed in all centers to collect the
status (dead/alive) of the patients at that date. The analysis of total mortality
considered the interval from the first procedure to the date of death, the date of the
speaking centers with CRS plus PIC (5 centers did not treat PMP). The number of
last news, or the cut-off date, whichever came first. The analysis of recurrence or
death was based on the interval from the first procedure to the date of first
recurrence, the date of death, the date of the last news, or the cut-off date,
whichever came first. Postoperative deaths were not excluded from the survival
analysis. Patients who had CCR-2 resections with residual tumor nodules exceeding 2.5
mm were considered as having immediate relapses. When the date of recurrence was
unknown in patients who had died, the date of death was used instead. Kaplan-Meier
survival estimates were calculated, and compared between strata with the log-rank
test. The influence of baseline risk factors on the hazard of death was assessed using a
multiple proportional hazard regression model stratified on centers. Stratification was
justified due to considerable heterogeneity in hazards between centers, and to a
strong potential confounding effect on other risk factors. In all multivariate analyses,
age and PCI were entered as continuous variables. The completeness of cytoreduction
was also entered as continuous when justified by a linear trend across categories. Risk
factors with a trend of significance (p ≤ 0.10) were retained in the final model. The SAS
statistical software (Windows, V9.1) was used for all analyses.
From March 1993 to December 2007, 301 patients with PMP were treated in 18 French-
patients treated per center ranged from 130 to 1. There were 190 females and 111
males and the mean age was 52 years (1st-3rd quartile: 44-59; median: 52, range: 26-
PMP originated from the appendix in 91% of the cases (ovary in 7%, and the origin was
unknown in 2%). The circumstances leading to the discovery of PMP were : unexplained
ascites leading to laparoscopy or laparotomy ( (32%), abdominal pain (27%), increasing
abdominal girth (15%), on imaging performed for an another reason (15%),
deterioration of general status (5%), wound hernia (4%), and appendicitis (2%).
The mean peritoneal index was 18.5 (1st- 3rd quartiles: 11-26, median:18, range:9-39).
Concerning the completeness of cytoreductive surgery, CCR-0 had been achieved in
73%, CCR-1 in 20%, and CCR-2 in 7%. Concerning intraperitoneal chemotherapy, 255
patients (85%) had been treated with HIPEC and 46 patients (15%) with EPIC. Hipec was
done even with mitomycin C based-regimens during 60-120 min. at 41-42°C, even with
oxaliplatin-based regimens during 30 min at 43°C. EPIC was done with mitomycinC
during day 1 and 5-fluorouracyl from day 2 to day 5. Systemic chemotherapy had been
administered before or after this combined treatment in 69 patients (23%), mainly
those presenting severity of illness criteria (mean number of courses: 7.1).
Mortality and Morbidity
Thirteen patients (4.4%) had died postoperatively. The causes of death were septic
shock (9 patients) linked to abdominal infection, respiratory complications (4 patients),
hematological toxicity (2 patients), pulmonary embolism (1 patient), and acute renal
insufficiency (1 patient).
Grade 3-4 complications had occurred in 119 patients (40%), and 17.5% (n=52) of all
patients had been re-operated on; 8% (n=24) had a digestive fistula, 9% (n=27) a
intraperitoneal abscess, 13% (n=38) hemorrhage, 14% (n=42) a pulmonary infection, and
20% (n=58) grade 3-4 hematological toxicity. The median duration of hospitalization
was 21 days (mean duration 26 ±14 days).
The logistic regression analysis of factors which significantly increased the risk of
mortality and morbidity identified two main factors, the peritoneal index (p=0.002,
correlated with the extent and duration of surgery), and the presence of a high
pathologic grade (PMCA) (p= 0.03).
With a mean follow-up of 88.4 months (1st – 3rd quartile: 28 – 125, median: 73), the
overall 1-year, 3-year, and 5-year survival rates were respectively 89.4%, 84.8% and
72.6% (figure 1). The 10-year survival rate was 54.8%. Median survival had not yet been
reached, but it will be longer than 100 months. The disease-free survival (DFS) rate
was 56% at 5 years (figure 1) and the median duration of DFS was 78 months.
There was a strong center effect (p=0.004) that justified an analysis of the impact of
other prognostic factors adjusted on centers (figure 2). Figure 2 shows 7 curves
corresponding to the six centers which treated at least 10 patients and one curve (the
lowest one) for the 11 remaining centers (which treated fewer than 10 patients per
In the univariate analysis (table 1), age did not exert an impact on survival unlike
gender (p=0.02, in favor of females), the extent of peritoneal seeding (p<0.001), the
completeness of CRS (p<0.001), the pathologic grade (p<0.001) and the presence of
invaded lymph nodes (p<0.001). Two other prognostic factors must be underlined:
HIPEC seemed more efficient than EPIC (p=0.001) and the results were different
according to the centers (p=0.0004). The prognosis was worse (p=0.01) for patients who
had received systemic chemotherapy (because it was given in advanced and high-grade
The multivariate analysis identified five factors with an independent influence on
survival (table 2). First, the center effect (p<0.01), second, the extent of peritoneal
seeding (p=0.004) (figure 3) with a significant decline in survival when the peritoneal
index was greater than 20. This index was strongly correlated with the completeness of
the CRS : CCR1 and CCR-2 were only observed when the peritoneal index was high.
Thirdly, the pathologic grade (p=0.03) was a poor prognostic factor for PMCA. No
significant difference in survival was found between DPAM and intermediate grades.
The other two factors were gender (p=0.02) in favor of females, and the use of HIPEC
rather then EPIC (p=0.04).
Multivariate analysis of disease-free survival identified two significant factors: the
extent of the peritoneal seeding (p<0.001) (strongly correlated with the CCR-status),
and the center’s experience (p=0.006).
In the group of 206 patients with a CCR-0 resection, the 5-year and 10-year survival
rates were 84% was 61% respectively. It is noteworthy that two factors became non-
significant in the univariate analysis : gender and the pathologic grade. Only one factor
continued to exert a prognostic impact : the extent of peritoneal seeding (p=0.004),
while HIPEC was marginally better (p=0.07) than EPIC.
This study reports the results of the first multicentric retrospective analysis of 301
patients with diffuse PMP treated by cytoreductive surgery combined with
perioperative intraperitoneal chemotherapy. Postoperative mortality was 4%, and
grade 3-4 morbidity was 40 %. The 5-year survival rate was 72.6% and median disease-
free survival was 78 months. Median survival has not yet been attained but was longer
than 100 months.
Survival. This study reports on the largest series of patients with PMP ever published
after the monocentric series of 501 cases reported by Sugarbaker (11). Compared to
the exceptional expertise of Sugarbaker, it presents results from 18 French-speaking
centers, and adequately reflects the management of PMP in real-life settings where the
experience of such treatment is very unequal from one center to another. It is always
difficult to extrapolate results from a monocentric experience to other centers, due to
considerable heterogeneity in selection criteria, in CRS and intraperitoneal
chemotherapy techniques. Our study effectively underlines the importance of the
experience of the center in the survival results. Due to a healthy degree of scepticism
among some oncologists, any multicentric study collecting data from a large number of
centers is particularly useful, even if the experience and the techniques differ
considerably from one center to another. This trial concerns a collaborative effort of
25 French-speaking centers, which included 1290 patients with primary or digestive
peritoneal carcinomatosis treated by CRS combined with PIC to evaluate the efficiency
of this combined approach (18 of them have treated PMP). Due to its retrospective
nature, this study suffers from many potential biases in the selection of patients and in
treatments. Nonetheless, it has allowed us to present a large number of patients
treated in many centers and to draw conclusions.
Also, the 5-year overall survival rate of 72.6% for PMP represents the worst results that
can be obtained with CRS combined with PIC because it takes into account the learning
curves of 18 centers, and because progress in patient selection and in the quality of
surgery increases with experience. This overall survival rate is identical to the 72%
reported by Sugarbaker in his cohort of 501 patients (11). Thus, our multicentric study
corroborates the results of the largest monocentric study. If we analyze the other four
trials which included more than 100 patients treated with this combined approach, our
survival results are better than those reported by Zoetmulder (13) and Loggie (14),
close to those of Chua et al. (16) but worse than those reported by Elias (15) who
included only patients whose resection was complete or sub-complete (CCR-0 or CCR-
Moreover, a mortality rate of 4% and a morbidity rate of 40% for this combined
approach are the sort of rates that would be obtained with any kind of complex
oncologic surgery such as oesophagectomy or duodeno-pancreatectomy. These rates
are therefore acceptable face to benefits for overall and disease-free survival.
The results of the conventional treatment of PMP mainly based on repeated surgical
debulking have been reported in two series (5,6). Gough et al. (5) reported a 5-year
survival rate of 53%, with a 97% recurrence rate at 10 years in 56 PMP patients who
underwent serial debulking and were selectively treated with intraperitoneal
radiotherapy or chemotherapy between 1957 and 1983. Minner et al. (6) reported an
88% recurrence rate at 5 years in 97 PMP patients treated with serial debulking,
systemic chemotherapy and/or delayed intraperitoneal 5-fluorouracil over a 22-year
period. The 3% and 12% disease-free survival rates are well below the 56% rate we
observed in our series. Sugarbaker and Zoetmulder also reported higher rates, 47% and
61%, respectively. All these results are probably due to the superiority of the combined
over the conventional approach. In France, this combined approach has become the
gold standard treatment when feasible (18).
Prognostic factors. The multivariate analysis of our series identified four significant
and independent prognostic factors, adjusted on centers: the extent of the disease,
the pathologic grade, gender, and the use of hyperthermia. One of these factors,
namely the extent of the disease, was not identified in the five earlier series of more
than 1OO patients treated with the combined approach. In contrast, the completeness
of cytoreduction was statistically significant in three (12,14,16). The pathologic grade
was not retained in three series (12-15), but selected in two (13,16) which reported
only 31 CCR-0 resections in 103 treated patients and 73 CCR-O resections in 106
treated patients for the first and the second study respectively.
Thus, new information is emerging from our multicentric study.
First, this multicentric study underlines the influence of the center in which the
procedure was performed and this led us to perform a multivariate analysis adjusted on
centers. Centers with less experience had a lower rate of complete cytoreduction and
poorer survival results. The fact that incomplete cytoreduction in the 174 patients
reported by Glehen et al (23), resulted in only 15% of patients alive at 5 years argues in
favor of exclusively entrusting this approach to experienced, good quality centers in
Second, the extent of peritoneal seeding has a potent prognostic impact, even if this
impact is probably lower than for patients with colorectal peritoneal carcinomatosis
(18-21). In our series, the 5-year survival rate was only 57% for patients with a
peritoneal index exceeding 19, whereas it was 83% for the other patients. We also
established that the completeness of cytoreduction was strongly correlated with the
disease extent (peritoneal cancer index), and also with the pathologic grades (data not
shown). Finally, extensive PMP with high pathologic grades are the most frequently
incompletely resected lesions and mainly in inexperienced centers.