Cytoreductive Surgery and Intraperitoneal Hyperthermic
Chemotherapy With Mitomycin C for Peritoneal Carcinomatosis
from Nonappendiceal Colorectal Carcinoma Perry Shen, MD, Jason Hawksworth, MD, James Lovato, PhD, Brian W. Loggie, MD, Kim R. Geisinger, MD, Ronald A. Fleming, PharmD, and Edward A. Levine, MD Background: Cytoreductive surgery (CS) and intraperitoneal hyperthermic chemotherapy (IPHC) are efficacious in patients with disseminated mucinous tumors of the appendix. We reviewed our experience using this approach for nonappendiceal colorectal cancer (NACC). Methods: We performed a retrospective chart review of a prospective database for patients undergoing CS and IPHC with mitomycin C for peritoneal carcinomatosis from colorectal primary lesions between December 1991 and April 2002. Results: There were 77 patients, with a median age of 54 years. Peritoneal carcinomatosis was synchronous and metachronous in 27% and 73% patients, respectively. Seventy-five percent of patients (n 58) had received chemotherapy prior to IPHC. Complete resection of all gross disease was accomplished in 37 patients (48%). The mean carcinoembryonic antigen level decreased from a preoperative value of 31.2 to a postoperative value of 6.9 (P .0001). Overall survival (OS) at 1, 3, and 5 years was 56%, 25%, and 17%, respectively. With a median follow-up of 15 months, the median OS was 16 months. Perioperative morbidity and mortality were 30% and 12%, respectively. Hematologic toxicity occurred in 15 patients (19%). Cox regression analysis identified poor performance status (P .018), bowel obstruction (P .001), malignant ascites (P .001), and incomplete resection of gross disease (P .011) as independent predictors of decreased survival. Patients with complete resection of all gross disease had a 5-year OS of 34%, with a median OS of 28 months. Conclusions: CS and IPHC with mitomycin C can improve outcomes for select patients with peritoneal spread from NACC. One third of patients who undergo complete resection of gross disease have long-term survival. Key Words: ChemotherapyColorectal cancerHyperthermiaPeritoneal carcinomatosis Surgery. The treatment of patients with peritoneal carcinomatosis (PC) from gastrointestinal malignancies is in evolution. PC is the most common cause of death in patients resected for intra-abdominal carcinomas. 1 A multicenter study prospec- tively following 370 patients with PC of nongynecologic origin revealed a mean and median overall survival (OS) of 6.0 and 3.1 months, respectively. 2 Surgical resection alone has been demonstrated to be ineffective for the treatment of PC, with median surviv- als of 1, 1, .7, and 6 months for PC from gastric, small bowel, pancreas, and colorectal cancer treated in this fashion. 3 Attempts at controlling PC with either external beam radiation therapy or brachytherapy have failed to demonstrate efficacy. 4 The use of systemic chemother- apy for PC has not been shown to be efficacious, as many patients present with PC after systemic chemotherapy fails. 5 Received May 8, 2003; accepted October 8, 2003. From Wake Forest University Baptist Medical Center (PS, JH, JL, KRG, EAL) and Kucera Pharmaceutical Company (RAF), Winston- Salem, NC; and Creighton University Cancer Center (BWL), Omaha, Nebraska. Address correspondence and reprint requests to: Perry Shen, MD, Surgical Oncology Service, Wake Forest University Medical Center, Medical Center Blvd., Winston-Salem, NC 27157; Fax: 336-716-9758; E-mail: pshen@wfubmc.edu. Published by Lippincott Williams & Wilkins 2004 The Society of Surgical Oncology, Inc. Annals of Surgical Oncology, 11(2):178186 DOI: 10.1245/ASO.2004.05.009 178 Intraperitoneal administration of chemotherapy has the benefit of delivering higher concentrations of cyto- toxic drug locally to the site of the tumor while mini- mizing systemic toxicity, in comparison with intravenous administration. Pharmacokinetic studies have demon- strated a 107-fold increase in the concentration of mito- mycin C (MMC) in the intraperitoneal perfusate versus plasma concentrations when administered systemically. 5 Recent studies have reported on the combination of cy- toreductive surgery and intraoperative intraperitoneal chemotherapy 612 administered under hyperthermic con- ditions (40C to 43C). The administration of intraperi- toneal chemotherapy at the time of surgery allows a potentially more even distribution of drug, without the potential of catheter-related complications and postoper- ative adhesions. Hyperthermia has been shown to poten- tiate the cytotoxicity of drugs such as MMC and cispla- tin. 13,14 These interactions are enhanced under hypoxic conditions, which is not true for most agents given after surgical resection. A report from Sugarbaker and colleagues 15 at the Washington Cancer Institute described the use of cytore- ductive surgery (CS) and intraoperative and periopera- tive intraperitoneal chemotherapy with 5-fluorouracil and MMC for 385 patients with PC from appendiceal malignancy. Patients with complete cytoreduction and pathology demonstrating adenomucinosis and adenocar- cinoma had 5-year OS rates of 86% and 50%, respec- tively. Piso et al. 16 also reported their experience with intraoperative hyperthermic intraperitoneal cisplatin af- ter peritonectomy procedures for PC from appendiceal carcinoma. They reported a mean OS time of 39 months, with a 4-year OS rate of 75%. Patients with complete versus incomplete cytoreduction had a 4-year OS rate of 92% and 40%, respectively (P .02). Another study report, by Zoetmulder and associates, 17 described their experience with IPHC with MMC for pseudomyxoma peritonei in 46 patients. The actuarial survival rate at 3 years was 81%. These published results for over 400 patients provide strong evidence for the use of an ag- gressive multimodality approach to this rare disease pro- cess, especially when standard surgical therapy results in only a 10-year OS of 10% to 30%. 18 In addition, we recently reviewed our experience with 109 patients treated with PC from various histologies and found that an appendiceal primary was an independent predictor of improved survival. 19 Data regarding the use of CS and IPHC for other types of histologies have not been as extensively studied. Spe- cifically, the treatment of PC from nonappendiceal colo- rectal carcinoma (NACC) with CS and IPHC with ad- ministration of MMC has been examined in only a limited fashion. Both Loggie et al. 20 and Sugarbaker et al. 21 have reported a significant difference in outcome for patients with appendiceal versus colorectal primaries. Other trials have suggested a benefit of CS and IPHC for PC from colorectal cancer, including one prospective randomized trial. 22,23 However, these trials included ap- pendiceal cancers in their study populations, making it difficult to determine the true effect of this approach. We reviewed our experience with CS and IPHC using MMC for PC from NACC to examine their demographics, clinical outcome, predictors of survival, factors contrib- uting to postoperative morbidity/mortality, and methods to improve patient selection. METHODS The study protocol was reviewed and approved by the Institutional Review Board of the Wake Forest Univer- sity School of Medicine. Eligible patients were enrolled in the protocol from December 1991 through April 2002. All patients with PC from NACC primaries were en- rolled in the study cohort. The Eastern Cooperative On- cology Group (ECOG) performance status was recorded for all patients enrolled in the protocol. 24 Cytoreductive Surgery All patients enrolled in the study protocol were oper- ated on by one of three surgeons (BWL, EAL, PS), each with significant experience with CS. CS consisted of the removal of all gross tumors with involved organs, peri- toneum, or tissue that was deemed technically feasible and safe for the patient. Any tumor adherent or invasive to vital structures that could not be removed was cytore- duced with the Cavitron Ultrasonic Surgical Aspirator device. Peritonectomy was performed as indicated. The resection status of patients was estimated following CS with use of the following classification: R 0 , complete removal of all visible tumor and negative cytology or negative microscopic margins; R 1 , complete removal of all visible tumor and positive cytology or microscopic margins; R 2a , minimal residual tumor, nodule(s) .5 cm; R 2b , gross residual tumor, nodule .5 cm but 2 cm; and R 2c , extensive disease remaining, nodules 2 cm. Intraperitoneal Hyperthermic Chemotherapy Patients were cooled to a core temperature of about 34C to 35C by passive measures (i.e., not warming airway gases or intravenous solutions and cooling the room). After CS was completed, peritoneal perfusion inflow and outflow catheters were placed percutaneously into the abdominal cavity. Temperature probes were placed on the inflow and outflow catheters. The abdom- 179 IPHC FOR COLORECTAL PERITONEAL CARCINOMATOSIS Ann Surg Oncol, Vol. 11, No. 2, 2004 inal skin incision was closed temporarily with a running suture to prevent leakage of peritoneal perfusate. A per- fusion circuit was established with approximately 3 L of Ringers lactate. Flow rates of approximately 600 to 900 mL/min were maintained with a roller pump managed by the pump technician. The circuit continued through a single roller pump, through a heat exchanger (SCI-MED, No. A-714; Gish Biomedical, Irvine, Ca), and then to the patient. Constant temperature monitoring was performed with temperature probes placed on both the inflow and out- flow catheters. Once inflow temperatures exceeded a temperature of 38.5C, 30 mg of MMC was added to the perfusate, and at 60 minutes an additional 10 mg of MMC was added to the perfusate to keep MMC perfus- ate concentrations 5 g/mL. A maximum inflow tem- perature of 40.5C was achieved during the perfusion, with outflow at the pelvis of 39.5C. The abdomen was gently massaged throughout the perfusion to improve drug distribution to all peritoneal surfaces. The total perfusion time after the initial addition of MMC was 120 minutes. In certain patients (elderly, those with extensive prior chemotherapy, those with inanition or poor perfor- mance status, and those having extensive peritoneal stripping during surgery), reductions in the dose of MMC (to 30 mg total) and/or the perfusion time (6090 min- utes) were made because of concerns about potential toxicity. The peritoneum was washed out with 3 L of lactated Ringer solution, and the abdomen was reopened for removal of perfusion catheters. In 2000, a new perfusion device (ViaCirq, Pittsburgh, PA) was introduced that allowed the perfusate to be heated to higher temperatures. Subsequent sessions of IPHC were conducted with a maximum inflow temper- ature of 43C and a minimum outflow temperature of 40.5C. Perioperative Evaluation Perioperative morbidity in the database was classified into four main groups: bowel leak, respiratory failure, infection, and sepsis. Bowel leak was defined by any occurrence of an anastomotic leak or bowel perforation, respiratory failure was defined by any extended period of mechanical respiration beyond the first 24 hours after surgery or any incidence of reintubation (regardless of cause), and infection was defined as any infection of the wound, abdomen, catheter, blood, or lung. Sepsis was defined as a worsening clinical condition requiring man- agement in the intensive care unit. Other complications not covered by these four categories were also included in the database. These included deep venous thrombosis, ileus, pneumonia, renal failure, and pleural effusion. Clinical Follow-Up Clinical follow-up occurred at 1 month, 3 months, and then every 3 months thereafter for up to 1 year. After 1 year, follow-up was at 3-month intervals or less fre- quently if the patient continued to remain without evi- dence of disease. Abdominal and pelvic CT scans were obtained at 3, 6, and 12 months postoperatively or when clinically indicated. Some patients received systemic chemotherapy after referral back to their medical oncologists. Statistical Analysis OS was calculated from the date of CS and IPHC to the last recorded date of follow-up or recorded date of death. All data were collected prospectively. Kaplan- Meier analysis was performed on all pertinent clinico- pathologic variables to determine estimates of survival over time. Group comparisons of OS were done with the log-rank test. Coxs proportional hazards regression model was used to perform multivariate analysis of clin- icopathologic factors to determine independent predic- tors of OS. Fishers exact test was used to correlate type of surgical procedure with postoperative morbidity. A P value .05 was considered significant for the purposes of this manuscript. RESULTS Patients A total of 77 patients with PC from NACC were enrolled in the protocol. Table 1 lists demographic char- acteristics and baseline data. The median age was 54 years. The primary sites were 74 in the colon and 3 in the rectum. All patients had a histologic diagnosis of adeno- carcinoma. PC was synchronous and metachronous in 27% and 73% of patients, respectively. In those with metachronous presentation, median disease-free interval between primary diagnosis and peritoneal disease was 14 months (range, 3 to 85 months). More than three quarters of patients had undergone prior surgery or chemother- apy. About one third of patients presented with either bowel obstruction or malignant ascites. Complete resec- tion of all gross disease was accomplished in 37 patients (48%). Table 2 correlates resection status with type of surgical procedure performed. Pre-IPHC cytology was not performed or the results were not available in ap- proximately one quarter of patients. Survival and Follow-Up For the cohort of 77 patients, 3-year and 5-year OS was 25% and 17%, respectively. The median OS was 16 months (95% confidence interval [CI], 1026) with a 180 P. SHEN ET AL. Ann Surg Oncol, Vol. 11, No. 2, 2004 median follow-up of 15 months. Table 3 displays the results of a log-rank analysis of multiple clinical and pathologic variables. Fifty-two patients (68%) have had a recurrence, with a median time to progression of 7 months (range, 331). The mean carcinoembryonic an- tigen decreased from a preoperative value of 31.2 to postoperative value of 6.9 (P .0001). Multivariate analysis demonstrated four clinicopatho- logic factors that were independent predictors of OS (Table 4): performance status (ECOG) of 0/1, R0/1 re- section, no bowel obstruction, and no malignant ascites. Figs. 14 depict the log-rank survival curves for these factors. Morbidity and Mortality Perioperative morbidity and mortality were 30% and 12%, respectively. Mortality was calculated as postoper- ative deaths directly attributable to the procedure, regard- less of the time interval since the operation. Causes of death included bowel perforation (n 2), bone marrow suppression (n 2), respiratory failure (n 4), and anastomotic leak (n 1). In the patient group with all the independent predictors of OS (ECOG performance status of 0/1, R0/1 resection, no bowel obstruction, no malig- nant ascites), the perioperative mortality was 4% (1/25). Fifteen subjects (19%) developed hematologic toxicity requiring growth factor support or platelet transfusion. Thirty-three patients (43%) required a blood transfusion, with a median of 2 units (range, 117) transfused. The median operative time and length of hospital stay for CS and IPHC were 9 hours (range, 516) and 10 days (range, 5150), respectively. Chi-square analysis and Fishers exact test were used to correlate the occurrence of a perioperative complica- tion (bowel leak, infection, respiratory failure, sepsis) with the performance of either a bowel resection (n 49) or bowel anastomosis (n 42). Patients with resec- tion but no anastomosis had ostomies created. There was no correlation of either bowel resection or anastomosis with perioperative complications as a group. However, when complications were individually analyzed against bowel resection and anastomosis, there was a significant correlation of sepsis with bowel anastomosis (P .0032) (see Table 5). Delivery of Intraperitoneal and Adjuvant Chemotherapy Sixty-one patients (79%) received the standard intra- peritoneal 2-hour perfusion of 40 mg MMC. Another three patients (4%) underwent a 2-hour perfusion, but two of them received 30 mg MMC and one especially large patient received 50 mg MMC. Twelve patients underwent a 1-hour perfusion with the following amounts of MMC administered: for three patients (4%), 40 mg, and for nine (12%), 30 mg. One patient received 40 mg MMC in a 1.5-hour perfusion. DISCUSSION In 1995 Sugarbaker et al. 21 published their experience using perioperative intraperitoneal 5-fluorouracil and MMC without hyperthermia for patients with PC sec- ondary to appendiceal and colorectal cancer. The 3-year OS rates for the appendiceal and colorectal PC patients were 70% and 30%, respectively (P .0001). In 2000, Loggie et al. 20 published a report from our institution examining the outcomes for patients with disseminated peritoneal cancer of gastrointestinal origin. The median OS of patients with colorectal primaries was 15 months. TABLE 1. Patient demographics and baseline data Baseline clinicopathologic variables No. (%) Total 77 (100) Age 55 y 41 (53) 55 y 36 (47) Sex Male 45 (58) Female 32 (42) Race White 73 (95) Black 4 (5) Previous treatment Surgery 67 (87) Chemotherapy 58 (75) Radiation 6 (8) Performance status (ECOG) 0 18 (23) 1 46 (60) 2 7 (9) 3 2 (3) 4 4 (5) Bowel obstruction Yes 22 (29) No 55 (71) Malignant ascites Yes 26 (34) No 51 (66) Liver metastases Yes 10 (13) No 65 (84) Unknown 2 (3) Histology Mucinous 43 (56) Signet ring cell 7 (9) Not otherwise specified 27 (35) Pre-IPHC cytology Positive 31 (40) Negative 28 (36) Unknown 18 (24) ECOG, Eastern Cooperative Oncology Group; IPHC, intraperitoneal hyperthermic chemotherapy. 181 IPHC FOR COLORECTAL PERITONEAL CARCINOMATOSIS Ann Surg Oncol, Vol. 11, No. 2, 2004 The median OS for appendiceal primaries had not been reached. A study in France by Elias and associates 22 used both early intraperitoneal and hyperthermic intraopera- tive intraperitoneal chemotherapy with a combination of agents: 5-fluorouracil, MMC, and cisplatin. The results of two separate trials were combined in one report. 22 The study cohort was patients with PC arising from colorectal adenocarcinomas, of which 14% were mucinous cancers of the appendix. All patients in this study had an R0/1 resection with all gross disease removed. The 5-year OS was 27%, with a median OS of 36 months. These results are comparable to those for our subset of patients with a similar type of resection status. A prospective study on the natural history of PC from colorectal cancer demonstrated a median OS of 5 months. 2 This subset of patients did not include pseudomyxoma peritonei. The median OS of patients in our study was 16 months, with a 5-year OS of 17%. Three quarters of the patients had already undergone systemic chemotherapy for metastatic disease that failed. Univariate predictors of survival were performance sta- tus, resection status, bowel obstruction status, malignant ascites status, liver metastases status, and pre-IPHC cy- tology status. The best outcomes were for patients un- dergoing an R0 resection, which yielded a 5-year OS of 55%, with a median OS that has not yet been reached. On multivariate Cox regression analysis, only four factors remained independent predictors of OS: ECOG performance status of 0/1, R0/1 resection status, absence of bowel obstruction, and absence of malignant ascites. Two of these factorsperformance status and bowel obstructionappear to favor patients who will best tol- erate the significant physiological stress induced by the multimodality application of not only extensive CS but also IPHC. The other two factorsresection status and ascitesappear to identify patients with more limited disease that best responds to CS and IPHC from an oncologic perspective. Three of these four factors can be determined preoperatively and help to optimize patient selection. The fourth factor, resection status, is a variable that is most accurately determined intraoperatively. The use of a peritoneal carcinomatosis index has been re- ported to be predictive of resectability in patients with PC. 15 However, because this index can be determined only at the time of laparotomy, it does not help prevent unnecessary surgery. Complications after CS and IPHC can be due to the extensive surgery or the intraperitoneal chemotherapy or the combination of both. Major peritonectomy proce- dures were not performed when cytoreduction with the Cavitron Ultrasonic Surgical Aspirator was possible. The most significant detrimental effect is on the hematologic, TABLE 2. Surgical procedures correlated with resection status Procedure R0 R1 R2a R2b R2c Total Tumor debulking 7 15 3 4 10 39 Retroperitoneal LND 2 3 5 Small bowel resection 3 9 7 5 6 30 Splenectomy 3 4 4 1 1 13 Colon resection 7 5 8 4 4 28 Rectum resection 2 6 3 1 12 Omentectomy 6 13 10 4 10 43 Bladder 1 1 Umbilicus resection 2 1 3 6 Pancreas resection 2 2 1 3 8 Colostomy 3 3 1 3 10 Gastrostomy 4 4 Tube thoracostomy 1 1 2 Salpingo-oopherectomy 1 7 3 4 15 Hysterectomy 4 1 5 Peritonectomy 1 1 2 Liver resection 1 2 2 2 7 Cholecystectomy 2 3 1 1 7 Hepatic cryoablation 2 1 3 Appendectomy 1 1 3 5 Diaphragm excision 1 1 2 Gastrectomy 1 1 2 Nephrouterectomy 1 1 Abdominal wall excision 1 1 Total 34 85 58 22 52 LND, lymph node dissection. 182 P. SHEN ET AL. Ann Surg Oncol, Vol. 11, No. 2, 2004 pulmonary, and gastrointestinal organ systems. 25 The perioperative morbidity and mortality in our study cohort were 30% and 12%, respectively. Although the morbid- ity compares favorably with that in other published se- ries, the mortality is substantial for any surgical proce- dure, especially one that is performed with palliative intent. However, we counted all deaths related to CS and IPHC, not limiting the number to deaths occurring only within 30 days of the procedure. Respiratory failure accounted for 44% of the peri- operative deaths. MMC has been reported to cause adult respiratory distress syndrome in patients exposed to FIO 2 concentrations greater than 50% periopera- tively. 26 This complication usually manifests after multiple doses in patients given the drug systemically with other chemotherapy combinations. The marrow toxicity that can occur with MMC usually occurs 48 weeks after systemic administration, 26 but the two patients who developed fatal myelosuppression had evidence of hematologic toxicity within 72 hours of IPHC. Previous studies have shown that platelet counts begin to decline after the procedure and typi- cally return to normal values within 12 weeks. 27,28 The differences in timing of nadir counts between IPHC and intravenous bolus MMC may be due to kinetic differences from the two administration schedules. We have previously reported that there is a direct relationship between hematologic toxicity and MMC concentrations in plasma. 29 The AUCs of MMC in TABLE 3. Univariate analysis of clinical and pathologic variables for overall survival Factor n Expired Percent Surviving (SE) Survival (mo) P value 3 y 5 y Median 95% CI Overall 77 45 25.0 (2.5) 17.3 (3.3) 15.9 10.225.5 Age .988 55 41 24 22.0 (3.8) 17.6 (4.4) 17.8 6.432.8 55 36 21 28.1 (3.3) 28.1 (4.8) 14.6 10.036.2 Race .716 White 73 42 26.2 (2.5) 18.2 (3.3) 15.9 10.225.5 Nonwhite 4 3 .0 (NA) .0 (NA) 18.6 3.85NA Gender .619 Male 45 25 25.9 (3.3) 17.3 (4.4) 12.7 5.432.8 Female 32 20 24.1 (3.8) 16.0 (5.6) 15.9 11.747.7 Prior surgery .284 No 10 6 40.5 (4.3) 27.0 (5.9) 25.5 10.0NA Yes 67 39 22.2 (3.0) 15.5 (3.9) 12.7 10.223.0 Prior chemotherapy .439 No 19 10 19.6 (6.2) 9.8 (9.5) 16.4 14.6NA Yes 58 35 26.1 (2.7) 19.3 (3.4) 10.4 5.432.3 Prior radiation .314 No 71 40 28.1 (2.4) 19.5 (3.3) 14.6 10.232.3 Yes 6 5 .0 (NA) .0 (NA) 16.4 3.5NA Performance status .0004 01 64 33 32.0 (2.4) 22.1 (3.2) 17.8 12.536.2 24 13 12 .0 (NA) .0 (NA) 3.8 2.0NA Resection status .0005 R0 13 4 69.2 (2.2) 55.3 (3.1) NR 16.4NA R1 24 11 19.1 (6.2) 19.1 (6.2) 17.8 10.2NA R2a 11 7 28.0 (5.9) 14.0 (9.2) 12.7 4.2NA R2b 9 6 .0 (NA) .0 (NA) 4.1 3.5NA R2c 20 17 6.0 (9.7) .0 (NA) 5.0 2.418.2 Bowel obstruction .0001 No 55 30 32.4 (2.4) 22.4 (3.2) 20.9 12.736.2 Yes 22 15 .0 (NA) .0 (NA) 4.1 2.0NA Malignant ascites .0001 No 51 27 32.7 (2.5) 22.1 (3.4) 20.9 15.938.9 Yes 26 18 6.9 (9.5) 6.9 (9.5) 4.2 2.410.4 Liver metastases .031 No 65 35 26.6 (2.7) 19.4 (3.5) 15.9 10.232.3 Yes 10 9 12.0 (9.3) .0 (NA) 12.7 1.5NA Pre-IPHC cytology .004 Negative 28 15 44.9 (2.4) 33.2 (3.2) 27.8 11.7NA Positive 31 23 9.4 (6.7) 4.7 (9.7) 9.9 4.219.6 SE, standard error; CI, confidence interval; NA, not applicable; NR, not reached; IPHC, intraperitoneal hyperthermic chemotherapy. 183 IPHC FOR COLORECTAL PERITONEAL CARCINOMATOSIS Ann Surg Oncol, Vol. 11, No. 2, 2004 plasma after a 2-hour IPHC perfusion are similar to that observed after an intravenous bolus of MMC. The dosing rationale for MMC during IPHC was to achieve a peri- toneal concentration of 510 g/mL in a fixed volume of perfusate, versus the standard practice of dosing by body surface area. Several recent reviews in the oncology literature question whether dosing by body surface area is any better than fixed dosing. 3032 Given the tremen- dous pharmacokinetic variations between patients, dos- ing differences by body surface area may not be signif- icant when the real sources of variability are the organs (or enzymes, etc.) responsible for drug elimination (renal function for renally eliminated drugs and hepatic func- tion for hepatically cleared drugs). Three of the patients died of either bowel leak or perforation, which appears more plausible given the of- ten extensive nature of the surgery and evidence from animal models that intraperitoneal chemotherapy with MMC has an adverse effect on anastomotic bowel heal- ing. 33 It is interesting that only one of these deaths was due to an anastomotic leak, whereas the other two pa- tients who died developed bowel perforations at a site other than the anastomosis. The two patients with perfo- rations presented with bowel obstruction, whereas the patient with an anastomotic leak had obstructive jaun- dice. Cox regression analysis has clearly identified bowel obstruction as an independent predictor of decreased OS, and we no longer perform IPHC in such patients. When we conducted a Fishers exact test to correlate the pres- ence of a bowel resection or anastomosis with compli- cations, we found that the presence of a bowel anasto- mosis was significantly correlated with sepsis. However, on Cox regression analysis (Table 4), the hazard ratio for bowel anastomosis was .9. It appears the creation of a FIG. 1. Overall survival as related to performance status (Eastern Cooperative Oncology Group). FIG. 3. Overall survival as related to bowel obstruction. TABLE 4. Cox regression analysis of overall survival Factor Hazard ratio 95% CI P value Age 55 vs. 55 .97 .501.85 .92 Race Nonwhite vs. white .84 .252.89 .79 Sex Female vs. male 1.69 .863.32 .13 Performance status 2 vs. 0/1 2.36 1.164.82 .018 Resection status R2 vs. R0/1 2.35 1.224.55 .011 Bowel obstruction present 3.47 1.627.44 .001 Ascites present 3.23 1.636.38 .001 Liver metastases present 2.13 .905.07 .087 Lymph node involvement 1.35 .712.58 .370 Anastamosis present .90 .431.89 .790 CI, confidence interval. FIG. 2. Overall survival as related to resection status. 184 P. SHEN ET AL. Ann Surg Oncol, Vol. 11, No. 2, 2004 bowel anastomosis may increase the risk of perioperative complications but not perioperative mortality. Although we do not believe bowel anastomosis during IPHC is contraindicated, we believe its association with periop- erative complications should be considered in light of the patients overall status, especially when the creation of a stoma is another option. When perioperative mortality among patients with the independent predictors of OS was assessed, the death rate was 4%. This conclusion should not be surprising, because these are the patients who presented in the best physiological condition with grossly resectable disease. Our experience with these 77 patients reflects an aggres- sive approach for patients with advanced cancer who have few meaningful therapeutic options. Some of these patients presented in poor health due to their malignant disease, and the perioperative mortality reflects the tre- mendous stress this multimodality approach places on the patient. However, our data clearly demonstrate that for select patients with PC from NACC, the application of CS and IPHC with MMC can extend survival. Other reports from our institution have shown that quality of life, a vital parameter for the evaluation of palliative interventions, is also preserved for the majority of pa- tients, both in the short term and in the long term. 34,35 One criticism of the use of CS and IPHC is the lack of randomized, controlled trials confirming its efficacy in comparison with supportive care, surgery alone, or sys- temic chemotherapy. A recent prospective trial in Europe randomized 104 patients with PC from colorectal adeno- carcinoma to CS and IPHC with MMC, followed by systemic therapy with 5-fluorouracil/leucovorin, in com- parison with CS and systemic 5-fluorouracil/leucovorin alone. With a mean follow-up of 24 months, the 2-year OS was 43% in the IPHC arm (median OS, 21 months) and 16% in the standard therapy arm (median OS, 10 months) (P .0145). 23 Previous studies of untreated patients with metastatic colorectal cancer have revealed median OS of 1419 months with the use of new sys- temic chemotherapeutic agents such as irinotecan and oxaliplatin. 3638 Our study revealed that patients under- going R0/1 resection had a median OS of 28 months. In addition, our study provides information on which clin- icopathologic factors predict improved outcome from this procedure. Future research directions may include the use of molecular markers to determine which patients will de- rive the most benefit. Just as the preoperative carcino- embryonic antigen level was seen to significantly de- crease after CS and IPHC, investigations with other surrogate markers may yield more prognostic informa- tion. We previously reported that the presence of a poly- morphism that causes reduced activity of quinone oxi- doreductase 1 (NQO1) is associated with a decrease in survival for patients undergoing CS and IPHC for PC. 39 NQO1 is important in the activation of MMC. Such a mutation found in tumor tissue of patients with PC may identify those who should be treated with agents other than MMC or undergo alternative treatments. Multi- center trials are needed to confirm these findings and improve perioperative and long-term outcome. REFERENCES 1. Pilati P, Rossi CR, Mocellin S, et al. Multimodal treatment of peritoneal carcinomatosis and sarcomatosis. Eur J Surg Oncol 2001;27:12534. 2. Sadeghi B, Arvieux C, Glehen O, et al. Peritoneal carcinomatosis from non-gynecologic malignancies: results of the EVOCAPE 1 multicentric prospective study. Cancer 2000;88:35863. 3. Chu DZ, Lang NP, Thompson C, Osteen PK, Westbrook KC. Peritoneal carcinomatosis in nongynecologic malignancy: a pro- spective study of prognostic factors. Cancer 1989;63:3647. 4. Wong CS, Harwood AR, Cummings BJ, Keane TJ, Thomas GM, Rider WD. Total abdominal irradiation for cancer of the colon. Radiother Oncol 1984;2:20914. 5. Dedrick RL. Theoretical and experimental bases of intraperitoneal chemotherapy. Semin Oncol 1985;12:16. 6. Schneebaum S, Arnold MW, Staubus A, Young DC, Dumond D, Martin EW Jr. Intraperitoneal hyperthermic perfusion with mito- TABLE 5. Fishers exact test correlating bowel anastomosis with sepsis a Sepsis Anastomosis Total No Yes No 35 33 68 Yes 0 9 9 Total 35 42 a P .0032 FIG. 4. Overall survival as related to malignant ascites. 185 IPHC FOR COLORECTAL PERITONEAL CARCINOMATOSIS Ann Surg Oncol, Vol. 11, No. 2, 2004 mycin C for colorectal cancer with peritoneal metastases. Ann Surg Oncol 1996;3:4450. 7. Yonemura Y, Fujimura T, Fushida S, et al. Hyperthermo-chemo- therapy combined with cytoreductive surgery for the treatment of gastric cancer with peritoneal dissemination. World J Surg 1991; 15:5305. 8. Bartlett DL, Buell JF, Libutti SK, et al. A phase I trial of contin- uous hyperthermic peritoneal perfusion with tumor necrosis factor and cisplatin in the treatment of peritoneal carcinomatosis. Cancer 1998;83:125161. 9. Park BJ, Alexander HR, Libutti SK, et al. Treatment of primary peritoneal mesothelioma by continuous hyperthermic peritoneal perfusion (CHPP). Ann Surg Oncol 1999;6:58290. 10. Stephens AD, Alderman R, Chang D, et al. Morbidity and mor- tality analysis of 200 treatments with cytoreductive surgery and hyperthermic intraoperative intraperitoneal chemotherapy using the coliseum technique. Ann Surg Oncol 1999;6:7906. 11. Cavaliere F, Di Filippo F, Botti C, et al. Peritonectomy and hyperthermic antiblastic perfusion in the treatment of peritoneal carcinomatosis. Eur J Surg Oncol 2000;26:48691. 12. Beaujard AC, Glehen O, Caillot JL, et al. Intraperitoneal chemo- hyperthermia with mitomycin C for digestive tract cancer patients with peritoneal carcinomatosis. Cancer 2000;88:25129. 13. Watanabe M, Tanaka R, Hondo H, Kuroki M. Effects of antineo- plastic agents and hyperthermia on cytotoxicity toward chronically hypoxic glioma cells. Int J Hyperthermia 1992;8:1318. 14. Teicher BA, Kowal CD, Kennedy KA, Sartorelli AC. Enhance- ment by hyperthermia of the in vitro cytotoxicity of mitomycin C toward hypoxic tumor cells. Cancer Res 1981;41:10969. 15. Sugarbaker PH, Chang D. Results of treatment of 385 patients with peritoneal surface spread of appendiceal malignancy. Ann Surg Oncol 1999;6:72731. 16. Piso P, Bektas H, Werner U, et al. Improved prognosis following peritonectomy procedures and hyperthermic intraperitoneal che- motherapy for peritoneal carcinomatosis from appendiceal carci- noma. Eur J Surg Oncol 2001;27:28690. 17. Witkamp AJ, de Bree E, Kaag MM, van Slooten GW, van Co- evorden F, Zoetmulder FA. Extensive surgical cytoreduction and intraoperative hyperthermic intraperitoneal chemotherapy in pa- tients with pseudomyxoma peritonei. Br J Surg 2001;88:45863. 18. Wirtzfeld DA, Rodriguez-Bigas M, Weber T, Petrelli NJ. Dissem- inated peritoneal adenomucinosis: a critical review. Ann Surg Oncol 1999;6:797801. 19. Shen P, Levine EA, Hall J, et al. Factors predicting survival after intraperitoneal hyperthermic chemotherapy with mitomycin C af- ter cytoreductive surgery for patients with peritoneal carcinoma- tosis. Arch Surg 2003;138:2633. 20. Loggie BW, Fleming RA, McQuellon RP, Russell GB, Geisinger KR. Cytoreductive surgery with intraperitoneal hyperthermic che- motherapy for disseminated peritoneal cancer of gastrointestinal origin. Am Surg 2000;66:5618. 21. Sugarbaker PH, Jablonski KA. Prognostic features of 51 colorectal and 130 appendiceal cancer patients with peritoneal carcinomatosis treated by cytoreductive surgery and intraperitoneal chemotherapy. Ann Surg 1995;221:12432. 22. Elias D, Blot F, El Otmany A, et al. Curative treatment of peritoneal carcinomatosis arising from colorectal cancer by complete resection and intraperitoneal chemotherapy. Cancer 2001;92:716. 23. Zoetmulder FAN, Verwaal V, Ruth S. Hyperthermic intraperito- neal chemotherapy (HIPEC) with mitomycin C significantly im- proves survival in patients with peritoneal carcinomatosis of colo- rectal origin. ASCO Prog Proc 2002;21:147a. 24. Schag CC, Heinrich RL, Ganz PA. Karnofsky performance status revisited: reliability, validity, and guidelines. J Clin Oncol 1984; 2:18793. 25. Loggie BW, Fleming R. Complications of heated intraperitoneal chemotherapy and strategies for prevention. In: Peritoneal Carci- nomatosis: Principles Of Management. Boston: Kluwer Academic Publishers, 1996:22233. 26. Berkery R, Cleri L, Skarin A. Chemotherapeutic agents. In: On- cology Pocket Guide To Chemotherapy. Philadelphia: SmithKline Beecham, 1997:17780. 27. Sayage AC, Gilly FN, Carry PY, et al. Intraoperative chemohy- perthermia in the management of digestive cancers. A general review of literature. Oncology 1993;50:3337. 28. Fujimoto S, Shrestha RD, Kokubun M, et al. Clinical trial with surgery and intraperitoneal hyperthermic perfusion for peritoneal recurrence of gastrointestinal cancer. Cancer 1989;64:15460. 29. Fleming R, Loggie BW, Russell GB, Rice MA. Systemic exposure of MMC during intraperitoneal administration is correlated with hematologic toxicity and survival in patients with peritoneal car- cinomatosis. Proc Am Assoc Cancer Res 1998;39:599. 30. Gurney H. Dose calculation of anticancer drugs: a review of the current practice and introduction of an alternative. J Clin Oncol 1996;14:2590611. 31. Ratain MJ. Body-surface area as a basis for dosing of anticancer agents: science, myth, or habit? J Clin Oncol 1998;16:22978. 32. Baker SD, Verweij J, Rowinsky EK, et al. Role of body-surface area in dosing of investigational anticancer agents in adults: 1991 2001. J Natl Cancer Inst 2002;94:18838. 33. Fumagalli U, Trabucchi E, Soligo M, et al. Effects of intraperito- neal chemotherapy on anastomotic healing in the rat. J Surg Res 1991;50:827. 34. McQuellon RP, Loggie BW, Fleming RA, Russell GB, Lehman AB, Rambo TD. Quality of life after intraperitoneal hyperthermic chemotherapy (IPHC) for peritoneal carcinomatosis. Eur J Surg Oncol 2001;27:6573. 35. McQuellon RP, Loggie BW, Lehman AB, et al. Long-term survi- vorship and quality of life after cytoreductive surgery plus intra- peritoneal hyperthermic chemotherapy for peritoneal carcinomato- sis. Ann Surg Oncol 2003;10:15562. 36. Saltz LB, Cox JV, Blanke C, et al. Irinotecan plus fluorouracil and leucovorin for metastatic colorectal cancer. Irinotecan Study Group. N Engl J Med 2000;343:90514. 37. de Gramont A, Figer A, Seymour M, et al. Leucovorin and flu- orouracil with or without oxaliplatin as first-line treatment in advanced colorectal cancer. J Clin Oncol 2000;18:293847. 38. Giacchetti S, Perpoint B, Zidani R, et al. Phase III multicenter randomized trial of oxaliplatin added to chronomodulated fluorou- racil-leucovorin as first-line treatment of metastatic colorectal can- cer. J Clin Oncol 2000;18:13647. 39. Fleming RA, Drees J, Loggie BW, et al. Clinical significance of a NAD(P)H: quinone oxidoreductase 1 polymorphism in patients with disseminated peritoneal cancer receiving intraperitoneal hy- perthermic chemotherapy with mitomycin C. Pharmacogenetics 2002;12:317. 186 P. SHEN ET AL. Ann Surg Oncol, Vol. 11, No. 2, 2004