- Email: [email protected]
Intra-arterial chemotherapy for unresectable pancreatic cancer
Annals of Oncology 11: 569-573, 2000. © 2000 Kluwer Academic Publishers. Primed in the Netherlands.
Original article Intra-arterial chemotherapy for unresectable pancreatic cancer M. Cantore,1 P. Pederzoli,2 G. Cornalba,3 G. Fiorentini,4 S. Guadagni,5 L. Miserocchi,1 A. Frassoldati,6 C. Ceravolo,7 F. Smerieri1 & J. H. Muchmore8 ' Departments of Oncology and Radiology 'C. Poma' Hospital, Manlova; 2 Department of Endocrine Surgery, University of Verona, 1Department of Radiology'Sacco' Hospital, University of Milan: *Department ofOncology Civil Hospital, Ravenna, s Department of Surgery, University of LAquila; 6Department of Oncology, University of Modena; 7Department of Surgery Civil Hospital, Crema, Italy; 8Department of Surgery, Tulane University School of Medicine, USA
Summary Background: A phase II trial of a new intra-arterial chemotherapy regimen for unresectable pancreatic cancer (UPC). Patients and methods: Ninety-six patients with UPC were treated with intra-arterial chemotherapy at three-weekly intervals. The schedule used was FLEC: 5-fluorouracil 1000 mg/m 2 , folinic acid 100 mg/m 2 , carboplatin 300 mg/m 2 ; epirubicin 60 mg/m 2 . Results: The overall response rates by CT-scan evaluation were: 15% partial response (PR), 44% stable disease (SD), 17% progressive disease (PD). The overall median survival was 9.9 months, and 10.6 and 6.8 for UICC stage III and IV, respectively. Pain reduction occurred in 42% of patients. A weight gain > 7% from baseline occurred in 8% of patients. A total of
Introduction
The incidence of pancreatic adenocarcinoma has increased over the last two decades in most Western countries [1]. Chemotherapy alone or in combination with radiotherapy has not significantly improved the survival of patients with UICC stage III-IV disease [2-4]. Thus, in addition to the discovery of new active drugs, current research involves an integrated strategy of new administration routes delivering high-dose chemotherapy to the tumor-bearing region drug. Pancreatic cancer is notably resistant to systemic chemotherapy, with most tumors expressing increased levels of the multidrug resistant (MDR) gene [5]. Currently, two other non-Pglycoprotein mechanisms of multidrug resistance are found in pancreatic cancers: the multidrug resistanceassociated protein (MRP) and the lung cancer resistanceassociated protein (LRP) [6]. Intra-arterial chemotherapy as a means of increasing drug delivery for the treatment of advanced, pancreatic cancer has been previously investigated. Preliminary results of non-randomized, small series studies demonstrate that pancreatic cancer is sensitive to regional chemotherapy [7-11]. This paper reports the results of a phase II study of a new and
341 courses of FLEC were administered. Grade 3-4 hematological toxicity was seen in 25% of patients; ematemesis in 4%; grade 3 gastrointestinal toxicity in 3%; and grade 3 alopecia in 16%. One sudden death, a pre-infarction angina, and a transitory ischemic attack were observed. The only complication related to the angiographic procedure was an intimal dissection of the iliac artery. Conclusions: The intra-arterial FLEC regimen was well tolerated and active. It requires only one day of hospitalization. Efficacy could only be assessed in a randomized study against a gemcitabine containing regimen.
Key words: chemotherapy, intra-arterial, liver metastasis, unresectable pancreatic cancer
simple locoregional approach in the treatment of unresectable pancreatic cancer (UPC).
Patients and methods Patients with biopsy-proven adenocarcinoma of the pancreas that was not suitable for curative resection were eligible for enrolment in the study. They were required to have a performance status less than 3 (Southwest Oncology Group (SWOG) rating), a pre-treatment white blood cell count > 3000/ul, platelet count > 120,000/nl and hemoglobin level >9.5 gm/dl. Patients' total bilirubin and serum creatinine levels were required to be <1.5 times the institutional upper limit of normal. Staging included abdominal sonography, total abdomen CT scan, chest X-ray. Weight, pain intensity and analgesic consumption, performance status and CA 19-9 were evaluated at study entry and after each cycle of regional therapy. All patients gave their informed consent according to our institutional guidelines, and the study received ethical approval. Three cycles of regional chemotherapy were administered at threeweekly intervals through an angiographic catheter (Simmons 2; 5 Fr) introduced via the femoral artery into the celiac axis. Each drug was diluted in 100 ml of normal saline and then infused over a period of 10 minutes, one after the other in the following order: folinic acid (FA) 100 mg/m 2 : 5-fluorouracil (5-FU) 1000 mg/m 2 ; carboplatin (CP) 300 mg/m 2 ; epirubicin (EPI) 60 mg/m 2 . The drugs were infused according to tumor location: i.e., the
570 gastroduodenal artery for tumors of the pancreatic head; the splenic artery for tumors of the pancreatic body-tail; and when liver metastases were present, half of the total dose was infused in the hepatic artery and half according to primary location. When it was not possible to introduce the tip of catheter into either the gastroduodenal or splenic arteries, the drugs were infused into the celiac axis. In addition, an antiemetic (granisetron 8 mg) and an H2-receptor antagonist (famotidine 40 mg) were given intravenously. Hematological growth factor (filgrastim) was given at a dose of 5 ug/kg/d from day 10 after chemotherapy, for six consecutive days. Patients were discharged the day after therapy. Complete blood cell counts were performed before each cycle and twice a week. The response was evaluated after three cycles. In cases of clinical response, independent of CT scan response, a further three cycles were administered. All patients were evaluated for response on an intention-to-treat basis. Modification of pain was objectively evaluated using a five-point rating scale [12]: grade 1 = no therapy required, grade 2 = only nonsteroidal anti-inflammatory drugs (NSAIDs); grade 3 = combination of NSAIDs, antidepressants, and anticonvulsants; grade 4 = morphine derivatives; grade 5 = combination of morphine derivatives, NSAIDs, antidepressants, and anticonvulsants. CR was considered when grade 1 was maintained for four weeks or more and PR was reported when there was a decrease by two or more degrees. Similarly, SD was noted when there was a decrease by one degree or no change appeared, and PD was considered when there was an increase of one or more degrees. Weight changes were evaluated just before the first treatment and before each cycle: a weight gain of > 7% from baseline, sustained for more than four weeks, was considered as a positive response. Antitumor efficacy was observed by CT-scan according to the standard SWOG response criteria. Response to treatment was assessed by monitoring CA 19-9 levels and was denned as: CR when the value entered into the normal range; PR when there was a decrease of more than 50% of initial value; PD for any increase; SD when the response did not meet the above criteria. Survival analysis was used to describe the global mortality of the cohort. Failure time analysis of prognostic factors was carried out by univariate and multivariate statistical methods [13]. Univariate analysis (product limit method) was used to calculate one to two-year survival probabilities in relation to seven clinical variables (gender, stage, tumor size, pain, jaundice, diabetic status, and weight loss) and six treatment-related variables (number of administered cycles, selective intra-arterial infusion, pain response, CT-scan response, CA 19-9 response, and weight gain). Cumulative probabilities of survival were adjusted by age, weighting the counts of survivors and non-survivors according to the predicted relative hazards for the observations based on age. Differences between survival curves were assessed using the log-rank test. Multivariate analysis was performed using the proportional hazards model, stepwise selection, after stratification according to age, on the 'dummy' variables obtained for each modality of the independent variables.
Results From January 1994 to June 1997, 96 consecutive patients with exocrine pancreatic cancer entered the study at seven institutions in Italy, following approval by an ethical review committee. A summary of the patient characteristics is given in Table 1. Seventy-three patients were treated by selective infusion according to tumor location; in twenty-three patients this procedure was not possible and the drugs were infused into the celiac axis. A total of 341 treatment cycles were administered.
Table 1. Patient characteristics. Characteristics
Number of patients
Patients entered Age (in years) Median Range Sex Male Female Performance status 0 1 2 Location of the tumor Head Body Tail Stage III IV Prior surgery GEA BDA Histology Adenocarcinoma Mucinous Undifferentiated
96 63 40-76 53 43 22 38 36 62 25 9 48 50 24 43 91 4 2
Abbreviations: GEA - gastroentero-anastomosis; BDA - biliary derivative anastomosis.
Thirty-seven percent of patients required no therapy for pain (CR). Pain improved, requiring less treatment (PR) in 5% of patients and was unchanged (SD) in 47%. PD, denned as the requirement for the use of increased pain-modifying therapy, was recorded in 11% of patients. Twenty-four percent of patients were not assessable by CTscan because of early death or because treatment was stopped before the third cycle. None had a CR to therapy, 15% had a PR (considering all the patients) and 44% were reported to have SD. In contrast, using measurements of CA 19-9 levels, 8% of patients had a CR, and the majority of patients (43%) had a PR. The overall median survival was 9.9 months, and 10.6 and 6.8 months for UICC stage III and IV, respectively. According to the following characteristics at baseline: tumor size, pain, jaundice, weight loss, diabetic status, derivative surgery, there was no significant difference in survival. Survival was found to be affected by gender (female vs. male) (P = 0.015), UICC stage (III vs. IV) (P = 0.05), number of administered courses (P - 0.0001), pain response (P - 0.001) and decrease of CA 19-9 >50% (P = 0.02). Prognostic factors found to be significant in multivariate analysis were: number of courses (P = 0.0003) and gender (P = 0.02). The median time to clinical progression wasfivemonths. Of the patients with UICC stage III disease we observed peritoneal progression in 25 of 48 (52%), progression of primary tumor in 23 of 48 (48%), liver progression in 17 of 48 (35%); only one patient developed
571 Table 2. Toxicity (n = 96). Toxicity
Any grade of toxicity, n (%)
Grade 3 + 4, n (%)
Anemia Thrombocytopenia Leukopenia Nausea/vomiting Alopecia Diarrhea Fever without infection Gastric pain Ematemesis Cardiac Sudden death Preinfarction angina Transitory ischemic attack Iliac intimal dissection
55 (57) 49(51) 43(44) 39 (40) 23 (24) 13(14) 12(12) 10(10) 4(4) 2(2) 1(1)
9(9) 24(25) 13(14) 3(3) 15(16) 1(1) 0(0) 4(4) 4(4)
KD 1(1) 1(1)
pulmonary metastases. Patients with UICC stage IV disease failed initially or progressed locally, within the liver, or regionally into the peritoneum. Two patients developed metastases outside the peritoneal cavity. Tolerability was assessed in all 96 patients (Table 2). Grade 3-4 anemia was observed in 9% of the patients; grade 3-4 thrombocytopenia in 25%; and grade 3-4 leukopenia in 14% of the patients; four cases of ematemesis were observed during the thrombocytopenic period. Non-hematological toxicity was mild: vomiting requiring therapy was observed in 3% of the patients, grade 3 diarrhea in one patient; severe abdominal pain in four patients; and complete, but reversible, alopecia in 16% of patients. One sudden death occurred on day 23 following the third cycle: the patient had well-compensated ischemic cardiopathy and had had a diaphragmatic myocardial infarction four years before treatment. Other reported events included a pre-infarction angina after the second cycle (n = 1) and a transitory ischemic attack on day 2 after the first cycle (n = 1). Only one patient did not complete treatment due to an intimal dissection of the iliac artery.
Discussion Pancreatic cancer is a notably chemoresistant malignant disease; there are two important reasons for this drug resistance: the existence of both a biological and a mechanical drug 'barrier'. The 'biological barrier' is related to the expression of moderate to high levels of diverse multidrug resistance genes within the tumor. The best defined is the multidrug resistance gene (MDR1) product, a 170-kD plasma membrane glycoprotein (P-gp) [14]. Also K-ras and/755 genetic mutations appear in the majority of pancreatic adenocarcinomas and more importantly, both the ras oncogene and p53 tumor suppressor genes activate the expression of the MDR1 promoter [15, 16]. Two other
non-P-gp mechanisms of multidrug resistance are found in pancreatic cancers: MRP, an ATP-dependent efflux protein similar to P-gp and the lung cancer resistanceassociated protein (LRP) [6]. The primary pancreatic tumor is normally encased in a very dense, poorly vascularized, fibrotic pseudocapsule ('mechanical barrier') and the tumor is relatively hypovascular [17]. One method to increase the local regional drug concentration within the tumor is to directly infuse the tumor and the tumor-bearing region via its arterial blood supply [18]. The rationale for delivering chemotherapeutic drugs via intra-arterial infusion is promising, since most agents demonstrate a clear concentration-response [19]. Regional chemotherapy has improved the response rates and quality of life in patients with colorectal liver metastases [20, 21]. This technique is anatomically practical for treating pancreatic cancer, and is justified by the local regional spread and the biological behavior of pancreatic cancer. The pattern of regional metastastic spread is mainly confined to the abdominal cavity, i.e., regional lymph nodes, the liver and the peritoneal surfaces, with only 27% present as extra-abdominal metastases [23, 24]. Also, the patterns of recurrence after radical surgery clearly shows that most patients develop local recurrence within the resection bed and almost all develop liver metastases [25]. The arterial supply of the celiac axis is such that it can direct chemotherapy to the primary tumor, the lymph nodes of the tumor-bearing area and the liver. Furthermore, any drug not taken up on the first pass through this region is then redelivered to the liver via the portal system. The feasibility and efficacy of intra-arterial chemotherapy for the treatment of pancreatic cancer has been evaluated with several combination regimens using different delivery techniques: celiac axis infusion for five days [7, 8], celiac axis stop-flow infusion, abdominal stop flow hypoxic perfusion [10], regional chemotherapy with hemofiltration [9], and one-day intra-arterial chemotherapy [11]. Based on previous experience, hepatic, intra-arterial CP improved the treatment of metastases from ocular melanoma [26]. Also, pre-operative intra-arterial treatment of advanced gastric cancer with CP and EPI showed a remarkable accumulation of the drugs in the regional lymph nodes [27]. Furthermore, we found that the complete response rate improved after intermittent intra-arterial infusion of 5-FU and CP for liver metastases due to pancreatic cancer [28]. This experience led us to add CP and EPI to the standard drug regimen of 5-FU and FA. The overall survival of all patients in this study was 9.9 months, and 10.6 and 6.8 for stage III and IV, respectively: the 12-, 18- and 24-month survival rates were 35%, 15%, 7%, respectively. Disease stage seems not to have a significant prognostic value from a multivariate analysis: pancreatic
572 cancer spreads early to regional lymph nodes and subclinical liver metastases might be present in the majority of patients at the time of diagnosis, even when findings from imaging studies are normal. In the UICC stage III patients, we denoted the sites of progression or failure to be peritoneal surfaces (52%), the primary tumor (48%), and the liver (35%). This is an interesting point because the expected incidence of hepatic progression is usually greater than 35%. These data seem to suggest that regional chemotherapy improves the control of hepatic progression and confirm the results of Muchmore, Ohigashi, and Link [9, 28, 29]. There is a relatively wide reported variance in the responses as determined from CT-scan data. In fact, the observed partial responses resulting from regional therapy vary from 15% in our study to 69% reported in studies by Aigner et al. [7], Link et al. [29] observed a 21% PR, and the patients treated in the study by Muchmore [9] achieved a 45% PR. This variance suggests that objective response rates may not be always uniformly determined. The objective response rate may not remain the gold standard in evaluating the results of treating advanced pancreatic cancer. The recent effort to identify a reproducible means to evaluate the quality of life by Clinical Benefit Respons [30] is justified and encouraging. The present study started before the introduction of Clinical Benefit Response. We evaluated pain intensity and weight changes. Regarding pain, we obtained a response (CR + PR) in 42% of the patients, SD in 47%, and only 11% of the patients had an increase in pain intensity. Comparing feasibility, compliance and results of our one-day intra-arterial regimen with the other more complicated techniques like abdominal stop-flow perfusion, regional chemotherapy with hemofiltration and five-days' celiac infusion, our technique is simple and offers cost-effective use of regional chemotherapy. The results indicate that this drug combination, administered as an intra-arterial infusion, is feasible and active, requires only one day of hospitalization and should be considered as a potential addition to integrated strategies in the treatment of pancreatic cancer.
References 1. Parker SI,TongT, Bolden S,Wingo PA. Cancer statistics 1997. CA 1997: 47: 5-27. 2. Schnall SF, Macdonald JS. Chemotherapy of adenocarcinoma of the pancreas. Semin Oncol 1996; 23: 220-8. 3. Evans DB, Abbruzzese JL, Rich TA. Cancer of the pancreas. In De Vita VT, Hellman S, Rosenberg SA (eds): Cancer, Principles and Practice of Oncology, fifth edn. Philadelphia. Pennsylvania: Lippincott 1997; 1054-87. 4. Burris HA, Moore MJ, Andersen J et al. Improvements in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreatic cancer: A randomized trial. J Clin Oncol 1997; 15: 2403-13. 5. Miller DW, Fontain M, Kolar C, Lawson T. The expression of multidrug resistance-associated protein (MRP) in pancreatic adenocarcinoma cell lines. Cancer Lett 1996: 107: 301-6.
6. Verovski VN, Van den Berge DL, Delvaeye MM et al. Low-level doxorubicin resistance in P-glycoprotein-negative human pancreatic tumor PSN1/ADR cells implicates a brefeldin A-sensitive mechanism of drug extrusion. Br J Cancer 1996; 73: 596-602. 7. Aigner KR, Muller H, Bassermann R. Intra-arterial chemotherapy with MMC, CDDP and 5-FU for non resectable pancreatic cancer: A phase II study. Reg Cancer Treat 1990; 3: 1-6. 8. Link KH, Gansauge F, Pillash J et al. Regional treatment of advanced nonresectable and of resected pancreatic cancer via celiac axis infusion. Dig Surg 1994; 11: 414-9. 9. Muchmore JH, Preslan JE, George WJ. Regional chemotherapy for inoperable pancreatic cancer. Cancer 1996; 78: 664-73. 10. Fiorentini G, Poddie D, Ricci S et al. Intra-aortic stop-flow infusion with hypoxic abdominal perfusion in UICC stage IIIIV pancreatic carcinoma. Report of a phase II study. Reg Cancer Treat 1996; 9: 88-91. 11. Cantore M, Fiorentini G, Bassi C et al. Intra-arterial chemotherapy for stage III—IV pancreatic cancer. Dig Surg 1997; 14: 113-8. 12. Muller H, Aigner KR, Basserman R. Influence of diagnostic techniques on response evaluation in regional chemotherapy histologic findings after isolated pelvic perfusion. In Sugarbaker PH (ed): Pelvic Surgery and Treatment for Cancer. St. Louis: Mosby 1994; 23-31. 13. Allgulander C, Fisher LD. Survival analysis (or time to an event analysis), and the Cox regression model-methods for longitudinal psychiatric research. Acta Psychiatr Scand 1986; 74: 529-35. 14. Goldstein LJ, Galski H, Fojo A et al. Expression of the multidrug resistance gene product P-glycoprotein in human cancer. J Natl Cancer Inst 1989; 81. 116-24. 15. Scarpa A, Capelli P, Mukai K et al. Pancreatic adenocarcinomas frequently show p53 gene mutations. Am J Pathol 1993; 142: 1534-43. 16. Chin KV, Ueda K, Pastan I, Gottesman MM. Modulation of the activity of the promoter of the human MDR1 gene by ras andp53. Science 1992, 255: 459-62. 17. Ishida H, Makino T, Kobayashi M, Tsuneoka K. Laparoscopic measurement of pancreatic blood flow. Endoscopy 1983; 15. 10710. 18. Collins JM. Pharmacologic rationale for regional drug delivery. J Clin Oncol 1984; 2: 498-504. 19. Link KH, Kindler D, Hummel M, Buchler M. Dose response treatment studies with two pancreatic carcinoma cell lines in vitro. Digestion 1992; 52: 102 (Abstr). 20. Meta-Analysis Group in Cancer. Reappraisal of hepatic arterial infusion in the treatment of nonresectable liver metastases from colorectal cancer. J Natl Cancer Inst 1996; 88 (5): 252-8. 21. Allen-Mersh TG, Earlam S, Fordy C et al. Quality of life and survival with continuous hepatic artery floxuridine infusion for colorectal liver metastases. Lancet 1994; 344: 1255-60. 22. Weiss L, Harlos JP, Hartveit F et al. Metastatic pattern from cancers of the pancreas: An analysis of 558 autopsies. Reg Cancer Treat 1992; 4: 265-71. 23. Griffin JF, Smalley SR, Jewell W et al. Patterns of failure after curative resection of pancreatic carcinoma. Cancer 1990; 66: 5661. 24. Andren-Sandberg A, Westerdhal J, Ihse I. Recurrence after pancreatectomy for pancreatic carcinoma. Digestion 1992; 52: 67-72. 25. Cantore M. Fiorentini G. Aitini E et al. Intra-arterial hepatic carboplatin based chemotherapy for ocular melanoma metastatic to the liver. Report of a phase II study. Tumori 1994; 80: 37-9. 26. Masuyama M. Taniguchi H. Takeuchi K et al. Recurrence and survival rate of advanced gastric cancer after preoperative intra-arterial infusion therapy. Gan To Kagaku Ryoho 1994; 21: 2253-5. 27. Tanaka H, Taniguchi H, Koyama H et al. A case of complete response by intermittent intra-arterial injection of 5-fluoruracicl and carboplatin for liver metastases due to pancreatic cancer. Gan To Kagaku Ryoho 1993; 20: 1669-71.
573 28. Ohigashi H. Hishikawa O, Nakamori S et al. Evaluation of intraarterial infusion chemotherapy and radical pancreatectomy in patients with locally advanced pancreatic cancer. Gan To Kagaku Ryoho 1993: 20: 1672-5. 29. Link KH, Gansauge F, Pillasch J et al. Multimodality therapies in ductal pancreatic cancer. The future. Int J Pancreatol 1997: 21: 71-83. 30. Anderson JS. Burris HA. Casper E et al. Development of a new system for assessing clinical benefit for patients with advanced pancreatic cancer. Proc Am Soc Clin Oncol 1994; 13: 461 (Abstr 1600).
Received 19 January 2000; accepted 21 March 2000.
Correspondence to: M. Cantore, MD Department of Oncology C.Poma Hospital Viale Albertoni 1 46100 Mantova Italy E-mail: [email protected]
Original article Intra-arterial chemotherapy for unresectable pancreatic cancer M. Cantore,1 P. Pederzoli,2 G. Cornalba,3 G. Fiorentini,4 S. Guadagni,5 L. Miserocchi,1 A. Frassoldati,6 C. Ceravolo,7 F. Smerieri1 & J. H. Muchmore8 ' Departments of Oncology and Radiology 'C. Poma' Hospital, Manlova; 2 Department of Endocrine Surgery, University of Verona, 1Department of Radiology'Sacco' Hospital, University of Milan: *Department ofOncology Civil Hospital, Ravenna, s Department of Surgery, University of LAquila; 6Department of Oncology, University of Modena; 7Department of Surgery Civil Hospital, Crema, Italy; 8Department of Surgery, Tulane University School of Medicine, USA
Summary Background: A phase II trial of a new intra-arterial chemotherapy regimen for unresectable pancreatic cancer (UPC). Patients and methods: Ninety-six patients with UPC were treated with intra-arterial chemotherapy at three-weekly intervals. The schedule used was FLEC: 5-fluorouracil 1000 mg/m 2 , folinic acid 100 mg/m 2 , carboplatin 300 mg/m 2 ; epirubicin 60 mg/m 2 . Results: The overall response rates by CT-scan evaluation were: 15% partial response (PR), 44% stable disease (SD), 17% progressive disease (PD). The overall median survival was 9.9 months, and 10.6 and 6.8 for UICC stage III and IV, respectively. Pain reduction occurred in 42% of patients. A weight gain > 7% from baseline occurred in 8% of patients. A total of
Introduction
The incidence of pancreatic adenocarcinoma has increased over the last two decades in most Western countries [1]. Chemotherapy alone or in combination with radiotherapy has not significantly improved the survival of patients with UICC stage III-IV disease [2-4]. Thus, in addition to the discovery of new active drugs, current research involves an integrated strategy of new administration routes delivering high-dose chemotherapy to the tumor-bearing region drug. Pancreatic cancer is notably resistant to systemic chemotherapy, with most tumors expressing increased levels of the multidrug resistant (MDR) gene [5]. Currently, two other non-Pglycoprotein mechanisms of multidrug resistance are found in pancreatic cancers: the multidrug resistanceassociated protein (MRP) and the lung cancer resistanceassociated protein (LRP) [6]. Intra-arterial chemotherapy as a means of increasing drug delivery for the treatment of advanced, pancreatic cancer has been previously investigated. Preliminary results of non-randomized, small series studies demonstrate that pancreatic cancer is sensitive to regional chemotherapy [7-11]. This paper reports the results of a phase II study of a new and
341 courses of FLEC were administered. Grade 3-4 hematological toxicity was seen in 25% of patients; ematemesis in 4%; grade 3 gastrointestinal toxicity in 3%; and grade 3 alopecia in 16%. One sudden death, a pre-infarction angina, and a transitory ischemic attack were observed. The only complication related to the angiographic procedure was an intimal dissection of the iliac artery. Conclusions: The intra-arterial FLEC regimen was well tolerated and active. It requires only one day of hospitalization. Efficacy could only be assessed in a randomized study against a gemcitabine containing regimen.
Key words: chemotherapy, intra-arterial, liver metastasis, unresectable pancreatic cancer
simple locoregional approach in the treatment of unresectable pancreatic cancer (UPC).
Patients and methods Patients with biopsy-proven adenocarcinoma of the pancreas that was not suitable for curative resection were eligible for enrolment in the study. They were required to have a performance status less than 3 (Southwest Oncology Group (SWOG) rating), a pre-treatment white blood cell count > 3000/ul, platelet count > 120,000/nl and hemoglobin level >9.5 gm/dl. Patients' total bilirubin and serum creatinine levels were required to be <1.5 times the institutional upper limit of normal. Staging included abdominal sonography, total abdomen CT scan, chest X-ray. Weight, pain intensity and analgesic consumption, performance status and CA 19-9 were evaluated at study entry and after each cycle of regional therapy. All patients gave their informed consent according to our institutional guidelines, and the study received ethical approval. Three cycles of regional chemotherapy were administered at threeweekly intervals through an angiographic catheter (Simmons 2; 5 Fr) introduced via the femoral artery into the celiac axis. Each drug was diluted in 100 ml of normal saline and then infused over a period of 10 minutes, one after the other in the following order: folinic acid (FA) 100 mg/m 2 : 5-fluorouracil (5-FU) 1000 mg/m 2 ; carboplatin (CP) 300 mg/m 2 ; epirubicin (EPI) 60 mg/m 2 . The drugs were infused according to tumor location: i.e., the
570 gastroduodenal artery for tumors of the pancreatic head; the splenic artery for tumors of the pancreatic body-tail; and when liver metastases were present, half of the total dose was infused in the hepatic artery and half according to primary location. When it was not possible to introduce the tip of catheter into either the gastroduodenal or splenic arteries, the drugs were infused into the celiac axis. In addition, an antiemetic (granisetron 8 mg) and an H2-receptor antagonist (famotidine 40 mg) were given intravenously. Hematological growth factor (filgrastim) was given at a dose of 5 ug/kg/d from day 10 after chemotherapy, for six consecutive days. Patients were discharged the day after therapy. Complete blood cell counts were performed before each cycle and twice a week. The response was evaluated after three cycles. In cases of clinical response, independent of CT scan response, a further three cycles were administered. All patients were evaluated for response on an intention-to-treat basis. Modification of pain was objectively evaluated using a five-point rating scale [12]: grade 1 = no therapy required, grade 2 = only nonsteroidal anti-inflammatory drugs (NSAIDs); grade 3 = combination of NSAIDs, antidepressants, and anticonvulsants; grade 4 = morphine derivatives; grade 5 = combination of morphine derivatives, NSAIDs, antidepressants, and anticonvulsants. CR was considered when grade 1 was maintained for four weeks or more and PR was reported when there was a decrease by two or more degrees. Similarly, SD was noted when there was a decrease by one degree or no change appeared, and PD was considered when there was an increase of one or more degrees. Weight changes were evaluated just before the first treatment and before each cycle: a weight gain of > 7% from baseline, sustained for more than four weeks, was considered as a positive response. Antitumor efficacy was observed by CT-scan according to the standard SWOG response criteria. Response to treatment was assessed by monitoring CA 19-9 levels and was denned as: CR when the value entered into the normal range; PR when there was a decrease of more than 50% of initial value; PD for any increase; SD when the response did not meet the above criteria. Survival analysis was used to describe the global mortality of the cohort. Failure time analysis of prognostic factors was carried out by univariate and multivariate statistical methods [13]. Univariate analysis (product limit method) was used to calculate one to two-year survival probabilities in relation to seven clinical variables (gender, stage, tumor size, pain, jaundice, diabetic status, and weight loss) and six treatment-related variables (number of administered cycles, selective intra-arterial infusion, pain response, CT-scan response, CA 19-9 response, and weight gain). Cumulative probabilities of survival were adjusted by age, weighting the counts of survivors and non-survivors according to the predicted relative hazards for the observations based on age. Differences between survival curves were assessed using the log-rank test. Multivariate analysis was performed using the proportional hazards model, stepwise selection, after stratification according to age, on the 'dummy' variables obtained for each modality of the independent variables.
Results From January 1994 to June 1997, 96 consecutive patients with exocrine pancreatic cancer entered the study at seven institutions in Italy, following approval by an ethical review committee. A summary of the patient characteristics is given in Table 1. Seventy-three patients were treated by selective infusion according to tumor location; in twenty-three patients this procedure was not possible and the drugs were infused into the celiac axis. A total of 341 treatment cycles were administered.
Table 1. Patient characteristics. Characteristics
Number of patients
Patients entered Age (in years) Median Range Sex Male Female Performance status 0 1 2 Location of the tumor Head Body Tail Stage III IV Prior surgery GEA BDA Histology Adenocarcinoma Mucinous Undifferentiated
96 63 40-76 53 43 22 38 36 62 25 9 48 50 24 43 91 4 2
Abbreviations: GEA - gastroentero-anastomosis; BDA - biliary derivative anastomosis.
Thirty-seven percent of patients required no therapy for pain (CR). Pain improved, requiring less treatment (PR) in 5% of patients and was unchanged (SD) in 47%. PD, denned as the requirement for the use of increased pain-modifying therapy, was recorded in 11% of patients. Twenty-four percent of patients were not assessable by CTscan because of early death or because treatment was stopped before the third cycle. None had a CR to therapy, 15% had a PR (considering all the patients) and 44% were reported to have SD. In contrast, using measurements of CA 19-9 levels, 8% of patients had a CR, and the majority of patients (43%) had a PR. The overall median survival was 9.9 months, and 10.6 and 6.8 months for UICC stage III and IV, respectively. According to the following characteristics at baseline: tumor size, pain, jaundice, weight loss, diabetic status, derivative surgery, there was no significant difference in survival. Survival was found to be affected by gender (female vs. male) (P = 0.015), UICC stage (III vs. IV) (P = 0.05), number of administered courses (P - 0.0001), pain response (P - 0.001) and decrease of CA 19-9 >50% (P = 0.02). Prognostic factors found to be significant in multivariate analysis were: number of courses (P = 0.0003) and gender (P = 0.02). The median time to clinical progression wasfivemonths. Of the patients with UICC stage III disease we observed peritoneal progression in 25 of 48 (52%), progression of primary tumor in 23 of 48 (48%), liver progression in 17 of 48 (35%); only one patient developed
571 Table 2. Toxicity (n = 96). Toxicity
Any grade of toxicity, n (%)
Grade 3 + 4, n (%)
Anemia Thrombocytopenia Leukopenia Nausea/vomiting Alopecia Diarrhea Fever without infection Gastric pain Ematemesis Cardiac Sudden death Preinfarction angina Transitory ischemic attack Iliac intimal dissection
55 (57) 49(51) 43(44) 39 (40) 23 (24) 13(14) 12(12) 10(10) 4(4) 2(2) 1(1)
9(9) 24(25) 13(14) 3(3) 15(16) 1(1) 0(0) 4(4) 4(4)
KD 1(1) 1(1)
pulmonary metastases. Patients with UICC stage IV disease failed initially or progressed locally, within the liver, or regionally into the peritoneum. Two patients developed metastases outside the peritoneal cavity. Tolerability was assessed in all 96 patients (Table 2). Grade 3-4 anemia was observed in 9% of the patients; grade 3-4 thrombocytopenia in 25%; and grade 3-4 leukopenia in 14% of the patients; four cases of ematemesis were observed during the thrombocytopenic period. Non-hematological toxicity was mild: vomiting requiring therapy was observed in 3% of the patients, grade 3 diarrhea in one patient; severe abdominal pain in four patients; and complete, but reversible, alopecia in 16% of patients. One sudden death occurred on day 23 following the third cycle: the patient had well-compensated ischemic cardiopathy and had had a diaphragmatic myocardial infarction four years before treatment. Other reported events included a pre-infarction angina after the second cycle (n = 1) and a transitory ischemic attack on day 2 after the first cycle (n = 1). Only one patient did not complete treatment due to an intimal dissection of the iliac artery.
Discussion Pancreatic cancer is a notably chemoresistant malignant disease; there are two important reasons for this drug resistance: the existence of both a biological and a mechanical drug 'barrier'. The 'biological barrier' is related to the expression of moderate to high levels of diverse multidrug resistance genes within the tumor. The best defined is the multidrug resistance gene (MDR1) product, a 170-kD plasma membrane glycoprotein (P-gp) [14]. Also K-ras and/755 genetic mutations appear in the majority of pancreatic adenocarcinomas and more importantly, both the ras oncogene and p53 tumor suppressor genes activate the expression of the MDR1 promoter [15, 16]. Two other
non-P-gp mechanisms of multidrug resistance are found in pancreatic cancers: MRP, an ATP-dependent efflux protein similar to P-gp and the lung cancer resistanceassociated protein (LRP) [6]. The primary pancreatic tumor is normally encased in a very dense, poorly vascularized, fibrotic pseudocapsule ('mechanical barrier') and the tumor is relatively hypovascular [17]. One method to increase the local regional drug concentration within the tumor is to directly infuse the tumor and the tumor-bearing region via its arterial blood supply [18]. The rationale for delivering chemotherapeutic drugs via intra-arterial infusion is promising, since most agents demonstrate a clear concentration-response [19]. Regional chemotherapy has improved the response rates and quality of life in patients with colorectal liver metastases [20, 21]. This technique is anatomically practical for treating pancreatic cancer, and is justified by the local regional spread and the biological behavior of pancreatic cancer. The pattern of regional metastastic spread is mainly confined to the abdominal cavity, i.e., regional lymph nodes, the liver and the peritoneal surfaces, with only 27% present as extra-abdominal metastases [23, 24]. Also, the patterns of recurrence after radical surgery clearly shows that most patients develop local recurrence within the resection bed and almost all develop liver metastases [25]. The arterial supply of the celiac axis is such that it can direct chemotherapy to the primary tumor, the lymph nodes of the tumor-bearing area and the liver. Furthermore, any drug not taken up on the first pass through this region is then redelivered to the liver via the portal system. The feasibility and efficacy of intra-arterial chemotherapy for the treatment of pancreatic cancer has been evaluated with several combination regimens using different delivery techniques: celiac axis infusion for five days [7, 8], celiac axis stop-flow infusion, abdominal stop flow hypoxic perfusion [10], regional chemotherapy with hemofiltration [9], and one-day intra-arterial chemotherapy [11]. Based on previous experience, hepatic, intra-arterial CP improved the treatment of metastases from ocular melanoma [26]. Also, pre-operative intra-arterial treatment of advanced gastric cancer with CP and EPI showed a remarkable accumulation of the drugs in the regional lymph nodes [27]. Furthermore, we found that the complete response rate improved after intermittent intra-arterial infusion of 5-FU and CP for liver metastases due to pancreatic cancer [28]. This experience led us to add CP and EPI to the standard drug regimen of 5-FU and FA. The overall survival of all patients in this study was 9.9 months, and 10.6 and 6.8 for stage III and IV, respectively: the 12-, 18- and 24-month survival rates were 35%, 15%, 7%, respectively. Disease stage seems not to have a significant prognostic value from a multivariate analysis: pancreatic
572 cancer spreads early to regional lymph nodes and subclinical liver metastases might be present in the majority of patients at the time of diagnosis, even when findings from imaging studies are normal. In the UICC stage III patients, we denoted the sites of progression or failure to be peritoneal surfaces (52%), the primary tumor (48%), and the liver (35%). This is an interesting point because the expected incidence of hepatic progression is usually greater than 35%. These data seem to suggest that regional chemotherapy improves the control of hepatic progression and confirm the results of Muchmore, Ohigashi, and Link [9, 28, 29]. There is a relatively wide reported variance in the responses as determined from CT-scan data. In fact, the observed partial responses resulting from regional therapy vary from 15% in our study to 69% reported in studies by Aigner et al. [7], Link et al. [29] observed a 21% PR, and the patients treated in the study by Muchmore [9] achieved a 45% PR. This variance suggests that objective response rates may not be always uniformly determined. The objective response rate may not remain the gold standard in evaluating the results of treating advanced pancreatic cancer. The recent effort to identify a reproducible means to evaluate the quality of life by Clinical Benefit Respons [30] is justified and encouraging. The present study started before the introduction of Clinical Benefit Response. We evaluated pain intensity and weight changes. Regarding pain, we obtained a response (CR + PR) in 42% of the patients, SD in 47%, and only 11% of the patients had an increase in pain intensity. Comparing feasibility, compliance and results of our one-day intra-arterial regimen with the other more complicated techniques like abdominal stop-flow perfusion, regional chemotherapy with hemofiltration and five-days' celiac infusion, our technique is simple and offers cost-effective use of regional chemotherapy. The results indicate that this drug combination, administered as an intra-arterial infusion, is feasible and active, requires only one day of hospitalization and should be considered as a potential addition to integrated strategies in the treatment of pancreatic cancer.
References 1. Parker SI,TongT, Bolden S,Wingo PA. Cancer statistics 1997. CA 1997: 47: 5-27. 2. Schnall SF, Macdonald JS. Chemotherapy of adenocarcinoma of the pancreas. Semin Oncol 1996; 23: 220-8. 3. Evans DB, Abbruzzese JL, Rich TA. Cancer of the pancreas. In De Vita VT, Hellman S, Rosenberg SA (eds): Cancer, Principles and Practice of Oncology, fifth edn. Philadelphia. Pennsylvania: Lippincott 1997; 1054-87. 4. Burris HA, Moore MJ, Andersen J et al. Improvements in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreatic cancer: A randomized trial. J Clin Oncol 1997; 15: 2403-13. 5. Miller DW, Fontain M, Kolar C, Lawson T. The expression of multidrug resistance-associated protein (MRP) in pancreatic adenocarcinoma cell lines. Cancer Lett 1996: 107: 301-6.
6. Verovski VN, Van den Berge DL, Delvaeye MM et al. Low-level doxorubicin resistance in P-glycoprotein-negative human pancreatic tumor PSN1/ADR cells implicates a brefeldin A-sensitive mechanism of drug extrusion. Br J Cancer 1996; 73: 596-602. 7. Aigner KR, Muller H, Bassermann R. Intra-arterial chemotherapy with MMC, CDDP and 5-FU for non resectable pancreatic cancer: A phase II study. Reg Cancer Treat 1990; 3: 1-6. 8. Link KH, Gansauge F, Pillash J et al. Regional treatment of advanced nonresectable and of resected pancreatic cancer via celiac axis infusion. Dig Surg 1994; 11: 414-9. 9. Muchmore JH, Preslan JE, George WJ. Regional chemotherapy for inoperable pancreatic cancer. Cancer 1996; 78: 664-73. 10. Fiorentini G, Poddie D, Ricci S et al. Intra-aortic stop-flow infusion with hypoxic abdominal perfusion in UICC stage IIIIV pancreatic carcinoma. Report of a phase II study. Reg Cancer Treat 1996; 9: 88-91. 11. Cantore M, Fiorentini G, Bassi C et al. Intra-arterial chemotherapy for stage III—IV pancreatic cancer. Dig Surg 1997; 14: 113-8. 12. Muller H, Aigner KR, Basserman R. Influence of diagnostic techniques on response evaluation in regional chemotherapy histologic findings after isolated pelvic perfusion. In Sugarbaker PH (ed): Pelvic Surgery and Treatment for Cancer. St. Louis: Mosby 1994; 23-31. 13. Allgulander C, Fisher LD. Survival analysis (or time to an event analysis), and the Cox regression model-methods for longitudinal psychiatric research. Acta Psychiatr Scand 1986; 74: 529-35. 14. Goldstein LJ, Galski H, Fojo A et al. Expression of the multidrug resistance gene product P-glycoprotein in human cancer. J Natl Cancer Inst 1989; 81. 116-24. 15. Scarpa A, Capelli P, Mukai K et al. Pancreatic adenocarcinomas frequently show p53 gene mutations. Am J Pathol 1993; 142: 1534-43. 16. Chin KV, Ueda K, Pastan I, Gottesman MM. Modulation of the activity of the promoter of the human MDR1 gene by ras andp53. Science 1992, 255: 459-62. 17. Ishida H, Makino T, Kobayashi M, Tsuneoka K. Laparoscopic measurement of pancreatic blood flow. Endoscopy 1983; 15. 10710. 18. Collins JM. Pharmacologic rationale for regional drug delivery. J Clin Oncol 1984; 2: 498-504. 19. Link KH, Kindler D, Hummel M, Buchler M. Dose response treatment studies with two pancreatic carcinoma cell lines in vitro. Digestion 1992; 52: 102 (Abstr). 20. Meta-Analysis Group in Cancer. Reappraisal of hepatic arterial infusion in the treatment of nonresectable liver metastases from colorectal cancer. J Natl Cancer Inst 1996; 88 (5): 252-8. 21. Allen-Mersh TG, Earlam S, Fordy C et al. Quality of life and survival with continuous hepatic artery floxuridine infusion for colorectal liver metastases. Lancet 1994; 344: 1255-60. 22. Weiss L, Harlos JP, Hartveit F et al. Metastatic pattern from cancers of the pancreas: An analysis of 558 autopsies. Reg Cancer Treat 1992; 4: 265-71. 23. Griffin JF, Smalley SR, Jewell W et al. Patterns of failure after curative resection of pancreatic carcinoma. Cancer 1990; 66: 5661. 24. Andren-Sandberg A, Westerdhal J, Ihse I. Recurrence after pancreatectomy for pancreatic carcinoma. Digestion 1992; 52: 67-72. 25. Cantore M. Fiorentini G. Aitini E et al. Intra-arterial hepatic carboplatin based chemotherapy for ocular melanoma metastatic to the liver. Report of a phase II study. Tumori 1994; 80: 37-9. 26. Masuyama M. Taniguchi H. Takeuchi K et al. Recurrence and survival rate of advanced gastric cancer after preoperative intra-arterial infusion therapy. Gan To Kagaku Ryoho 1994; 21: 2253-5. 27. Tanaka H, Taniguchi H, Koyama H et al. A case of complete response by intermittent intra-arterial injection of 5-fluoruracicl and carboplatin for liver metastases due to pancreatic cancer. Gan To Kagaku Ryoho 1993; 20: 1669-71.
573 28. Ohigashi H. Hishikawa O, Nakamori S et al. Evaluation of intraarterial infusion chemotherapy and radical pancreatectomy in patients with locally advanced pancreatic cancer. Gan To Kagaku Ryoho 1993: 20: 1672-5. 29. Link KH, Gansauge F, Pillasch J et al. Multimodality therapies in ductal pancreatic cancer. The future. Int J Pancreatol 1997: 21: 71-83. 30. Anderson JS. Burris HA. Casper E et al. Development of a new system for assessing clinical benefit for patients with advanced pancreatic cancer. Proc Am Soc Clin Oncol 1994; 13: 461 (Abstr 1600).
Received 19 January 2000; accepted 21 March 2000.
Correspondence to: M. Cantore, MD Department of Oncology C.Poma Hospital Viale Albertoni 1 46100 Mantova Italy E-mail: [email protected]