Factors Predicting Survival After Intraperitoneal Hyperthermic Chemotherapy With Mitomycin C After Cytoreductive Surgery for Patients With Peritoneal Carcinomatosis FREE

TREATMENT OF patients with peritoneal carcinomatosis (PC) from gastrointestinal and nongastrointestinal malignancies is in evolution. Peritoneal carcinomatosis is the most common cause of death in patients resected for intra-abdominal carcinomas.¹ A multicenter study² prospectively following 370 patients with PC reported mean and median overall survival (OS) of 6.0 and 3.1 months, respectively.

Surgical resection alone has been shown³ to be ineffective for the treatment of PC, with median survival of 1.0, 1.0, 0.7, and 6.0 months for PC from gastric, small-bowel, pancreas, and colorectal cancer treated in this fashion. Attempts at controlling PC with either external beam radiation therapy or brachytherapy have not demonstrated efficacy.⁴ The use of systemic chemotherapy for PC has not been shown to be efficacious, as many patients present with PC after failing systemic chemotherapy.⁵ Intraperitoneal administration of chemotherapy has the benefit of higher concentrations of cytotoxic drug delivered locally to the site of the tumor while minimizing systemic toxic effects compared with intravenous administration. Pharmacokinetic studies⁵ have demonstrated a 107-fold increase in the concentration of mitomycin C (MMC) in the intraperitoneal perfusate vs plasma concentrations when administered systemically. Initial studies from Sugarbaker and colleagues^6- 7 at the Washington Cancer Institute described the use of cytoreductive surgery (CS) and early postoperative intraperitoneal chemotherapy using fluorouracil and MMC. Three-year OS of 66% for men and 84% for women with PC from appendiceal cancer has been reported with such a multimodality approach.⁶

Recent studies^8- 14 have reported on the combination of CS and intraoperative intraperitoneal chemotherapy administered under hyperthermic conditions (40°C-43°C). The use of intraperitoneal chemotherapy at the time of surgery allows a potentially more even distribution of drug without having to deal with catheter-related complications and postoperative adhesions. Hyperthermia has been shown to potentiate the cytotoxicity of drugs such as MMC and cisplatin.^15- 16 These interactions are enhanced under hypoxic conditions, which is not true for most agents given after surgical resection.

Data on prognostic factors for this procedure are limited. Factors such as completeness of resection, lymphatic or hematogenous spread, histologic features of the tumor, site of primary tumor, previous surgery, and malignant ascites have been proposed as significant variables determining outcome.^14,17- 19 We examined the results of a prospective phase 2 protocol of CS and intraperitoneal hyperthermic chemotherapy (IPHC) with MMC for PC to determine the clinicopathologic factors that were independent predictors for OS.

The study protocol was reviewed and approved by the institutional review board of the Wake Forest University School of Medicine. Eligible patients entered the protocol between December 1991 and November 1997. The protocol was open to patients with PC from gastrointestinal and nongastrointestinal primary sites without evidence of extra-abdominal disease. Eastern Cooperative Oncology Group performance status was recorded in all patients entering the protocol.²⁰

All patients in the study were operated on by the same surgeon (B.W.L.). Cytoreductive surgery consisted of the removal of all gross tumors and involved organs, peritoneum, or tissue deemed technically feasible and safe for the patient. Any tumors adherent or invasive to vital structures that could not be removed were cytoreduced using the cavitational ultrasonic surgical aspirator (CUSA; Valleylab, Boulder, Colo). Peritonectomy was performed as indicated. The resection status of patients was estimated after CS using the following classification: R0, complete removal of all visible tumor and negative cytologic findings or microscopic margins; R1, complete removal of all visible tumor and positive cytologic findings or microscopic margins; R2a, minimal residual tumor, nodule(s) measuring 0.5 cm or less; R2b, gross residual tumor, nodule greater than 0.5 cm but less than or equal to 2 cm; and R2c, extensive disease remaining, nodules greater than 2 cm.

Patients were cooled to a core temperature of approximately 34°C to 35°C by passive measures (ie, 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 catheter tips. The abdominal skin incision was closed temporarily with a running suture to prevent leakage of peritoneal perfusate. A perfusion circuit was established with approximately 3 L of Ringer lactate. Flow rates of approximately 600 to 800 mL/min were maintained using 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, Calif), and then to the patient.

Constant temperature monitoring was performed at all temperature probes. Once inflow temperatures exceeded 38.5°C, 30 mg of MMC was added to the perfusate, and at 60 minutes an additional 10 mg of MMC was added to keep MMC perfusate concentrations higher than 5 µg/mL. A maximum inflow temperature of 40.5°C was realized during perfusion, with an outflow temperature at the pelvis of 39.5°C. The abdomen was gently massaged throughout perfusion to improve drug distribution to all peritoneal surfaces. Total perfusion time after the initial addition of MMC was 120 minutes. In certain patients (elderly individuals, those with extensive previous chemotherapy, those with inanition or poor performance status, and patients having extensive peritoneal stripping during surgery), reductions in the dose of MMC (to 30 mg total) or perfusion time (to 60-90 minutes) were made owing to concerns about potential toxic effects.

Clinical follow-up occurred at 1 and 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 frequently if the patient continued to remain without evidence of disease. Abdominal and pelvic computed tomographic scans were obtained 3, 6, and 12 months after surgery or when clinically indicated. Some patients received systemic chemotherapy after referral back to their medical oncologists.

The rate of 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 clinicopathologic variables to determine estimates of survival over time. Group comparisons of OS rates were performed using the log-rank test. The Cox proportional hazards regression model was used to perform multivariate analysis of clinicopathologic factors to determine independent predictors of OS.

A total of 109 patients with PC from gastrointestinal and nongastrointestinal origins were entered into the protocol. Table 1 lists patient demographics and baseline data. The site of primary tumor categorized as "other" includes 1 patient each with pancreas, small intestine, periampullary, and bladder tumors.

For the cohort of 109 patients, 1- and 3-year OS was 61% and 33%, respectively. Median OS was 16 months (95% confidence interval, 12-24 months), with median follow-up of 52 months. Table 2 gives the results of a log-rank analysis of the multiple clinicopathologic variables listed in Table 1.

Multivariate analysis demonstrated 4 clinicopathologic factors that were independent predictors of OS (Table 3): the appendix as a primary site, complete resection of all gross tumor, absence of liver metastases, and nonadenocarcinoma histologic features. Figure 1, Figure 2, Figure 3, and Figure 4 depict the survival curves for these factors. When the appendix was the primary site, the histologic findings consisted of adenocarcinoma (n = 13), pseudomyxoma peritonei (n = 9), and carcinoid (n = 1).

Postoperative morbidity and mortality were 36% and 8%, respectively. Mortality was calculated as postoperative deaths directly attributable to the procedure, regardless of time since the operation. Causes of death included bowel perforation (n = 3), bone marrow suppression (n = 2), respiratory failure (n = 2), methicillin-resistant Staphylococcus aureus infection (n = 1), and pulmonary embolus (n = 1). Twenty-three patients (21%) developed hematologic toxic effects requiring growth factor support or platelet transfusion. The median number of units of single donor platelets transfused was 2 (range, 1-104 U). Sixty-nine patients (63%) required a blood transfusion, with a median of 3 U (range, 1-38 U) transfused. The median operative time and length of hospital stay for CS and IPHC was 9 hours (range, 4-18 hours) and 9 days (range, 5-105 days), respectively.

Of 103 patients whose IPHC dosages and perfusion times were available, 67 (65%) received the standard intraperitoneal perfusion of 2 hours and 40 mg of MMC. Another 14 patients (14%) underwent 2 hours of perfusion but received only 30 mg of MMC. The other 22 patients underwent 1 hour of perfusion with the following amounts of MMC administered: 40 mg in 2 patients (2%), 30 mg in 19 (18%), and 20 mg in 1 (1%). When the clinical outcome of patients was stratified by those who received a standard protocol perfusion vs those who did not, there was no significant difference (P = .60).

As surgical techniques and perioperative care have improved, there has been a greater trend toward more aggressive surgical treatment of solid tumors. The role of cytoreduction, which implies treatment that incompletely eradicates tumor, has traditionally been reserved for chemotherapy or radiation therapy because it was believed that using surgery for this purpose was associated with excessive morbidity and mortality.²¹ However, there is a theoretic and a clinical basis to suggest that the role of CS should be reexamined.

The benefit of CS has been demonstrated in the treatment algorithm for peritoneally disseminated ovarian cancer. The combination of aggressive surgical debulking of disease with systemic chemotherapy and radiation therapy resulted in the best survival advantage.²² The experience in ovarian cancer illustrates the treatment of a disease for which chemotherapeutic options exist and in which surgery is performed only to reduce the bulk of disease to a level at which the chemotherapy could be expected to be most effective.

The benefit of intraperitoneal chemotherapy in the setting of PC can be explained by gompertzian cellular kinetics: in the initial stages, tumoral cell growth is exponential, but as the tumor enlarges, its blood supply and growth slow down, and a gradually larger percentage of tumor cells enters a nonproliferative phase of the cellular cycle.²³ In theory, debulking should rid the body of more cells, thereby stimulating the remaining cells to enter the proliferating phase of the cell cycle, potentially becoming more responsive to administration of antineoplastic agents.^24- 25 Studies²⁶ of intraperitoneal delivery of cytotoxic agents have shown that direct tumor absorption of drugs occurs to a level of 5 mm beneath the tumor surface. Low transperitoneal absorption, owing to the plasma-peritoneal barrier,²³ allows systemic drug concentrations 18- to 620-fold lower than intraperitoneal concentrations.²⁷

Such historic and theoretic considerations gave credence to the concept of intraperitoneal chemotherapy as an adjunct to CS. Initial studies^7,28 reported the use of early postoperative intraperitoneal chemotherapy given through Tenckhoff catheters placed at the time of surgery. Although there have been no controlled studies, to our knowledge, comparing postoperative and intraoperative intraperitoneal therapy, a potential criticism of postoperative therapy is inhomogeneous drug distribution caused by early postoperative adhesions. The resultant tumor entrapment by fibrin in dissected areas decreases its exposure to chemotherapeutic agents.^29- 30

Spratt et al³¹ were the first researchers to associate hyperthermia with intraperitoneal chemotherapy after CS. Experimental evidence³² suggests that tumor tissue is more sensitive to heat than normal tissue because of intrinsic thermosensitivity and lower efficacy in heat exchange through vasodilation. The effects of hyperthermia on malignant tissue seem to be mediated by direct cytotoxicity and the microcirculation peculiar to neoplasms.³³ Moreover, hyperthermia synergistically enhances the chemosensitivity of tumor cells to MMC.¹⁵ Mechanisms of action include increased cellular accumulation and activation of MMC and altered repair of DNA damage caused by MMC.¹⁶

The most significant predictor of outcome in our series, which is consistent with previous studies,^14,17- 19 is the completeness of the cytoreduction. The ability of hyperthermia and intraperitoneal administration of chemotherapy to work synergistically directly depends on the residual tumor volume.³⁴ Patients who underwent complete CS usually had disease limited to the parietal peritoneum or structures that could be completely removed, such as the colon, omentum, spleen, or gallbladder. When there was extensive involvement of the small intestine or its mesentery, an R0/1 resection was not possible. In fact, one relative contraindication to CS and IPHC is radiologic evidence of substantial tumor infiltration of the small-bowel mesentery.

An appendiceal primary site is known to be an important factor in predicting a favorable outcome.^7,19 The appendix as a primary cancer site for the seeding on peritoneal surfaces is unique in that the tumor is often of low biological aggressiveness.³⁵ Disseminated tumors from this primary site have a low incidence of lymphatic and hematogenous spread and tend to spare the small bowel, which makes complete cytoreduction more feasible.

Although nonadenocarcinoma histologic findings were found to be an independent predictor for improved survival, only approximately 20% of patients in the study cohort fit this category. The favorable prognosis of patients with pseudomyxoma peritonei has been well described.⁶ We³⁶ recently described our experience using CS and IPHC with MMC for patients with malignant peritoneal mesothelioma and reported median survival of 34 months. Owing to the rarity of presentation, no mature outcome data are available for patients with PC from sarcoma and carcinoid, except as part of a larger series such as this.

Frequently, patients will be found to have metastatic implants on the liver surface during abdominal exploration. These are thought to occur from the same process of intraperitoneal dissemination that leads to PC, and their subsequent removal is part of the process of CS. However, when hematogenous spread of tumor leads to hepatic parenchymal metastases, patient outcome is uniformly poor.^7,37 This indicates spread of tumor beyond the peritoneal cavity, which makes it a systemic problem. Based on the present analysis, we generally do not perform CS and IPHC in this subset of patients.

Nearly a third of the patients in this series did not receive the full application of heat and chemotherapy as described in the protocol. Log-rank analysis did not find any significant difference in outcome between patients who underwent 2 hours of heated application of 40 mg total of MMC and those who were subjected to less hyperthermia, less chemotherapy, or both. It is not clear whether administering a shorter period of hyperthermia with or without a decreased dosage of MMC would have had the same cytotoxic effect with less hematologic toxic effects. In addition, the literature¹ reports that achieving an intraperitoneal temperature of at least 41°C is desirable for optimizing drug diffusion into tissues and for maximizing the synergistic effect with chemotherapy. In our protocol, the maximum inflow temperature achieved was 40.5°C, with an outflow temperature of 39.5°C. At the time of this study, the perfusion machines being used were primarily designed for cardiopulmonary bypass and had safety cutoff temperatures that limited the maximum inflow temperature that could have been achieved. The final plateau temperature was ultimately determined by equipment limitations and any heat losses from the perfusion catheters. Currently, our perfusions are performed using a machine (ViaCirQ, Pittsburgh, Pa) that does not have temperature limitations, and increased intraperitoneal temperatures are being studied.

Table 4 lists previously published series of CS and IPHC for PC, including the present study. Most studies used an intra-abdominal perfusion temperature greater than or equal to 41°C. However, despite using a lower intraperitoneal temperature, our clinical results compare favorably with those from these other studies. In addition, our postoperative morbidity and mortality rates are consistent with those from the other series as well.

The morbidity and mortality of this procedure, as demonstrated in this study and others in the literature, are not insignificant. As this is essentially a palliative procedure, with most patients eventually succumbing to their disease process, an important factor to consider besides OS is the effect of this procedure on the individual's quality of life (QOL). Two recent studies from our institution (Wake Forest University Health Sciences, Winston-Salem) examined short- and long-term QOL outcomes after CS and IPHC. Sixty-four patients in the short-term study⁴¹ reported decreased overall QOL after surgery compared with baseline but then returned to baseline or better within 3 to 6 months of surgery. A follow-up study⁴² of 17 patients surviving more than 3 years after CS and IPHC reported that more than 90% had minimal to no limitations of activity, with functional assessments that compared favorably to national reference values for their respective age groups. Currently, all patients undergoing CS and IPHC are entered preoperatively into an ongoing QOL study.

Probably the most difficult aspect of this study to interpret is the heterogeneity of the patient population. Patients in this phase 2 trial presented with PC from various primary tumor types, including appendiceal primaries with good prognosis and gastric cancers with poorer prognosis. Yet, despite the variety of solid malignancies represented in this study, they all presented at a similar stage of disease. As the number of patients undergoing this procedure increases, more emphasis will need to be placed on studying the outcome of patients with PC from specific primary tumor sites with uniform histologic features.

Findings from this trial and available data indicate that applying a multimodality approach to patients with PC can significantly alter the natural history of the disease, palliate symptoms, and even produce long-term survival. A recent prospective trial⁴³ randomized 104 patients with PC from colorectal adenocarcinoma to receive CS and IPHC with MMC followed by systemic fluorouracil and leucovorin calcium therapy or CS and systemic fluorouracil and leucovorin calcium therapy alone. With mean follow-up of 24 months, 2-year OS was 43% in the IPHC arm and 16% in the standard therapy arm (P = .01).⁴³

Cytoreductive surgery and IPHC with MMC is an aggressive multidisciplinary approach to a difficult oncologic situation with few meaningful therapeutic options. Although it is clearly not a treatment for all patients with PC, we believe that selected patients may benefit from improved QOL and extended OS.

Corresponding author and reprints: Perry Shen, MD, Surgical Oncology Service, Wake Forest University Baptist Medical Center, Medical Center Blvd, Winston-Salem, NC 27157 (e-mail: pshen@wfubmc.edu).

Accepted for publication September 7, 2002.

This study was presented in part at the American Society of Clinical Oncology annual meeting, San Francisco, Calif, May 12, 2001.