Toward curative treatment of peritoneal carcinomatosis from nonovarian origin by cytoreductive surgery combined with perioperative intraperitoneal chemotherapy

A multi-institutional study of 1290 patients




Peritoneal carcinomatosis (PC) from nonovarian malignancies long has been regarded as a terminal disease. Over the past decade, new locoregional therapeutic approaches combining cytoreductive surgery with perioperative intraperitoneal chemotherapy (PIC) have evolved that have demonstrated improved survival.


A retrospective, multicenter cohort study was performed in French-speaking institutions to evaluate toxicity and principal prognostic factors after cytoreductive surgery and PIC (hyperthermic intraperitoneal chemotherapy [HIPEC] and/or early postoperative intraperitoneal chemotherapy [EPIC]) for PC from nongynecologic malignancies.


The study included 1290 patients from 25 institutions who underwent 1344 procedures between February 1989 and December 2007. HIPEC was performed in 1154 procedures. The principal origins of PC were colorectal adenocarcinoma (N = 523), pseudomyxoma peritonei (N = 301), gastric adenocarcinoma (N = 159), peritoneal mesothelioma (N = 88), and appendiceal adenocarcinoma (N = 50). The overall morbidity and mortality rates were 33.6% and 4.1%, respectively. In multivariate analysis, patient age, the extent of PC, and institutional experience had a significant influence on toxicity. The overall median survival was 34 months; and the median survival was 30 months for patients with colorectal PC, not reached for patients with pseudomyxoma peritonei, 9 months for patients with gastric PC, 41 months for patients with peritoneal mesothelioma, and 77 months for patients with PC from appendiceal adenocarcinoma. Independent prognostic indicators in multivariate analysis were institution, origin of PC, completeness of cytoreductive surgery, extent of carcinomatosis, and lymph node involvement.


A therapeutic approach that combined cytoreductive surgery with PIC was able to achieve long-term survival in a selected group of patients who had PC of nonovarian origin and had acceptable morbidity and mortality. The current results indicated that this treatment should be centralized to institutions with expertise in the management of PC. Cancer 2010. © 2010 American Cancer Society.

The primary peritoneal malignancies, such as peritoneal mesothelioma or papillary serous carcinoma, are rare. In contrast, peritoneal dissemination from digestive cancers is common. In the past, carcinomatosis was regarded as a terminal disease, and most oncologists viewed it as a condition only to be palliated. However, recent reports describe curative treatment options for selected patients with carcinomatosis from nongynecologic cancer.1-5

Over the past 2 decades, novel therapeutic approaches to peritoneal carcinomatosis (PC) have emerged that combine cytoreductive surgery and peritonectomy procedures6 with perioperative intraperitoneal chemotherapy (PIC), including hyperthermic intraperitoneal chemotherapy (HIPEC) and/or early postoperative intraperitoneal chemotherapy (EPIC).7-9 Theoretically, cytoreductive surgery is performed to treat macroscopic disease, and PIC is used to treat microscopic residual disease with the objective of removing disease completely in a single procedure. Many consider it the standard of care for diseases like pseudomyxoma peritonei or peritoneal mesothelioma.2, 5, 10 Moreover, it is the only plan that has produced curative results for carcinomatosis in international registries,11 in several phase 2 trials, and in 1 phase 3 trial.3, 12-14 Despite all of these studies suggesting a survival benefit, oncologists remain skeptical regarding this combined therapeutic approach partly because of its perceived high toxicity. A collaborative effort of 25 French-speaking institutions involved in the treatment of peritoneal surface malignancies was performed on a large number of patients with PC from nongynecologic carcinomatosis to evaluate the tolerance and the efficacy of this treatment, to answer questions regarding patient selection, and to identify principal prognostic indicators.


Patient Population

One thousand two hundred ninety patients who had undergone 1344 procedures combining cytoreductive surgery and PIC for the treatment of PC from nonovarian origin between February 1989 and December 2007 made up the study population. The inclusion criteria were primary PC or PC of nonovarian origin confirmed by pathologic examination with HIPEC, EPIC, or both administered within 7 days of surgery. The exclusion criteria were PIC performed >7 days after the surgery and the presence of extra-abdominal metastases.

Standardized clinical data on consecutive patients from each of 25 institutions involved in the management of peritoneal surface malignancies (France, 20 institutions; Belgium, 2 institutions; Quebec, Canada, 2 institutions; and Switzerland, 1 institution) were received and entered into a central database. Institutions provided data on between 1 and 400 patients. Figure 1 lists the periods during which patients were studied. The same author reviewed all data sheets before their entry into the database in an effort to make this a uniform interpretation of retrospective data.

Figure 1.

This chart illustrates the evolution of 1 year's inclusions of procedures combining cytoreductive surgery and perioperative intraperitoneal chemotherapy for the treatment of peritoneal carcinomatosis.

Data Forms

A standard data form was created to retrieve information on the origin of PC and the status of the patient before undergoing the combined procedure, including age, sex, extent of PC, and previous treatment with systemic chemotherapy. The extent of PC was assessed intraoperatively. Institutions used 2 different tools, the Gilly classification15 and the Sugarbaker Peritoneal Cancer Index (PCI),16 to assess the extent of PC. By using the Gilly classification, carcinomatosis was classified into 5 stages: stage 0, positive peritoneal cytology; stage I, malignant tumor nodules <5 mm in greatest dimension localized to 1 part of the abdomen; stage II, diffuse throughout the whole abdomen; stage III, tumor nodules between 5 mm and 2 cm in greatest dimension; and stage IV, large tumor deposits (>2 cm in greatest dimension). For the descriptive analysis, 4 subgroups according to the PCI score were created: 1 to 6, 7 to 12, 13 to 19, and >19. Information recorded about the combined procedure included the date, the completeness of cytoreductive surgery, the simultaneous resection of primary tumor or of liver metastases, the presence or absence of lymph node metastases, the type of perioperative intraperitoneal chemotherapy (HIPEC or EPIC) and its modalities (drugs used, “closed” or “open” HIPEC, temperatures, duration), and treatment with adjuvant systemic chemotherapy. The assessment of the completeness of the cancer resection (CCR) by cytoreductive surgery was performed by the surgeon at the end of the procedure and was classified in 3 categories4: CCR-0 (no macroscopic residual cancer remaining), CCR-1 (no residual nodule >2.5 mm), and CCR-2 (residual nodules >2.5 mm in greatest dimension). The institutions were classified regarding their length of experience with performing cytoreductive surgery at the time of the procedure (0-3 years, 3-7 years, 7-11 years, and >11 years). Four groups also were analyzed regarding the period of the procedure (before 1997, between 1997 and 1999, between 2000 and 2002, between 2003 and 2004, and after 2005). Information was obtained regarding the postoperative course, including postoperative death (within 30 postoperative days) and its cause, major complications (grade 3 and 4 complications according to the National Cancer Institute's Common Toxicity Criteria), and reoperations. Follow-up data recorded included the date of the most recent follow-up, the status of the patient (alive with disease, alive without disease, dead with disease, or dead without disease), the site of initial recurrence, and all other subsequent sites of recurrence.

Statistical Analysis

The descriptive analyses of the patients who were included in the registry and the survival analyses were performed on a patient basis (1 record per patient). However, all procedures were considered for the analyses of postoperative morbidity or mortality events (death or major complications that occurred within 30 days after surgery). Categorical variables are described in terms of frequency and percentages. The distributions of continuous variables are described as the mean, standard error, median, and first and third quartiles. The influence of patient, disease, and treatment characteristics was related to the risk of postoperative morbidity or mortality events using univariate and multivariate logistic regression models that were adjusted by institution. The analysis of long-term mortality censored information after the cutoff date of December 31, 2006, because an active enquiry was performed in all institutions to collect the status (dead/alive) of the patients at that date. There were 72 patients included after the cutoff date, 4 patients had an unknown date of death, and 46 patients were lost to follow-up (the last follow-up before the cutoff date). The analysis of total mortality considered the interval from the first procedure to the date of death, the date of last follow-up, or the cutoff 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 last follow-up, or the cutoff date, whichever came first. When the date of recurrence was unknown in patients who died, the date of death was used instead. For this analysis, 32 patients were lost to follow-up. 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 hazards regression model stratified by institution. Stratification was justified by a large heterogeneity of hazards between institutions and by a strong potential confounding effect on other risk factors. In all multivariate analyses, age and PCI were entered as continuous variables. Completeness of cytoreduction (4 categories) also was entered as a continuous variable when justified by a linear trend across categories. Risk factors with a significance degree ≤.10 were retained in the final model. SAS statistical software for Windows (version 9.1; SAS Institute, Inc., Cary, NC) was used for all analyses.


Twenty-5 institutions treated 1290 patients who underwent 1344 procedures. One institution recorded >403 patients, 3 institutions recorded 100 to 300 patients, 3 institutions recorded 50 to 100 patients, 4 institutions recorded 20 to 50 patients, 5 institutions recorded 10 to 20 patients, and 9 institutions recorded <10 patients.

Patient Characteristics

There were 728 women (56.5%) and 562 men (43.5%), and the mean ± standard deviation (SD) patient age was 52 ± 12 years. The Eastern Cooperative Oncology Group performance status was 0 or 1 for 95% of patients. The origin of PC was mainly colorectal adenocarcinoma (40.5%), followed by pseudomyxoma peritonei (23.3%), gastric cancer (12.3%), peritoneal mesothelioma (6.8%), and appendiceal adenocarcinoma (3.9%) (Table 1). In the majority of patients, the diagnosis of carcinomatosis was an incidental finding at laparoscopy or laparotomy (Table 2). Radiologic examinations were performed for diagnosis or to assess disease extent and included abdominal computed tomodensitometry (CT) for 1047 patients (81.2%), positron emission tomography CT for 331 patients (25.7%), and magnetic resonance imaging for 48 patients (3.7%). One hundred thirteen patients (8.8%) underwent laparoscopy. Six hundred seventy-five patients (52.3%) had received previous neoadjuvant or palliative systemic chemotherapy with a mean ± SD number of 7.4 ± 4.8 cycles.

Table 1. Origin of Carcinomatosis in Patients Who Underwent Cytoreductive Surgery Combined With Perioperative Intraperitoneal Chemotherapy
EtiologyNo. of Patients%
  1. GIST indicates gastrointestinal stromal tumor.

Colorectal cancer52340.5
Pseudomyxoma peritonei30123.3
Gastric cancer15912.3
Peritoneal mesothelioma886.8
Appendiceal adenocarcinoma503.9
Small bowel adenocarcinoma453.5
Primary peritoneal serous carcinoma302.3
Peritoneal sarcomatosis282.2
Uterine adenocarcinoma13 
Unknown primary adenocarcinoma8 
Small bowel carcinoid7 
Urachal fold4 
Uterine epidermoid carcinoma4 
Appendiceal carcinoid3 
Desmoplastic tumor (small round cell)3 
Hepatocellular carcinoma2 
Malignant teratoma2 
Table 2. Modes of Carcinomatosis Diagnosis
DiagnosisNo. of Patients%
Incidental finding (laparotomy or laparoscopy)43934
Abdominal pain25119.5
Radiologic examinations13810.3
Ascites or increased abdominal girth1249.2
Weight loss or lethargy675.1
Intestinal obstruction503.7
Elevated tumor markers201.5
Symptomatic hernia191.4
Anemia or gastrointestinal hemorrhage131
Appendicitis or peritonitis131
Abdominal mass70
Constipation or diarrhea50


The assessment of carcinomatosis extent was performed using the Gilly classification for 678 patients (53%) and the PCI for 1152 patients (89%). The Gilly stage was 0 in 42 patients (6.2%), 1 in 89 patients (13.1%), 2 in 124 patients (18.3%), 3 in 165 patients (24.3%), and 4 in 258 patients (38.1%). The mean ± SD PCI was 13.1 ± 8.9.

After the best surgical effort at cytoreductive surgery, 950 patients were classified with CCR-0 reductions (74.7%), 218 patients were classified with CCR-1 reductions (17.1%), and 104 patients were classified with CCR-2 reductions (8.2%). Synchronous liver metastases were present (and were resected) in 125 patients (11%). The cytoreductive surgery was synchronous with resection of the primary tumor in 328 patients (32.9%). One thousand one hundred six patients (85.8%) received HIPEC, and 211 patients (16.6%) received EPIC. HIPEC was combined with EPIC in 37 patients (3.3%). The drugs and regimens that were used for this intraperitoneal chemotherapy are reported in Table 3. All HIPEC procedures were performed intraoperatively after cytoreductive surgery but with variations in exposure techniques (open abdomen in 700 patients [65.2%] and closed abdomen in 373 patients [35%]), drugs, duration (30-120 minutes; mean ± SD duration, 59.5 ± 28.7 minutes), intraperitoneal temperature (40-43°C; mean ± SD intraperitoneal temperature, 42.5 ± 0.8°C), type of perfusate, and flow rates.

Table 3. Types of Drugs and Regimens Used for Hyperthermic Intraperitoneal Chemotherapy and Early Postoperative Intraperitoneal Chemotherapy
  1. HIPEC indicates hyperthermic intraperitoneal chemotherapy; EPIC, early postoperative intraperitoneal chemotherapy.

HIPEC (1154 procedures; 86.4%)
 1. Mitomycin C-based regimens
 Mitomycin 30-50 mg/m² with or without cisplatin  50-100 mg/m² delivered over 60-120 min at 41-42.5°C
 2. Oxaliplatin-based regimens
 Oxaliplatin 360-460 mg/m² with or without irinotecan  100-200 mg/m² with or without intravenous 5-fluouracil  and leucovorin delivered over 30 min at 43°C
EPIC (190 procedures; 13.6%)
 Abdominal cavity filled at the end of surgery with 1 L/m²  Ringer lactate
 EPIC lasted 5 days (Days 1-5), drains clamped at 23 h/24 h
 Day 1, Mitomycin C 10 mg/m²
 Days 2-5, 5-fluorouracil 600 mg/m²

Fifty patients underwent a second combined procedure that involved cytoreductive surgery and perioperative surgery, and 4 patients underwent a third procedure. Four hundred seventeen patients (34.8%) received postoperative adjuvant systemic chemotherapy when they demonstrated an objective response to preoperative chemotherapy (if administered) or when they had bad prognostic features (CCR-1 or CCR-2 reduction, involved lymph nodes, or liver metastases), with a mean ± SD of 7.6 ± 4.5 cycles.

Postoperative Mortality and Morbidity

Fifty-two patients (4.1%) died postoperatively. The main causes of death were multiorgan failure (11 patients), septic shock (10 patients), respiratory complications (10 patients), digestive fistula or peritonitis (4 patients), cardiorespiratory failure (3 patients), pulmonary embolism (2 patients), cardiac arrhythmia (2 patients), hemorrhagic shock (2 patients), hematologic toxicity (2 patients), and acute renal insufficiency (2 patients).

Major complications (grade 3 and 4 complications) occurred in 403 patients (33.6%). Details of postoperative complications are reported in Table 4. A reoperation was necessary in 173 patients (14%). Neutropenia was the principal complication and occurred in 157 patients (13.3%). Digestive fistula occurred in 119 patients (9.7%). The mean ± SD duration of hospitalization was 24.1 ± 17.4 days.

Table 4. Details of Major Complications (Grade 3/4 According to the National Cancer Institute Common Toxicity Criteria)
Type of ComplicationNo. of Patients%
Grade 3-4 complications42233.6
Digestive fistula1239.7
Postoperative bleeding957.7
Intra-abdominal abscess907
Systemic sepsis322.3
Bowel obstruction201.5
Renal insufficiency141

In univariate analysis, several factors had no significant influence on postoperative complications: sex, origin of carcinomatosis, performance status, treatment with a second or third procedure, treatment with neoadjuvant systemic chemotherapy, synchronous resection of liver metastasis, and period of the procedure. The impact of the PIC techniques (drugs, drug concentration, temperature, duration) could not be analyzed because of significant variations across centers. The logistic multiple regression analysis of factors that significantly increased the risk of postoperative morbidity and mortality revealed 3 important factors: an increased risk of death or postoperative complication with increasing age (P = .02), the extent of carcinomatosis assessed by the PCI (P < .0001), and the institution where the treatment was performed (P < .0001).


The median follow-up was 45.3 months (first to third quartiles, 20.3-90.9 months). The overall 1-year, 3-year, and 5-year survival rates were 77%, 49%, and 37%, respectively; and the corresponding disease-free survival rates were 55%, 28%, and 22%, respectively (Fig. 2). The overall median survival was 34 months, and it was 30 months for colorectal PC, not reached for pseudomyxome peritonei, 9 months for gastric PC, 41 months for peritoneal mesothelioma, and 77 months for PC from appendiceal adenocarcinoma (Fig. 3).

Figure 2.

Overall and disease-free survival rates are illustrated for 1290 patients who had peritoneal carcinomatosis of digestive or primary origin.

Figure 3.

Overall survival rates are illustrated for patients with colorectal peritoneal carcinomatosis (PC), pseudomyxoma peritonei, peritoneal mesothelioma, gastric PC, and PC from appendiceal adenocarcinoma.

In univariate analysis, the only factor that did not have prognostic impact was treatment with EPIC and the PCI. All other clinical and therapeutic factors had a significant prognostic impact (Tables 5 and 6). Patients aged >61 years had a significantly shorter survival than younger patients (P < .001). There was a strong prognostic influence of the center at which procedures were performed (P < .001). Institutions with experience >7 years produced significantly improved survival (P < .001) (Fig. 4). There was a significant improvement in survival from 1989 to 2004 (Fig. 5). The impact of PIC techniques (drugs and concentrations used, temperature, duration) could not be analyzed because of marked variations between institutions.

Figure 4.

Overall survival is illustrated according to the extent of institutional experience.

Figure 5.

Overall survival is illustrated according to the period when the procedure was performed.

Table 5. Prognostic Impact of Clinical Factors on Overall Survival by Univariate Analysis
Clinical FactorNo. AnalyzedMedian Survival, mo3-Year Survival, %5-Year Survival, %Log-rank P
  • NR indicates not reached; ECOG, Eastern Cooperative Oncology Group; PCI Peritoneal Cancer Index.

  • a

    Independent prognostic factor by multivariate analysis.

Age, y    <.001
Sex    <.001
Etiology    <.001a
 Pseudomyxoma peritonei276NR8573 
 Peritoneal mesothelioma84416243 
 Small bowel adenocarcinoma41324733 
 Appendiceal adenocarcinoma49775656 
 Primary peritoneal serous carcinoma23525843 
ECOG Performance status    <.001
Gilly classification    <.001
Sugarbaker PCI    .201a
Lymph node involvement    <.001
Center's experience, y    <.001
Period of procedure    <.001
Institution <.001
Table 6. Prognostic Impact of Clinical Factors on Overall Survival by Univariate Analysis
Therapeutic FactorNo. AnalyzedMedian survival, mo3-Year Survival, %5-Year Survival, %Univariate P
  • CCR indicates complete cancer resection; EPIC, early postoperative intraperitoneal chemotherapy; HIPEC, hyperthermic intraperitoneal chemotherapy; PIC, perioperative intraperitoneal chemotherapy; NR, not reached.

  • a

    Independent prognostic factor by multivariate analysis.

Preoperative systemic chemotherapy    .008
Synchronous resection of liver metastasis    .003
CCR category    <.001a
Treatment with EPIC    .805
HIPEC duration, min    <.001
PIC technique    <.001
 EPIC alone168314635 
 “Open” procedures614515844 
 “Closed” procedures391183727 
Intraperitoneal temperatures    .006
The use of oxaliplatin during HIPEC    <.001
Adjuvant systemic chemotherapy    .018

In multivariate analysis, the principal independent prognostic factors in addition to the effects of treatment center were the etiology of PC (P < .001), the carcinomatosis extent assessed with the Sugarbaker PCI (P < .001), the completeness of cytoreduction (P < .001) (Fig. 6), and the presence of lymph node involvement (P = .001) (Table 7). The multivariate analysis of disease-free survival selected the same prognostic factors plus the presence of liver metastases (P = .008) with an increased risk of recurrence of 44% (95% confidence interval, 9.8%-88.8%).

Figure 6.

Overall survival is illustrated according to the completeness of cytoreduction achieved.

Table 7. Multivariate Analysis of Prognostic Factors for Overall Survival in 1290 Patients With Peritoneal Carcinomatosis of Digestive or Primary Origin Who Underwent Cytoreductive Surgery Combined With Perioperative Intraperitoneal Chemotherapy
VariableRR95% CIP
  • RR indicates relative risk; CI, confidence interval; PCI, Peritoneal Cancer Index.

  • a

    For each point in the PCI; each additional point increased the RR of death, ie, by 4.3%.

  • b

    In 3 complete cancer resection (CCR) categories: CCR-0, CCR-1, and CCR-2/CCR-3; moving from 1 category to another increases the risk of death by 46%.

  • c

    In 2 groups (yes or no). The RR indicates the additional risk of death.

Colorectal cancer1.000
Pseudomyxoma peritonei0.1570.104-0.237<.0001
Peritoneal mesothelioma0.3830.222-0.661.0006
Primary peritoneal serous carcinoma0.4500.141-1.438.1777
Appendiceal adenocarcinoma0.5890.331-1.048.0716
Gastric cancer1.8701.343-2.603.0002
Small bowel adenocarcinoma0.6590.389-1.116.1211
Peritoneal sarcomatosis1.6080.963-2.684.0695
Other etiology0.7460.462-1.205.2313
Completeness of surgeryb1.4601.192-1.787.0003
Positive lymph nodesc1.3661.060-1.761.0160

We also specifically studied the prognostic factors in patients who had undergone complete cytoreductive surgery (CC-0 resection). The multivariate analysis identified 3 significant prognostic factors: the extent of carcinomatosis assessed by the PCI (P < .001), the etiology of PC (P < .001), and age >52 years (P = .029), with an increased risk of death of 10.7% for every 5-year increase in age (95% confidence interval, 1.1%-21.2%).


Peritoneal carcinomatosis of gastrointestinal or primary origin long has been considered a fatal clinical entity to be treated palliatively. Since the 1980s, the development of new surgical techniques (cytoreductive surgery and peritonectomy procedures) combined with PIC (and, preferentially, HIPEC) have provided new hope of a potential cure for patients with PC. Now, curing PC is a realistic goal. The experiences of many single institutions, combined with phase 2 studies,12-14, 17, 18 an international registry,11 and a phase 3 trial,3 have produced encouraging survival results after treatment of PC with this therapeutic strategy. This has led to the development of new specialized centers for the management of peritoneal surface malignancies that are treating increasing numbers of patients (Fig. 1). This collaborative effort of 25 institutions collected data from 1290 patients and, to our knowledge, represents the largest experience in the treatment of PC published to date. With a median follow-up of 45.3 months and a low rate of patients lost to follow-up (4%), the overall median survival was 34 months, and the disease-free survival rate at 5 years was 22%. For rare diseases like pseudomyxoma peritonei or peritoneal mesothelioma, the combination of cytoreductive surgery with PIC already is considered by many as the standard of care2, 5, 10 The current large, retrospective experience confirms that 5-year survival rates >70% may be obtained for patients with pseudomyxoma peritonei, and a median survival of >40 months may be obtained for patients with peritoneal mesothelioma. Both of these etiologies have a very different biology than other malignancies and represent the best indication for such a therapeutic strategy. For patients with PC from colorectal cancers, a median survival of 30 months and a 5-year survival rate of 26% were observed with this combined treatment. Palliative systemic chemotherapy reportedly produces a median survival between 7 months and 24 months; and, in these patients, long-term survival is rarely, if ever, achieved.13, 19, 20 These results are similar to those obtained in a previous multicenter retrospective study.11 For PC from gastric cancer, the median survival reported was only 9 months, reflecting either a more aggressive disease process that was less responsive to this combined treatment modality or the need for better patient selection.

One of the most important issues arising from our analysis is the strong influence of the institution not only on survival but also on morbidity and morbidity rates. The institution in which the procedure was performed was an independent prognostic indicator of postoperative complications. When comparing centers that performed cytoreductive surgery for the shortest time with the most experienced centers, median survival increased from 23 months to 42 months, but that effect was not replicated when the analysis was adjusted based on the center. However, it is reasonable to assume that experience may provide better patient selection, surgical expertise, and postoperative management. A learning curve already had been reported by several authors. Moran reported a decreased mortality rate from 18% down to 3%,21 The Netherlands Cancer Institute reported a decrease in mortality from 8% to 4%,22 and Yan et al reported a decrease from 7% to 1%.23 All interventional complex procedures have an inherent risk, and experience undoubtedly diminishes but can never abolish this risk.24

The morbidity and mortality rates of 33.6% and 4.1%, respectively, are similar to those reported after other major surgical procedures, such as esophagectomy or pancreaticoduodenectomy, With the potential survival benefit demonstrated after cytoreductive surgery for PC, these rates are acceptable. Morbidity increased significantly with age, and this treatment approach may need careful consideration in individuals aged >70 years who have multiple comorbidities. The extent of PC assessed by the PCI remains the most reliable tool in the evaluation of carcinomatosis extent. A higher PCI tended to result in more extensive cytoreduction. Although PIC has its own morbidity (neutropenia, renal insufficiency with the use of cisplatin) and may have increased the rate of some surgical complications (fistula), most complications that we observed were caused by the surgery itself (digestive fistula, intraperitoneal hemorrhage, and abscesses).

The extent of carcinomatosis also had a significant influence on survival and was correlated strongly with the completeness of cytoreduction (ie, the size of residual tumor nodules). For intraperitoneal chemotherapy to be effective, residual disease after attempted cytoreductive surgery must be of low volume. For colorectal or gastric carcinomatosis, long-term survival was obtained only by patients who underwent complete macroscopic resection (CCR-0). Elias et al13 recently reported a 5-year survival rate of >50% and a median survival of 63 months for patients with colorectal PC who underwent complete cytoreduction and received HIPEC with oxaliplatin. Glehen et al12 observed a median survival of 21 months for patients with gastric PC who underwent CCR-0 reduction and received HIPEC using mitomycin C, whereas no patients who underwent CCR-2 reduction remained alive at 2 years. For less aggressive tumors, such as pseudomyxoma peritonei or peritoneal mesothelioma, CCR-1 resection (residual tumor nodules <2.5 mm in greatest dimension) may be acceptable and may result in long-term survival when combined with PIC.2, 10

Independent negative prognostic indicators included lymph node involvement. This traditionally indicates systemic spread of disease. A locoregional therapeutic approach for a disease that is no longer confined to the peritoneum will be inadequate. Additional disease-specific, systemic chemotherapy is indicated for these patients. Future prospective studies should determine whether this systemic chemotherapy should be given in either the neoadjuvant or adjuvant setting or both. Preoperative identification of systemic disease would allow clinicians to tailor neoadjuvant therapy before attempting cytoreductive surgery and PIC. Although extraperitoneal sites of disease spread can be identified radiologically, lymph node involvement still is primarily a pathologic assessment. Outcomes may be improved if techniques for detecting systemic disease can be developed. However neoadjuvant systemic chemotherapy for patients with resectable PC who have no evidence of systemic spread may jeopardize the chance of achieving complete cytoreduction or may increase morbidity in an attempt to achieve such reduction.

The impact of PIC is difficult to assess because of the heterogeneity in techniques used. None of the variables in PIC that we studied (closed or open technique, drug, temperature, duration) were significant in multivariate analysis in their impact on morbidity, mortality, or survival. Some authors have reported improved survival with hyperthermia (but always in retrospective studies) for patients with pseudomyxoma peritonei and colorectal PC.7, 25 Because of its design, no firm conclusions can be drawn regarding PIC techniques from this retrospective study. Institutional experience and expertise are likely to be the most important variables in limiting procedure-related toxicity and improving survival.9

Despite its bias because of its design and the exclusion of all patients who were inoperable or who underwent palliative surgery, this large multicenter study demonstrated that a potential cure of PC is a realistic goal in selected patients and that PC should not be considered a terminal event. The principal prognostic factors that we identified were the origin of PC, carcinomatosis extent assessed by the PCI, the completeness of cytoreductive surgery, and the involvement of lymph nodes. Complete cytoreductive surgery with the objective of removing all macroscopic disease, combined with PIC performed in an experienced peritoneal surface malignancy center, may provide long-term survival with acceptable morbidity and mortality rates.


We acknowledge Faheez Mohamed for help with English editing, Nadine Bossard for statistical analysis, and other authors of the Association Française de Chirurgie (Abboud Karine, Bellevue Hospital, St. Etienne, France; Turrini Olivier, Padi Calmetteo Centre, Marseilles, France; Arvieux Catherine, Michallon Hospital, Grenoble, France; Rat Patrick, Bocage Hospital, Dijon, France; Gertsch Philippe, San Giovanni Hospital, Bellinzana, Switzerland; Ferron Gwenael, Claudias Regaud Centre, Toulouse, France; Meeus Pierre, Leon Berard Center, Lyon, France; Brigand Cécile, Hautepierre Hospital, Strausbourg, France; Marchal Frederic, Alexis Vautrin Center, Nancy, France; Tuech Jean Jacques, Charles Nicolle Hospital, Rouen, France; Pocard Marc, Lariboisiére Hospital, Paris, France; Loungnarath Rasmy, Saint Luc Hospital, Montreal, Canada; Saint Luc Hospital, Montreal, Canada; Tasseti Vincent, Emile Muller Hospital, Mulhouse, France; Lermite Emily, Angers Hospital, Angers, France; Durand Sylvaine, Dupuytren Hospital, Limoges, France; Van der Speeten Kurt, Ziekenhuis Oost-Limburg, Genk, Belgium).


The authors made no disclosures.