early postoperative intraperitoneal chemotherapy as an...

(CANCER RESEARCH 50. 5790-5794. September 15. I990|

Early Postoperative Intraperitoneal Chemotherapy as an Adjuvant Therapy toSurgery for Peritoneal Carcinomatosis from Gastrointestinal Cancer:Pharmacological StudiesPaul H. Sugarbaker,1 T. Graves,2 E. A. DeBruijn,1 W. J. Cunliffe,4 R. E. Mullins,2 W. E. Hull,5 L. Oliff,2 andP. Schlag6

Cancer Invitale, H'ashington Hospital Center, H ashington, DC 20010, and Pharmacy and Clinical Laboratories, Emory I'niversitv Hospital and School of Medicine,

Atlanta, Georgia 30322

ABSTRACT

Gastrointestinal malignancy may spread to peritoneal surfaces in theabsence of lymphatic or hematogenous mÃ©tastases. To treat peritonealcarcinomatosis, a uniformly lethal disease process, extensive cytoreduc-

tive surgery and i.p. chemotherapy were combined. Early postoperativei.p. chemotherapy Â»as instilled in the first few days after the surgicalprocedure in an attempt to treat anatomic sites that would be sealed offby postoperative adhesions. Mitomycin C was given on the first postoperative day at two doses, 10 and 12 mg/m2. 5-Fluorouracil was given on

postoperative days 2-5 at 15 and 20 mg/kg, respectively. Median areaunder the curve ratio i.p./i.v. was 117 for 5-11immuraci I and 21.6 for

mitomycin C. Elevated intraportal levels of drug were observed for i.p.5-fluorouracil but not for mitomycin C. The marked pharmacokinetic

advantage of postoperative i.p. suggests that this treatment strategyshould be considered in a clinical trial in patients at risk for progressionof peritoneal carcinomatosis.

INTRODUCTION

Surgery for gastrointestinal cancer is associated with anextremely high local recurrence rate. In a review of autopsycases at the Roswell Park Memorial Institute, 90% of patientswhose gastric cancer recurred had disease identified at theresection site (I). From the report of Tepper et al. (2), recurrence within the pancreas bed should be expected in all patientswho fail surgical removal of pancreatic cancer. The majority ofpatients whose cancer recurred had cancer in the retroperito-neum at the same site from which the primary pancreas malignancy had been removed. Cass et al. (3) found that two-thirdsof patients with colorectal malignancy have resection site recurrence. A review of the local failure rates with gastrointestinalmalignancy are summarized in Ref. 4. If these clinical data areaccurate, then gastrointestinal cancer presents a significantlydifferent natural history from many other tumors. For example,local recurrence is unusual with breast cancer or extremitysarcoma. Sugarbaker et al. (5) hypothesized that the surgicalprocedure itself contributes substantially to the natural historyof gastrointestinal cancer. The mechanism whereby a largeproportion of patients have disease recurrence confined to theresection site and peritoneal surfaces is related to traumaticdissemination of tumor emboli within the peritoneal cavity, and

Received 12/5/89; accepted 6/19/90.The costs of publication of this article were defrayed in pan by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1To whom requests for reprints should be addressed, at Cancer Institute.Washington Hospital Center. 110 Irving St.. N.W., Washington. DC 20010.

2Present address: 1364 Clifton Road. N.E., Atlanta. GA 30322.' Present address: University of Antwerp. Universiteitsplein 1. Antwerp. Bel

gium.4 Present address: Queen Elizabeth Hospital. Sheriff Hill, Gateshead. Tyne

and Wear NE9 6SX. England.5Present address: General Cancer Research Center. C'entrai Spectroscopy

Department. Neuenheimer Feld 280. D-6900 Heidelberg. Federal Republic ofGermany.

6 Present address: Chirurgische Klinik. Klinikum der UniversitÃ¤tHeidelberg.

Heidelberg. West Germany.

the implantation of these tumor emboli within the fibrinousexÃºdate that accumulates at the resection site and on abradedperitoneal surfaces. Sources for these intraabdominal tumoremboli include severed lymphatic channels, disrupted tissueinterstices at the lateral margins of tumor dissection, and tumoremboli within venous blood lost from the tumor specimen (5).By the theory of metastatic inefficiency described by Weiss (6),even a small number of tumor cells dispersed within the rawsurfaces of the abdominal cavity would be expected to result inrecurrent disease; conversely, large numbers of tumor cellsconfined to endothelial lined lymphatic or vascular channelswith an intact basement membrane would not be expected toresult in tumor implantation.

In an attempt to design treatments that would eliminate localtumor spread as a mechanism of gastrointestinal cancer recurrence, we performed pharmacological studies with i.p. chemotherapy early in the postoperative period. The rationale forEPIC7 is: (a) the resection site and abraded peritoneal surfaces

are at high risk for tumor cell implantation in the postoperativeperiod; (b) all intraabdominal surfaces are fully exposed to i.p.chemotherapy if the surgeon has been careful to separate alladherent structures and if these treatments are instituted priorto the formation of abdominal adhesions; (c) in the postoperative period the surgical techniques required for i.p. drug deliveryare extremely simple. With the abdomen open, insertion of aperitoneal access device is safe and associated with virtually nomorbidity; (d) regional chemotherapy may result in markedlyincreased local responses without compromising systemic effects. Drug delivery to the liver is markedly increased if agentsexhibit a "single pass effect" through this organ; (e) the cost to

the patient in terms of time and money is minimal since thetherapy is instituted and completed within a normal postoperative time frame. The drugs themselves are relatively inexpensive compared to the hospitalization. In order to perform phaseI and pharmacological studies, patients with peritoneal carcinomatosis were selected. These patients inevitably have recurrences with surgery alone as a treatment modality. These earlystudies build a strong pharmacological rationale for the moregeneral applications of EPIC. This management plan may helpcontrol early spread of cancer on peritoneal surfaces and thetumor spillage that may occur as a result of surgical trauma. IfEPIC can eliminate resection site and peritoneal surface recurrence, these treatments may be associated with a change in thenatural history of surgically treated gastrointestinal cancer.

MATERIALS AND METHODS

Patients. For these early studies, patients with advanced primary orrecurrent gastrointestinal cancer confined to the abdominal cavity were

7The abbreviations used are: EPIC, early postoperative i.p. chemotherapy:MMC. mitomycin C; 5-FUra, 5-fluorouracil; AUC, area under the concentrationversus time cune from zero to infinity; HPLC. high performance liquid chroma-tography: MRS. magnetic resonance spectroscopy.

5790

Research. on August 16, 2019. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

EPIC AS ADJUVANT THERAPY FOR PERITONEAL CARCINOMATOSIS

selected for protocol treatments. Diagnoses included recurrent coloncancer (7 patients), perforated appendiceal cancer (16 patients), gallbladder cancer (1 patient), and biliary tract cancer (1 patient). Allpatients had biopsy confirmed tumor on peritoneal surfaces. Patientswere excluded if they had liver mÃ©tastasesor tumor identified systemi-cally. Altogether, 26 patients underwent an extensive surgical procedurefollowed by EPIC with MMC and 5-FUra.

Treatments. After completing the cytoreductive surgical procedureand prior to closing the abdominal incision, a Tenckhoff catheter wasplaced through the abdominal wall. A purse string suture was used atthe peritoneal level in order to minimize leakage of peritoneal fluid.After the abdomen was closed abdominal lavage with a 1.5% dextrosedialysate was instituted to remove blood products and tissue debris thatresulted from surgery. No heparin or potassium was added to thedialysis fluid. One liter of the fluid was run into the abdominal cavityas rapidly as possible. The liter of fluid was immediately drained bygravity. This procedure was repeated on an hourly basis until theeffluent was clear. Abdominal lavage was repeated every 4 h until thei.p. chemotherapy was begun on the first postoperative day. The materials utilized to lavage the abdominal cavity or to deliver i.p. chemotherapy are shown in Fig. 1.

In these early studies MMC and 5-FUra were utilized i.p. For MMCa single dose of 12 mg/m2 (7 patients) or 10 mg/m2 (16 patients) wasutilized. Extensive previous experience with 5-FUra i.p. 6-10 weeksafter surgery was available (7, 8). The initial i.p. dose of 5-FUra was 20mg/kg (7 patients) with a maximum dose of 2 g. To decrease toxicity,the dose of 5-FUra was reduced to 15 mg/kg (16 patients) with amaximum dose of 1800 mg. Three patients with reduced renal orhepatic function were given MMC at one-half the calculated dose. Eachdrug was administered as a single daily instillation in 1.5% glucosedialysis fluid. Two liters of dialysate fluid were instilled i.p. initially butthis was associated with excessive abdominal discomfort. In subsequentpatients, 1 liter of dialysate was used. Drugs were allowed to dwell for23 h and removed by closed suction drains placed in dependent partsof the abdomen. When drainage ceased, another container of i.p.chemotherapy was attached and secured with a sterile clamp to theconnection site and infused by gravity as rapidly as possible into theperitoneal cavity. All abdominal drains were clamped during instillationand dwell. On the sixth postoperative day all fluid was drained fromthe peritoneal cavity and the Tenckhoff catheter was withdrawn fromits tract; closed suction drains were removed as surgically indicated.Occasionally, there was a small seepage of fluid during the chemotherapy instillations and for a few days thereafter from the abdominalincision and from the Tenckhoff catheter skin exit site. This wasthought to be an inconsequential problem except that it could causeinadvertent exposure of hospital personnel to chemotherapeutic agents.

Fig.nation

. Methods for maintaining peritoneal access free of bacterial contami-Â¡nthe early postoperative period.

Monitoring. Sampling of peritoneal fluid and plasma was performedin selected patients. A total of 17 complete pharmacokinetic studieswere performed with monitoring for 8-12 h in most patients. In threepatients, a portal venous catheter was inserted so that portal levels ofMMC or 5-FUra could be determined. These catheters were securedwith elastic suture material and were removed upon completion of thepharmacokinetic monitoring without complications (9). Blood, peritoneal fluid, or portal vein samples were drawn into heparinized tubes at0, 10, 20, 40, 60, 90, 120, 180, 240, 360, 480, and 720 min. The tubeswere centrifuged at 150 x g for 10 min and the plasma was stored at-70Â°C. Samples were assayed in batch format to minimize run to run

variation.Pharmacokinetic Analysis. R-strip, an integrated software program

by Micro Math Scientific Software (Salt Lake City, UT), was utilizedto determine pharmacokinetic parameter estimates for half-life (/./,),AUC, and elimination rate constants (A,). This software package utilizes linear regression equations to determine the pharmacokineticmodel which best fits the set of data points. Once the "best fit" model

is determined, values are calculated for f./,, Kâ€žand AUC (via thetrapezoidal rule).

5-FUra Assays by HPLC and Magnetic Resonance Spectroscop). 5-FUra was assayed by HPLC. Briefly, 5-FUra was quantitated in bodyfluids and tissue extracts using liquid-liquid extraction and reversephase chromatography. All organic solvents were HPLC grade. Type 1water was used Â¡nall Chromatographie and extraction applications. Allother reagents were the highest purity. Five hundred ^' of sample(serum, plasma, or peritoneal fluid), 25 n\ of internal standard (50 n\i5-chlorouracil in water), and 50 n\ 1.0 M potassium phosphate buffer,pH 7.0, was added to a tube and vortexed. The drug was extracted with8 ml ethyl acetate; the organic phase was recovered and evaporated todryness. Samples were reconstituted with ISO ^1 of mobile phase.Mobile phase consisted of 20 mivi acetic acid in 1% acetonitrile.Instrumentation included a 510 HPLC pump, a 7IO-B WISP autosampler, a RCM100 Radial Compression Module containing a Radial-Pak Cis-MBondapak column, a model 481 UV detector, (all fromWaters, Inc., Milford, MA), and a C-R6A integrator/recorder (Shi-madzu, Columbia, MD). The flow rate was 1.0 ml/min and the detectorwas set at 266 nm and 0.001 ultraviolet absorbance (AUFS). Lateeluding peaks were flushed from the system by injection of 300 Â¿ilofacetonitrile between each analytical run.

For one patient 19FMRS was used to determine the levels of 5-FUra

and its catabolites in plasma and peritoneal fluid samples, which werefrozen and shipped to Heidelberg for analysis at the German CancerResearch Center (10). "F nuclear magnetic resonance measurementswere performed at 470 MHz (11/7 T) using a Bruker AM-500 FTnuclear magnetic resonance spectrometer and 10-mm sample tubes.Approximately 1.5 ml of the biological fluid were measured directly,without any chemical treatment, at a controlled temperature of 4Â°C.The techniques used to obtain "F MRS spectra, their interpretation,and the quantitative analysis for 5-FUra and catabolites have beendescribed in detail elsewhere (10). Data acquisition times per sampleranged from I h (detection level for fluorine-containing metaboliteswere calculated to be 4 /Â¿M)to 8 h (detection level, about 1 ^M) (10).

MMC Assays by HPLC. The assays of peripheral blood, portalblood, peritoneal fluid, and urine for MMC were carried out as described by Tjaden et al. (11). Briefly, a fully automated liquid Chromatographie system for the bioanalysis of MMC was used. The isolationof MMC from its biological matrix (plasma, peritoneal fluid, or urine)was performed using a continuous flow system equipped with a dialysismembrane in order to remove proteins. By using a reverse phaseprecolumn, the samples are concentrated and subsequently introducedonto the reverse phase analytical column by applying column switchingtechniques.

RESULTS

Pharmacological Comparisons of Early versus Delayed i.p. 5-FUra. The studies of Dedrick et al. (12), Collins (13), andSpeyer et al. (14) all suggest a marked regional concentrationadvantage using i.p. chemotherapy administration. We wished

5791




to see if the same differences in 5-FUra concentration in theplasma and peritoneal fluid compartment existed in the earlypostoperative period when peritoneal surfaces are extensivelytraumatized by the cytoreductive surgical procedure. Studies onperitoneal fluid and plasma were performed on the secondpostoperative day and on the second day of a chemotherapycycle at 3 months postoperatively in an individual patient. Theresults for early and delayed i.p. chemotherapy are comparedin Fig. 2. These data clearly show a marked regional pharmacological advantage of i.p. chemotherapy of comparable magnitude with both early and delayed i.p. 5-FUra chemotherapy.A comparison of early and delayed i.p. 5-FUra pharmacoki-netics was repeated and similar results were obtained (Table 1).

Changes in 5-FUra Pharmacokinetics over a 5-Day TreatmentSchedule. In previous studies Sugarbaker et al. (15) showed thatthe local-regional nature of 5-FUra drug distribution was modified over a 5-day course of treatment. These authors suggestedthat changes in the peritoneal surfaces were responsible for theincreased clearance of drug from the abdominal cavity. In orderto determine the pharmacokinetics of a treatment schedule inthe early postoperative period, we sampled i.p. 5-FUra onpostoperative days 2 and 5. These results are shown in Fig. 3.The differences between 5-FUra concentration obtained in anindividual patient (peritoneal fluid versus plasma) on the secondpostoperative day as compared to the fifth postoperative day-

are less marked. Similar results were seen in the study of asecond patient given EPIC and a patient given delayed i.p.chemotherapy (Table 1). AUC differences were approximately1.5-3 times greater.

Elevated Intraportal 5-FUra in the Early Postoperative Period.In a single patient blood samples were drawn from the portalvein as well as from the peritoneal fluid and peripheral blood

001 0.077

TIME (HOURS)

Fig. 2. Pharmacological studies of early versus delayed Â¡.p.5-FUra. Peritonealfluid day 2 postoperatively (O): peritoneal fluid day 2 3 months later (â€¢);plasmaday 2 postoperative (D): plasma day 2 of cycle obtained 3 months later (â€¢).

on the second postoperative day. Fig. 4 shows the intraportallevels of 5-FUra to be nearly 10 times that present in theperipheral blood.

Comparison of HPLC Data and MRS Data. On a singlepatient, blood and peritoneal fluid samples were drawn on thesecond postoperative day and samples were equally divided.The 5-FUra concentration was determined by HPLC and alsoby |gF MRS. The data are shown in Fig. 5. The close correlation

of time courses observed by the two assay techniques suggeststhat a high level of accuracy and reproducibility can be achievedfor both methods when monitoring 5-FUra (10).

However, MRS analysis offers the advantage of measuringboth the parent compound and the catabolites of 5-FUra (10).Dihydrofluorouracil, Â«-fluoro-itf-ureidopropionate, and the finalcatabolite Â«-fluoro-/i-alanine were present in plasma at levelsof 10-20 n\i 10 min after treatment. Â«-Fluoro-ÃŸ-alaninereached its maximum of approximately 50 ^M 2 h after treatment at which point 5-FUra represented only 4% of the circulating fluorine-containing metabolites. The three catabolites of5-FUra appeared in the peritoneal fluid at detectable levels 30-60 min posttreatment and remained at nearly constant levels(ca. 50 Â¿/Mtotal) 2-6 h after treatment. The MRS method alsoallows the detection of free fluoride aniÃ³n(F~, often a contam

inant in 5-FUra solutions), which was found in both fluids atabout 5 fiM throughout the monitoring period (data not shown).

AUC and f., with Early Postoperative i.p. 5-FUra. Table 1summarizes the pharmacokinetic parameter estimates for patients receiving 5-FUra i.p. The AUC and the tv, for i.p. andsystemic 5-FUra in eight patients are shown. One sees that themedian area under the curve ratio for i.p. versus systemic 5-FUra in the early postoperative period was 117:1. The elimination tv, calculated from the data sets best described by a one-compartmental model ranges from 33 to 88 min (mean, 65.8)and for simultaneously sampled systemic 5-FUra it ranges from41.5 to 96 min (mean, 67.4). The magnitude of these differenceswill varying depending upon the dose of chemotherapy (concentration of drug in dialysate fluid), permeability of the i.p. barrierand the day chemotherapy is administered during the 4- or 5-day treatment cycle.

Assay of Four Body Compartments after Early Postoperativei.p. MMC. In the patient shown in Fig. 6, not only wereperitoneal fluid and plasma MMC concentrations determinedbut also portal blood and urine MMC levels were assayed.Marked concentration differences in peritoneal fluid as compared to systemic blood were seen. The concentration advantagefor i.p. MMC (AUC ;.P./AUCÂ¡,) was 21.6. The difference in the

Table 1 i.p. 5-FL 'ra pharmacokinetic parameter estimates

tv,(min)Patient

(cycle,day)I

(C1D2)2(C1D2)3IC1D2)(C1DS)(C2D2)(C2D5)4(C1D1)5

(CIDI)(C1D5)6

(CIDS)7(CIDI)8

(CIDI)Dose

(mg)1,1001

.6001.3001.3001.3001.300l.SOO1.8001.800620500800Plasma"58.972.868.567.973.79641.563.159.859.979.2Peritoneal

fluid032.688.171.763.279.6104.4(1)11.3(2)13.2(1)19.6(2)14.9(3)72.144.454.086.366.4AUC

(xg/ml/h)Plasma329.4146.4415.6714.3181.0255.851.14337.8953.4163.373.6Peritoneal

fluid32,971.687.954.245.324.242,597.047.891.867.542.514.682.482.799.158.419.419.I44.I8.674.717,789.5AUC,.p.:AUCi...100.1:1600.8:1109.1:159.6:1264.6:1264.0:1264.3:1245.1:161.3:1117.2:1118.1:1Correlation

coefficientPlasma0.640.950.930.930.910.800.980.990.960.990.98Peritoneal

fluid0.9970.980.990.990.960.990.990.960.980.970.970.98

Â°Ali data sets were best described as a one-compartment pharmacokinetic model except the peritoneal fluid data sets for patient 3 (C2D5) and 4 (CIDI) which

were describd as two and three compartmental. respectively, and as indicated.

5792



EPIC' AS ADJUVANT THERAPY FOR PERITONEAL ( ARC INOMATOSIS

100 769

0.77

TIME (HOURS)

Fig. 3. Changes in Â¡.p.5-Fl'ra pharmacokinetics from postoperative day 2 to

day 5. Day 2 peritoneal fluid (O); day 5 peritoneal fluid (â€¢);day 2 plasma (D):day 5 plasma (â€¢).

7688

769

i01

0.01 0077

TIME (HOURS)

Fig. 4. Elevated intraporlal 5-Fl:ra in the early postoperathe period. Dataobtained on the seeond postoperative day. Peritoneal fluid (O); portal plasma (A):plasma (D).

1000 7688

0.77

0.01 I,, , i I 0.0771234

TIME (HOURS)Fig. 5. C'omparison of 5-Fl'ra levels determined by "F MRS and HPL.C.

Data obtained from a single patient on the second postoperative day. Peritonealfluid by HI'I.C (O); peritoneal fluid by MRS (â€¢);plasma by HP1.C (D): plasmaby MRS(B).

0.001

299

0.029

00031 23456

TIME (HOURS)

Fig. 6. Early posloperathe intraperitoneal mitomyein C. Assay of four bod>compartments on the first postoperative day. Peritoneal fluid (O); urine (A):plasma (G); portal plasma (A).

i.p. M MC concentration advantage as compared to 5-FUra mayhe related to the smaller molecular weight, smaller dose, andincreased absorption of MMC across the abdominal cavitymembrane. The peripheral blood and portal blood MMC levelsare essentially the same after i.p. MMC administration. Contrast this to the differences observed for intraportal versussystemic 5-FUra (Fig. 4). The high levels of MMC in the urineshould be noted. Urine samples were obtained through a Foley

catheter and a continuous urine output of 120 ml/h was maintained over 12 h. The elevated concentration of MMC in theurine was noted to persist for at least 6 h in a second patientwho was studied.

MMC Area under the Curve and /,,. Table 2 shows the AUCand /,. estimates for five patients treated on their first postoperative day with i.p. MMC. The median AUC ratio of peritonealfluid to plasma was 21.6. The median d. for the patients whosedata indicated a one compartment decay in the peritoneal fluidwas 96.5 min and that for plasma was 290.8 min.

DISCUSSION

The major cause for limited effectiveness of cancer chemotherapy may be the development of resistant cells. One strategyfor minimizing this mechanism of treatment failure involvesdose intensive regimens administered to patients with an absolute minimum tumor burden. In our patients with peritonealcarcinomatosis, we used cytoreductive surgery to reduce thetumor burden. An attempt was made to resect cancer until nomacroscopic evidence of disease could be appreciated by theunaided eye. For patients with primary gastric, pancreatic, orcolorectal cancer, surgical resection of the primary malignancyrepresents the optimal cytoreductive procedure. In our patientswith peritoneal carcinomatosis and in future patients withprimary gastrointestinal cancer, EPIC may be the ultimate indose intensive chemotherapy. Our studies suggest that thisunique timing for i.p. chemotherapy administration presentsadditional possibilities as an adjuvant to gastrointestinal cancersurgery. High doses of regional chemotherapy over prolonged(120 h) time periods should translate into a high fraction ofcell kill and a small likelihood of drug resistance. Portions ofthe abdominal cavity sealed off by scar formation should nowhave more adequate exposure to chemotherapy. Even if EPICdoes not result in prolonged survival, it may change the naturalhistory of primary gastrointestinal cancer by eliminating resection site and peritoneal surface recurrence.

One may wish to compare the exposure of tumor in vivowhen i.p. chemotherapy was used to treat patients postopera-tively to the exposure required to kill tumor cells in in vitrotests. Park et al. (16) determined the exposure required toreliably kill colon cancer cell lines in vitro by colorimetrie assayin a 96-h time period. The concentration of MMC required was2.5 Mg/ml. The MMC drug concentration used in these clinicalstudies was between 12 and 20 /ig/ml. Because of the peritonealplasma barrier, high drug concentrations were maintained forapproximately 12 h. The studies of Jol et al. (17) suggested thatthere was a linear cell kill even at exposure times of 1 h. Theexposure time in our patients compared to Park's in vitro studies

was considerably reduced but the initial concentration of MMCwas nearly 8 times that required for complete cell kill.

For 5-FUra, peritoneal surfaces were treated for 96 h. Theinitial dose of drug was 500-1800 ng/m\. Park found reliablecell kill with 5-FUra at 200 ng/m\. For this drug, the exposureachieved with i.p. drug therapy should compare favorably withthat show ing high levels of tumor destruction by chemosensi-tivity testing.

There is a marked difference in the portal venous and systemic venous concentration of 5-FUra after i.p. drug delivery.The high level of 5-FUra in the portal blood has been relatedto the first pass effect through the liver. Pharmacokinetic studies with 5-FUra p.o. have shown a high hepatic extraction ratioand low bioavailability when the drug is slowly absorbed. Withrapid absorption, saturation of hepatic catabolic enzymes allows

5793




Table 2 Â¡.p.milomycin C pharmacokinetic parameter estimates

Patient Dose(mg)1

272 303 16.8426.25

27.6fÂ»'Plasma265.3

577.6290.8407.0159.7(min)0

AUC(jig/ml/h)Peritoneal

fluid96.5

127.284.1

117.0(1)10.7(2)88.6Plasma53,579.665.0

109,390,455.066,431,849.353,905,084.138,014,030.9Peritoneal

fluid847.227.746.0

1,624,374,610.014,358,091,580.0

1,175,736,880.01,807,090,350.0AUC,.P.:AUC,.V.15.8:1

14.8:121.6:121.8:147.5:1Correlation

coefficientPlasma0.80

0.470.700.710.87Peritoneal

fluid0.99

0.990.990.990.90

" All data sets were best described as a one compartment pharmacokinetic model except the peritoneal fluid data points for patient 4 which was best described as

a two compartment model as indicated.

for a smaller extraction ratio and thus higher systemic concentrations. 5-FUra diffusion i.p. across the peritoneal membraneis thought to be slow because of its low lipid solubility andchemical bulkiness. Therefore, out data support a high liverextraction of 5-FUra. If absorption was increased due to alteration in the peritoneal barrier one might expect to see anenzyme saturation or Michaelis-Menton picture leading to excessive systemic toxicity (18). Our data also indicate that i.p.5-FUra is preferably taken up by visceral peritoneal surfacesinto the portal blood stream (13). This is confirmed by ourMRS data which showed that 10-20 min after treatment only40% of the fluorine-containing metabolites in peripheral bloodis 5-FUra. The remaining fraction is catabolites which increasedto 96% at 2 h after treatment as the peritoneal 5-FUra concentration dropped below 1 mivi. This builds a strong rationale fori.p. 5-FUra to be used in the immediate postoperative periodas an adjuvant against tumor cell implantation within the livervasculature. Adjuvant chemotherapy i.p. may protect againstrecurrent cancer in the liver as well as implantation of tumorcells at the resection site and on peritoneal surfaces.

Some considerations concerning the nature of the peritonealplasma barrier should be apparent from this experience. Somepeople have suggested that this barrier was due to the singlecell layer mesothelium. In many of these patients studied themesothelium was almost completely removed by the cytore-ductive surgery, and yet marked concentration differences persisted between the peritoneal space and the plasma. We suggestthat this anatomic structure cannot be the peritoneal-plasmabarrier. Rather we suggest that a layer of tissue itself and thecapillary basement membrane that keeps interstitial fluids outof the vascular and lymphatic compartments may be the barrier.

One may wish to speculate regarding the alterations in local-regional pharmacokinetics of i.p. 5-FUra over a 5-day scheduleof drug administration. This may be due to saturation or fatigueof enzyme systems within the liver, or the peritoneal-plasma

barrier may be modified by the chemotherapy itself. The changing pharmacology with repeated instillations of 5-FUra chemotherapy presents a caveat for all drug monitoring studies. Theplasma 5-FUra concentrations are 10 times higher on day 5 ofi.p. 5-FUra as they are on day 2; AUC are increased 1- to 3-fold. Of course the same dose of drug was instilled i.p. everyday. The same phenomenon was seen when 5-FUra was usedas delayed postoperative treatments (15). For prolonged coursesof chemotherapy, the pharmacokinetics may change over time.Frequent and repeated monitoring over time is required todetermine safe regimens that provide maximal drug dosage.

In order to knowledgeably plan chemotherapy for gastrointestinal cancer, one must determine the actual delivery of drugsto tumor and normal tissues. Only with these types of investigations can the optimal route of drug delivery be determined.

We have begun to do this by instilling the i.p. chemotherapyprior to the surgical procedure and then sampling not only theperitoneal fluid and plasma but also the tissues within theabdominal cavity which are bathed by the chemotherapy. It isquite possible that optimal treatment of gross tumor on peritoneal surfaces should used both i.p. and i.v. drugs.

REFERENCES

1. Douglass. H. O.. Jr. Gastric cancer: overview of current therapies. Semin.Oncol. 12 (Suppl. 4): 57-62. 1985.

2. Tepper, J., Nardi, G., and Suit, H. Carcinoma of the pancreas: review of M.G. H. experience from 1963-1973. Analysis of surgical failure and implications for radiation therapy. Cancer (Phila.), 37: 1519-1524. 1976.

3. Cass, A. W., Million, R. R.. and Pfaff. W. W. Patterns of recurrence followingsurgery alone for adenocarcinoma of the colon and rectum. Cancer (Phila.),37: 2861-2865. 1976.

4. Cunliffe, W. J.. and Sugarbaker, P. H. Gastrointestinal malignancy: rationalefor adjuvant therapy using early postoperative intraperitoneal chemotherapy(EPIC). Br. J. Surg.. 76: 1082-1090. 1989.

5. Sugarbaker. P. H.. Cunliffe. W., Belliveau. J.. et al. Rationale for earlypostoperative intraperitoneal chemotherapy as a surgical adjuvant for gastrointestinal malignancy. Regul. Cancer Treat.. 7:66-79, 1989.

6. Weiss. L. Metastatic inefficiency: causes and consequences. Cancer Rev., 3:1-24. 1986.

7. Sugarbaker. P. H., Gianola, F. J., Speyer. J. L., Wesley, R., Barofsky, I., andMeyers, C. E. Prospective randomized trial of intravenous versus intraperitoneal 5-fluorouracil in patients with advanced primary colon or rectal cancer.Surgery, 98: 414-421, 1985.

8. Speyer, J. L., Collins, J. M., Dedrick, R. L.. Brennan, M. F., Lander, H.,DeVita. V. T., and Myers, C. E. Phase 1 and pharmacologie studies of 5-fluorouracil administered intraperitoneally. Cancer Res., 40: 567-572, 1980.

9. Sugarbaker, P. H. Prevention of hemorrhage by elastic suture material usedfor securing intravenous catheters within major branches of the portal vein.J. Surg. Oncol., 39: 269-270. 1988.

10. Hull, W. E., Post, R. E.. Hermann, R., Britsch, B., and Rung, W. Metabolitesof 5-fluorouracil in plasma and urine as monitored by 19F nuclear magnetic

resonance spectroscopy, for patients receiving chemotherapy with or withoutmethotrexate pretreatment. Cancer Res.. 48: 1680-1688. 1988.

11. Tjaden, U. R., DeBruijn, E. A.. Van der Hoeven. R. A. M., Jol, C., Van derGreef, J.. and Lingeman. H. Automated analysis of mitomycin-C in bodyfluids by high-performance liquid chromatography with one-line sample pretreatment. J. Chromatogr.. 420: 53-62, 1987.

12. Dedrick, R. L., Meyers, C. E.. Bungay. P. M., et al. Pharmacokinetic rationalefor peritoneal drug delivery in the treatment of ovarian cancer. Cancer Treat.Rep., 62: 1-11, 1978.

13. Collins. J. M. Pharmacologie rationale for regional drug deliver. J. Clin.Oncol.. 2:498-504. 1984.

14. Speyer, J. L., Sugarbaker, P. H., Collins, J. M.. Dedrick, R. L., Klecker, R.W., Jr., and Meyers, C. E. Portal levels and hepatic clearance of 5-fluorouracilafter intraperitoneal administration in humans. Cancer Res., 41: 1916-1922,

1981.15. Sugarbaker, P. H., Klecker, R. W., Gianola. F. J., and Speyer, J. L. Prolonged

treatment schedules with intraperitoneal 5-fluorouracil diminish the local-regional nature of drug distribution. Am. J. Clin. Oncol., 9: 1-7, 1986.

16. Park. J. G.. Kramer. B. S.. Steinberg, S. M., Carmichael. J., Collins, J. M..Minna. J. D., and Gazdar, A. F. Chemosensitivity testing of human eolorectalcarcinoma cell lines usually a tetrazolium-based colorimetrie assay. CancerRes., 47: 5875-5879, 1987.

17. Jol, C, Kuppem, P.. Leeflang, P. A., van Oosterom, A. T., Tjaden, U. R.,and de Bruijn, E. A. In vitro human ovarian cancer. In: T. Taguchi and O.Andryse (eds.). New Trends in Cancer Chemotherapy with Mitomycin C. pp.181-191, Amsterdam: Excerpta Medica, 1987.

18. El Sayed, Y. M., and Sadee, W. The fluoropyrimidines. In: M. M. Ames. G.Powis, and J. S. Kivach (eds.), Pharmacokineties of Anticancer Agents inHumans, pp. 209-227. Amsterdam: Elsevier Science Publishers. BV. 1983.

5794



1990;50:5790-5794. Cancer Res Paul H. Sugarbaker, T. Graves, E. A. DeBruijn, et al. from Gastrointestinal Cancer: Pharmacological StudiesAdjuvant Therapy to Surgery for Peritoneal Carcinomatosis Early Postoperative Intraperitoneal Chemotherapy as an

Updated version

http://cancerres.aacrjournals.org/content/50/18/5790

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/50/18/5790To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



early postoperative intraperitoneal chemotherapy as an...

Documents