This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Ensminger, W. Articles by Maybaum, J. Search for Related Content PubMed PubMed Citation Articles by Ensminger, W. Articles by Maybaum, J.

Effects of Dexamethasone or Celecoxib on Biliary Toxicity after Hepatic Arterial Infusion of 5-Fluorodeoxyuridine in a Canine Model

¹Departments of Pharmacology, ²General Surgery, ³Upjohn Center, ⁴Unit for Laboratory Animal Medicine, ⁵Internal Medicine, and ⁶Radiology, University of Michigan Medical School, Ann Arbor, Michigan

	ABSTRACT

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Previous work has shown that in humans the dose-limiting toxicity for fluorodeoxyuridine [2-fluoro-5'-deoxyuridine (FdUrd)] when administered by hepatic arterial infusion is biliary sclerosis. The current study was undertaken to attempt to modify this toxicity in a canine model that has been demonstrated to closely mimic the clinical situation. Unlike previous studies using this model, in which animals were sacrificed after extensive fibrosis had already occurred, the current experiments were designed so that observations of pathology were made at an earlier time, when the initial inflammatory injury underlying the fibrotic process was still taking place. Implantable pumps were used to deliver FdUrd into the hepatic artery of animals at a rate of 0.3 mg/kg/day in the presence or absence of 10 mg/week dexamethasone or 100 mg/day of celecoxib for 35 days, at which time the animals were beginning to show signs of toxicity. After evaluation for radiological evidence of biliary obstruction, the animals were sacrificed and portions of their livers were processed for examination of microscopic pathology and 2-bromo-5'deoxyuridine labeling index. Dexamethasone treatment protected the animals from biliary sclerosis determined radiologically, further validating this model as being representative of the response in humans. Similarly the Cox-2 inhibitor, celecoxib, appeared to provide protection against radiological changes of biliary stricture, although possibly to a lesser degree than the resultant from dexamethasone. In addition, FdUrd treatment caused elevation of the DNA 2-bromo-5'deoxyuridine labeling index above control levels in biliary epithelial cells. Dexamethasone and celecoxib each significantly attenuated the FdUrd-induced elevation of DNA labeling index in biliary epithelium. These findings demonstrate the usefulness of this canine model for studying the mechanisms of drug-induced biliary sclerosis and reinforce the hypothesis that blocking inflammation may retard the progression of injury that eventually leads to fibrosis. This study suggests that clinical testing of celecoxib as a preventive for hepatic arterial-FdUrd induced biliary damage could prove valuable.

	INTRODUCTION

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Colorectal cancer frequently metastasizes to the liver. Surgical resection can cure approximately one-third of patients in whom three or fewer metastases are found in the liver. Unfortunately, the majority of patients have more than three liver metastases and, thus, are treated with chemotherapy. For those patients with colorectal cancer confined to the liver, regional administration of chemotherapy by direct hepatic arterial (HA) infusion has been used to deliver higher drug doses to the tumor than are possible with i.v. therapy to achieve greater antitumor effect (1) . Application of HA infusion of fluorodeoxyuridine [2-fluoro-5'-deoxyuridine (FdUrd)] became widely practiced in the mid-1980s with the availability of a totally implanted pump and catheter for that purpose (2 , 3) . The implanted system allowed chronic HA FdUrd infusion as a treatment for colorectal metastases and was associated with a response rate that was approximately twice that observed with i.v. fluorouracil (4 , 5) .

In initial studies with the implanted system in the early 1980s, we found that serum liver enzymes became elevated and jaundice developed 4–6 weeks into a constant infusion of HA FdUrd. To moderate this toxicity, we promulgated a cycle of 2 weeks of drug infusion alternating with 2 weeks of a saline infusion as a means to lower the incidence and severity of the liver enzyme elevation and jaundice (2) . Subsequently, Kemeny et al. (6) described biliary strictures with HA FdUrd in a few patients. In another study, Hohn et al. (7) found that all of 35 patients treated with HA FdUrd developed significant increases in serum alkaline phosphatase levels and, at least 7 had sclerosis of the intrahepatic and/or extrahepatic biliary systems. Liver biopsies from toxic patients demonstrated cholestasis and pericholangitis as well as hepatitis and hepatocyte necrosis (8 , 9) . Hepatocyte necrosis, as measured by serum levels of hepatic enzymes, generally was found to resolve. However, biliary epithelial damage resulted in fibrosis and the development of unrelenting jaundice because of permanent biliary obstruction in some of the patients.

In later investigations, it was demonstrated that both the tumor response rate and the incidence of biliary sclerosis was increased when leucovorin was given with increasing dosages of HA FdUrd (10 , 11) . Subsequently, the Sloan-Kettering group added dexamethasone in an attempt to ameliorate pericholangitis and, thereby, to decrease biliary toxicity, allowing for more HA FdUrd/leucovorin to be administered (12 , 13) . When dexamethasone was coadministered, biliary toxicity was diminished, more intense chemotherapy was tolerated, and the response rate improved. It appears that increasing the dose of HA FdUrd can improve the response rate if biliary toxicity can be controlled. The demonstration that adjuvant HA FdUrd administered after hepatic resection of colorectal liver metastases significantly improves survival makes an understanding of biliary toxicity of increasing importance (14, 15, 16) . A serious concern is that HA FdUrd adjuvant therapy treatment-induced jaundice, biliary obstruction, resultant infections, and cirrhosis could negatively impact the quality of life and survival of otherwise cured patients.

In an attempt to further understand and prevent biliary toxicity, we developed a dog model of biliary sclerosis associated with chronic HA FdUrd infusion (17 , 18) . In this dog model, serum liver enzymes progressively increase starting 2–3 weeks into HA FdUrd infusion, followed by hyperbilirubinemia beginning 4–6 weeks into infusion. Cholangiograms reveal focal strictures involving the central bile ducts and diffuse attenuation of the intrahepatic ducts. Hence, the time course and type of abnormalities seen in this model mimic the toxicities seen in the clinical setting. To further dissect the process and to work toward prevention of the biliary toxicity associated with HA FdUrd infusion, we have investigated the effect of adding dexamethasone or the anti-inflammatory Cox-2 inhibitor, celecoxib, on the generation of biliary damage in this dog model. Through pathological and radiological examination at an earlier time than was studied previously, we find that the principal locus of injury by HA FdUrd is biliary epithelium and not hepatocytes and that both dexamethasone and celecoxib have salutory effects in ameliorating biliary sclerosis.

	MATERIALS AND METHODS

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Animals and Surgical Procedures.
Twelve purpose-bred dogs (20 kg body weight) from Covance (Kalamazoo, MI) were housed in the Unit for Laboratory Animal quarters for at least 1 week before surgery. Animal husbandry was provided by the staff of the Unit for Laboratory Medicine under the guidance of supervisors who are certified as Animal Technologists by the American Association of Laboratory Animal Service. This research project was reviewed and approved by an appropriate protocol committee of Unit for Laboratory Medicine following standards defined in an assurance statement filed with the Office of Protection from Research Risk at the NIH. Blood samples were obtained to determine baseline liver enzyme levels. Infusaid Model 400 pumps (Arrow Corp., Walpole, MA) were implanted as described previously (19) . Through a midline laparatomy, the common hepatic artery was isolated from its proximal portion to distal to the gastroduodenal artery. The gastroduodenal artery and other nonhepatic arterial branches were ligated. The abdomen was temporarily closed with towel clips and the dog was placed on its left side. A transverse s.c. pocket was dissected on the lower right anterolateral thorax. A primed Infusaid infusion pump filled with heparinized saline (10,000 units of heparin in 50 ml of 0.9% bacteriostatic saline) was implanted into the pocket with the catheter tunneled into the abdomen. A venous access-type port was implanted in another s.c. pocket in the right flank with the catheter tunneled into the abdomen. The pockets were closed with two layers of absorbable suture, and the skin was closed with skin staples. The dog was repositioned to a supine position, and the abdomen was reaccessed. The gall bladder and cystic duct were dissected. The catheter of the venous-type access port was inserted prograde into the cystic duct through a ductotomy and was fixed with surrounding ligatures with its tip at the junction of the cystic and common bile duct (17) . The gallbladder was removed. A silastic small-diameter tip was attached to the catheter from the pump and was inserted into the common hepatic artery through a proximal arteriotomy and was fixed with fine sutures to the arterial wall. The incision was closed in three layers, and the dog was allowed to recover under a heating lamp.

Two weeks after surgery, a radionuclide HA perfusion scan was obtained to confirm catheter placement. A baseline cholangiogram was also obtained before the start of the HA FdUrd infusion (see below). Pumps were filled with heparinized (10,000 units), normal saline (50 ml) for at least one 14-day cycle to establish an in vivo flow rate. Immediately before, and twice weekly after the initiation of HA FdUrd, blood samples were obtained to monitor serum levels of total bilirubin, aspartate aminotransferase and alkaline phosphatase. There were a total of four evaluable animals for each treatment group.

HA Perfusion Scanning.
Before initiation of HA FdUrd treatment, dogs received an i.v. injection of 4 mCi technetium-99m-sulfur colloid after they were anesthetized with a short acting barbiturate. An image over the abdomen was taken with a large field-of-view camera to obtain a standard liver-spleen scan. The Sideport on the hepatic artery infusion pump was then injected with 5 mCi of technetium-99m-labeled macroaggregated albumin over 1 min and similarly imaged. The macroaggregated albumin HA scan was compared (using subtraction) with the liver-spleen scan to confirm that at least 70% of the liver was being perfused.

Cholangiograms.
Pretreatment baseline cholangiograms were obtained. Cholangiograms were repeated at 35 days into drug infusion, or earlier if drug toxicity developed. One FdUrd-only dog developed toxicity rapidly on day 29 and had to be euthanized before a final cholangiogram could be obtained. To obtain the cholangiograms, dogs were anesthetized with a short-acting barbiturate. Contrast was injected through the cystic duct access port to a volume that displayed the entire biliary tree under fluoroscopic control. Spot films were obtained. The films were read, with treatments blinded, by a board-certified radiologist (D. W.).

Drug Treatment.
FdUrd was administered as a constant HA infusion via the implanted pump to all animals at a delivered dose of 0.3 mg/kg/day. Dexamethasone was added to the infusate in the pumps to deliver a dose of 10 mg/week to four of the animals. Celecoxib at 100 mg orally once per day was given to another two animals and twice daily to another two animals. A proton pump inhibitor (omeprazole, 20 mg/day) was started p.o. on Day 0 and continued on a daily basis until euthanasia in all animals. One dog was given the standard dose of 20 mg daily, but developed blood in the stool. In this dog, the omeprazole dose was then increased to 20 mg twice daily, after which bleeding stopped. All of the dogs were scheduled to be sacrificed at day 36 of HA FdUrd infusion. Two dogs were sacrificed early because of drug toxicity, one on day 29, the other on day 35. An autopsy was performed on every dog. Liver samples were collected for analysis of pathology and immunohistochemistry.

Pathological Changes Induced by Drug Treatment.
Formalin-fixed liver samples were processed into standard H&E slides. Slides were read with treatments blinded by a board-certified veterinary pathologist (E. W.). Grading of portal fibrosis and bile duct injury followed the grading previously described clinically (9) .

Determination of Labeling Index in Hepatocytes and Biliary Epithelial Cells.
All of the dogs received 500 mg of 2-bromo-5'deoxyuridine (BrdUrd) 1–4 h before euthanasia. Tissue specimens were prepared as described previously (19) . Briefly, samples were fixed with 70% ethanol, embedded in paraffin, and sliced into 3–5-µm sections. After tissue sections were deparaffinized, they were treated with 0.1 M HCl and 0.7% Triton X-100, followed by boiling in water for 10 min. The tissue sections were plunged into cold distilled water for 5 min and then were treated with 0.3% hydrogen peroxide in methanol for 30 min. After being blocked with 1% BSA, slides were incubated with mouse anti-BrdUrd antibody (1:200; Sigma-Aldrich), followed by incubation with antimouse IgG conjugated with hydrogen peroxidase (1:500; Sigma). The chromogenic reaction was developed by exposing slides to diaminobenzidine and hydrogen peroxide. After rinsing with distilled water, tissue sections were counterstained with hematoxylin.

The BrdUrd labeling index was determined for cubical cells in bile ducts and for hepatocytes. The number of labeled and unlabeled cells in each bile duct and the number of labeled and unlabeled hepatocytes within a x400 field centered about the portal triad were recorded. The labeling index was calculated as the number of positive cells divided by the total number of positive and negative cells.

	RESULTS

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
FdUrd-Induced Changes in Circulating Markers for Hepatic and Biliary Injury.
As expected from previous studies (17 , 18) , elevated levels of aspartate aminotransferase and alkaline phosphatase were detected in some dogs, beginning about 30 days into the infusions (Fig. 1) . Bilirubin was elevated at the time of sacrifice in two of the animals in the FdUrd-only group and two in the celecoxib group. Addition of dexamethasone or celecoxib to the treatment regimen did not change the other biochemical profiles uniformly. These results indicate that the time course of HA FdUrd-induced biochemically detectable injury in the dogs in the present study follows the time course seen in earlier studies.

View larger version (29K): [in this window] [in a new window] [Download PPT slide]   Fig. 1. Serum levels of aspartate aminotransferase (SGOT), alkaline phosphatase (Alk. Phos.), and bilirubin as a function of time during infusion of 2-fluoro-5'-deoxyuridine (FdUrd) alone versus FdUrd plus dexamethasone (Dex.) or celecoxib (Cel.). Each line represents an individual animal. FdUrd: , 1F; , 2F; , 3F; , 4F. FdUrd + Dex.: , 1D; , 2D; , 3D; , 4D. FdUrd + Cel.: , 1C; , 2C; , 3C; , 4C.

View larger version (52K): [in this window] [in a new window] [Download PPT slide]   Fig. 2. Radiological evidence of the protective effect of dexamethasone in the canine model for hepatic arterial 2-fluoro-5'-deoxyuridine (HA FdUrd)-induced biliary sclerosis. Representative cholangiograms are shown from dogs receiving FdUrd only (A) or FdUrd plus dexamethasone (B). R. Hepatic Bile Duct, right hepatic bile duct.

	DISCUSSION

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The present investigation involved sacrificing dogs at an earlier time in the infusion course than in our previous studies, a time when the serum aspartate aminotransferase is rising (indicating hepatocyte damage) and the serum alkaline phosphatase is rising (indicating damage to biliary epithelium), yet jaundice is not prominent, reflecting clinically less significant biliary obstruction.

As Table 3 suggests, there is an attempt at regeneration at the same time that damage is occurring. The BrdUrd labeling indicates recruitment of cells into S phase. It is noteworthy that with HA FdUrd alone, the labeling index for biliary epithelium is greater than 10 times the hepatocyte labeling index. With continued FdUrd exposure, it is possible that with passing time, more and more biliary epithelial cells are recruited into S phase, in which they become extremely sensitive to the continued cytotoxic effects of the S-phase-specific agent FdUrd. Separation of cytotoxicity (reflected in enzyme elevations) from regeneration (reflected in BrdUrd labeling indices) is relevant to the clinical situation in which hepatic enzyme levels fall rapidly after HA FdUrd is stopped. Yet, clinically, reinstitution of HA FdUrd immediately after serum enzymes have returned to baseline, causes a rapid reelevation of hepatic enzymes. In this situation, it is likely that the enlarged regenerating and cycling cell fraction is specifically and extremely sensitive to HA FdUrd. These results in the dog model substantiate our clinical practice of waiting for 1 month after hepatic enzymes return to baseline before reinstituting HA FdUrd treatment. The much larger fraction of biliary epithelial cell compartment that is in S phase as compared with the hepatocyte compartment is likely to account for the greater and more significant toxicity arising in the biliary tree.

The lower level of hepatocyte BrdUrd labeling is consistent with the pathology results shown in Table 1 . BrdUrd labeling might be more sensitive than standard pathological examination at earlier time points.

As shown in Table 1 , dexamethasone and celecoxib largely prevent cholangiographic evidence of bile duct injury. Evidence that celecoxib seems to have an effect on reducing cholangiographic abnormalities induced by FdUrd (Table 1) and on reducing BrdUrd labeling in biliary epithelium (Table 3) is discordant with the pathological results (Table 2) . It is possible that pathological examination at yet earlier time points would show a more consistent result.

Consistent with the data from treated patients (12 , 13) , HA dexamethasone specifically protects the biliary epithelium in treated dogs as shown in Tables 2 and 3 . Dexamethasone produced a marked decrease in portal fibrosis and bile duct injury (Table 2) . A similar selectivity is found when examining the BrdUrd labeling index, in which dexamethasone had a significant effect in lowering the labeling index only for biliary epithelium (Table 3) . Thus, the clinical use of dexamethasone to decrease biliary toxicity is substantiated by the results in this model. However, a break or rest period off HA FdUrd as part of treatment is still essential, because the labeling index shows that DNA synthesis, albeit at a lower level, continues in biliary epithelium despite dexamethasone.

Although this study has demonstrated the usefulness of dexamethasone in preventing portal fibrosis, it appears that the time point chosen was later than would be ideal to examine earlier stages of inflammation. It is reasonable to assume that dexamethasone and celecoxib work by decreasing or delaying inflammation. Pathological evidence that celecoxib is measurably effective in preventing inflammation was not seen in this study, although cholangiographic evidence seem to show that damage was avoided. Celecoxib is likely to be a less effective anti-inflammatory agent than dexamethasone. On the other hand, celecoxib generally has a preferable toxicity profile, compared with dexamethasone, and has potential antitumor efforts on its own (21) . Pathological examination at an earlier time point might more clearly show a beneficial pathological effect of celecoxib.

The evidence of celecoxib protection against bile duct injury, shown radiologically and through effects on BrdUrd labeling in this dog model, lends support to the initiation of a clinical trial determining whether celecoxib decreases the incidence of biliary sclerosis in patients receiving HA FdUrd.

	FOOTNOTES

 
Grant support: NIH-CA77391.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Requests for reprints: William D. Ensminger, University of Michigan, 3701 Upjohn Center, 1310 East Catherine, Ann Arbor, MI 48109-0504. Phone: (734) 764-5468; Fax: (734) 763-3438; E-mail: ensminge{at}umich.edu

Received 8/21/03. Revised 10/ 2/03. Accepted 10/29/03.

	REFERENCES

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 

1. Ensminger W. D. Intraarterial chemotherapy for the treatment of hepatic metastases DeVita V. T. Hellman S. Rosenberg S. A. eds. . Cancer: Principles and Practice of Oncology, Update Series, Vol. 1: 1-11, Lippincott Philadelphia 1987.
2. Ensminger W., Niederhuber J., Dakhil S., Thrall J., Wheeler R. Totally implanted drug delivery system for hepatic arterial chemotherapy. Cancer Treat. Rep., 65: 393-400, 1981.[Medline]
3. Niederhuber J. E., Ensminger W., Gyves J., Thrall J., Walker S., Cozzi E. Regional chemotherapy of colorectal cancer metastatic to the liver. Cancer (Phila.), 53: 1336-1343, 1984.
4. Meta-Analysis Group in Cancer. Reappraisal of HA infusion in the treatment of nonresectable liver metastases from colorectal cancer. Meta-Analysis Group in Cancer. J. Natl. Cancer Inst. (Bethesda), 88: 252-258, 1996.[Abstract/Free Full Text]
5. Ensminger W. D. Intraarterial therapy Perry M. C. eds. . The Chemotherapy Source Book, Ed. 2 253-269, Williams & Wilkins Baltimore 1996.
6. Kemeny M. M., Battifora H., Blayney D. W., Cecchi G., Goldberg D. A., Leong L. A., Margolin K. A., Terz J. J. Sclerosing cholangitis after continuous hepatic artery infusion of FUDR. Ann. Surg., 202: 176-181, 1985.[Medline]
7. Hohn D., Melnick J., Stagg R., Altman D., Friedman M., Ignoffo R., Ferrell L., Lewis B. Biliary sclerosis in patients receiving HA infusions of floxuridine. J. Clin. Oncol., 3: 98-102, 1985.[Abstract]
8. Shea W. J., Jr., Demas B. E., Goldberg H. I., Hohn D. C., Ferrell L. D., Kerlan R. K. Sclerosing cholangitis associated with HA FUDR chemotherapy: radiographic-histologic correlation. Am. J. Roentgenol., 146: 717-721, 1986.[Abstract/Free Full Text]
9. Doria M. I., Jr., Shepard K. V., Levin B., Riddell R. H. Liver pathology following HA infusion chemotherapy. Hepatic toxicity with FUDR. Cancer (Phila.), 58: 855-861, 1986.
10. Kemeny N., Seiter K., Conti J. A., Cohen A., Bertino J. R., Sigurdson E. R., Botet J., Chapman D., Mazumdar M., Budd A. J. HA floxuridine and leucovorin for unresectable liver metastases from colorectal carcinoma. New dose schedules and survival update. Cancer (Phila.), 73: 1134-1142, 1994.
11. Kemeny N., Cohen A., Bertino J. R., Sigurdson E. R., Botet J., Oderman P. Continuous intrahepatic infusion of floxuridine and leucovorin through an implantable pump for the treatment of hepatic metastases from colorectal carcinoma. Cancer (Phila.), 65: 2446-2450, 1990.
12. Kemeny N., Seiter K., Niedzwiecki D., Chapman D., Sigurdson E., Cohen A., Botet J., Oderman P., Murray P. A randomized trial of intrahepatic infusion of fluorodeoxyuridine with dexamethasone versus fluorodeoxyuridine alone in the treatment of metastatic colorectal cancer. Cancer (Phila.), 69: 327-334, 1992.
13. Kemeny N., Conti J. A., Cohen A., Campana P., Huang Y., Shi W. J., Botet J., Pulliam S., Bertino J. R. Phase II study of hepatic arterial floxuridine, leucovorin, and dexamethasone for unresectable liver metastases from colorectal carcinoma. J. Clin. Oncol., 12: 2288-2295, 1994.[Abstract/Free Full Text]
14. Kemeny N., Conti J. A., Sigurdson E., Cohen A., Seiter K., Lincer R., Niedzwiecki D., Botet J., Chapman D., Costa P., Budd A. A pilot study of hepatic artery floxuridine combined with systemic 5- fluorouracil and leucovorin. A potential adjuvant program after resection of colorectal hepatic metastases. Cancer (Phila.), 71: 1964-1971, 1993.
15. Kemeny M. M., Adak S., Lipsitz S., Gray B., MacDonald J., Benson A. B., III Results of the intergroup [Eastern Cooperative Oncology Group (ECOG) and Southwest Oncology Group (SWOG)] prospective randomized study of surgery alone versus continuous hepatic artery infusion of FUDR and continuous systemic infusion of 5FU after hepatic resection for colorectal liver metastases. Proc. Am. Soc. Clin. Oncol., 18: 264a 1999.
16. Kemeny N., Huang Y., Cohen A. M., Shi W., Conti J. A., Brennan M. F., Bertino J. R., Turnbull A. D., Sullivan D., Stockman J., Blumgart L. H., Fong Y. Hepatic arterial infusion of chemotherapy after resection of hepatic metastases from colorectal cancer. N. Engl. J. Med., 341: 2039-2048, 1999.[Abstract/Free Full Text]
17. Zeng Z. S., Zhang Z. F., Sigurdson E. R. Toxicity study of hepatic artery infusion chemotherapy with massive dose FUDR rats. Surg. Oncol., 4: 147-155, 1995.[Medline]
18. Andrews J. C., Knol J., Wollner I., Knutsen C., Smith P., Prieskorn D., Ensminger W. Floxuridine-associated sclerosing cholangitis. A dog model. Investig. Radiol., 24: 47-51, 1989.[CrossRef][Medline]
19. Andrews J. C., Knutsen C., Terio P., Prieskorn D., Ensminger W. D. Hepatobiliary toxicity of 5-fluoro-2'-deoxyuridine. Intra-arterial versus portal venous routes of infusion. Investig. Radiol., 26: 461-464, 1991.[CrossRef][Medline]
20. Wollner I. S., Knutsen C. A., Ullrich K. A., Niederhuber J. E., Crudup J. W., Juni J. E., Warber S. L., Gyves J., Stetson P. L., Ensminger W. D. A dog model using an implanted system for protracted hepatic arterial chemotherapy. J. Surg. Res., 41: 510-517, 1986.[CrossRef][Medline]
21. Thun M. J., Henley S. J., Patrono C. Nonsteroidal anti-inflammatory drugs as anticancer agents: mechanistic, pharmacologic, and clinical issues. J. Nat. Canc. Inst. (Bethesda), 94: 252-266, 2002.

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Ensminger, W. Articles by Maybaum, J. Search for Related Content PubMed PubMed Citation Articles by Ensminger, W. Articles by Maybaum, J.