Curative and palliative aspects of regional chemotherapy in combination with surgery

Herwart Müller · Ralf Hilger

H. Müller
Department of Surgical Oncology, Hammelburg, Germany
R. Hilger
Department of Internal Medicine, University Hospital, Essen, Germany
E-mail: h. mueller@ surgicaloncology. de
Phone: +49-9732-900156
Fax: +49-9732-900159

Published online:

Keywords: Advanced tumours -Locoregional treatment -Metastatic disease -Non-small-cell lung cancer -Induction therapy

Introduction
Two major changes that will be important in the management of cancer over the next decade are emerging: first, 'adequate cancer surgery' will be newly defined, with surgery as just one component of a multimodal strategy. Secondly, most chemotherapies will be administered regionally, in addition to systemic treatments. However, not only will there be a change in the route used for chemotherapy administration, but its timing will also be changed. The majority of regional chemotherapy treatments will be administered during the preoperative and perioperative periods, and not in the follow-up period.

It is what the surgeon does not see that kills the patient. Cancer is a cellular disease that disseminates in most patients as it progresses, even though initially it is anatomically contained as a primary tumour mass. Very often, and perhaps in most patients, circulating cancer cells from this primary tumour may be present for years before the cells adhere, implant, and lead to macroscopically manifest disease.

This dissemination process is not equivalent throughout the vascular and lymphatic system; there is a marked regional predominance in this metastatic process. There is a great theoretical advantage in targeting the highest concentration of cytostatic agents at the anatomical site that is situated most proximally in relation to the primary cancer and that is at greatest risk of disease progression.

Regional chemotherapy is the form best suited for eradication of microscopic residual disease located in anatomical proximity to the primary malignancy. In addition to preventing local recurrence, regional chemotherapy has the great advantage of also providing systemic prophylaxis against the cascade phenomenon in which cells metastasise from the primary tumour to different organs. This brings about the unnecessarily rapid demise of the host. In this instance, not only is the optimal systemic dose of chemotherapy given, but also the desired magnified locoregional effect is achieved, based on the first-pass phenomenon.

Principles of regional chemotherapy
Regional administration of anticancer drugs may be recommended if the site of the cancer is known and if it is relatively localised and not disseminated. Many attempts have been made to enhance response by administering high concentrations of anticancer drugs selectively to the tumour tissue. It is advantageous for anticancer treatment if the toxic effects of the cytostatic on normal tissue are minimised by administering the drugs to the cancer tissue only, in such a way as to ensure that the drugs do not reach the normal tissue. Today, different formulations are available for administration by various methods, such as intra-arterial infusion, chemoembolisation, intraperitoneal administration or isolated perfusion techniques.

Dose-response relationship

Since the fundamental work of Collins et al. in 1981, it has been widely assumed that cytotoxic drugs have steep dose-response curves, i. e. there can be a large increase in cytotoxicity with a relatively small increase in drug exposure.
This has two different implications, namely (1) that attainment of very high drug concentrations in the target organ being regionally perfused will increase the proportion of tumour cells killed, and (2) that if less drug is delivered into the systemic vascular compartments there will be reduced exposure of rapidly dividing normal cell populations, such as bone marrow progenitors which will lead to less toxicity

Fig. 1. Dose-response relationship

Based on this dose-response relationship for cytotoxic drugs, the aims of regional chemotherapy will be
• To increase efficacy in the treated area by raising local drug concentrations.
• To avoid systemic toxicity by decreasing systemic drug exposure.

If there is no need to prevent systemic toxicity, it will be possible to use combined systemic and regional chemotherapy in order to reach a high local efficacy associated with systemic drug levels high enough to be effective against micrometastatic cancer.

Different administration techniques

Intraperitoneal administration
During the last three decades, the arsenal of regional administration forms could be greatly enlarged.
In the early 1950s, Weissberger et al. treated ovarian cancer patients intraperitoneally with nitrogen mustard and demonstrated impressive control of malignant ascites. Unfortunately, this study and other early clinical investigations were unable to demonstrate any therapeutic impact of the i. p. approach on peritoneal tumour masses. In spite of the convincing pharmacokinetic and experimental data indicating that i. p. chemotherapy is more effective than systemic administration of cytostatics in patients with cancer restricted to the peritoneal cavity, it was still unclear up to 1995 whether i. p. treatment would be superior to i. v. treatment in terms of survival. Alberts et al. presented a carefully controlled prospective trial showing that the median survival of patients with minimal disease ovarian cancer (tumours <2 cm) was 8 months longer after i. p. cisplatin plus i. v. cyclophosphamide than after i. v. cisplatin plus i. v. cyclophosphamide.
Results of this study were confirmed by a large Intergroup Study in 2001, which showed a survival advantage for the group of patients treated by i. p. chemotherapy (i. v. paclitaxel + i. p. carboplatin) who survived for 28 as opposed to 22 months.

Intra-arterial administration

In 1950, Klopp et al. began to infuse anticancer drugs directly into regional arteries supplying the tumours to minimize systemic adverse reactions and to achieve higher concentrations of the drug in tumour tissue.
Since then, a large variety of different administration forms has been established for intra-arterial infusion in different target areas. The easiest mode of access to arteries supplying the tumour with blood is the transient insertion of catheters via a percutaneous angiography with the Seldinger technique. This technique makes catheterisation of many different arteries of the body readily possible in different areas.

Table 1. Arteries accessible for intra-arterial infusion

The problem involved in such angiographic catheterisation of different arteries supplying the tumour is the necessity of repeating the treatment and procedure. This leads to restrictions in the patient's daily life, which is the reason for implanting catheter devices that can be used repeatedly. Despite the advantages of such implantable devices in terms of secure placement and application whilst restrictions in the patient's life are avoided, such catheters also have several drawbacks. Moreover, an operation is necessary. There is also the risk of arterial thrombosis and infections. Last but not least, the catheter tip may become displaced. Several different catheter devices have been produced and several different implantation techniques have been published in recent years. This has led to an improvement in the success rate and in lower morbidity.

Regional perfusion techniques

In 1958, Ryan et al. and Creech et al. used extracorporeal circulation to perfuse anticancer drugs locally in organs that could be separated from the systemic circulation.
They attempted active intervention to eradicate all tumour cells and to decrease systemic adverse reactions to a greater degree than can be achieved by arterial infusion therapy. Both groups first established this technique for primary or recurrent tumours located in the pelvic area. This technique of isolated perfusion has been established for different areas of the body, such as the limbs, the liver, the thoracic region, the abdomen and the pelvis. Depending on the anticancer agent used and its dosage, varying degrees of leakage can occur, leading to a more or less pronounced systemic toxicity. As long as we can accept this leakage, it is possible to use special balloon catheter techniques for such isolated perfusions.
This special and minimally invasive administration form leads to a considerable reduction in side effects and morbidity.

Fig. 2. Scheme of isolated hypoxic abdominal perfusion

Such an abdominal perfusion can be established with two special balloon catheters, which are introduced into the aorta and vena cava via an inguinal approach under general anaesthesia. The balloons are placed under X-ray control in such a way that when the diaphragm is level they reliably occlude both vessels completely. Additionally, two Esmarch bandages are insufflated, one at the root of each leg. Via these inserted catheters, an isolated circuit of the abdominal region can be established for 30 min.
Using more or less the same technique, but with the bandages placed on the roots of the upper limbs, a reduction of the systemic circulation can be achieved. This technique is called isolated thoracic perfusion.

Fig. 3. Scheme of isolated thoracic perfusion

Pharmacokinetics

Intraperitoneal administration

The concentration differential between the peritoneal cavity and the systemic circulation arises because the rate of movement of the drug from the peritoneal cavity into the plasma (peritoneal clearance) is generally slow relative to the total body clearance. The pharmacokinetics of i. p. treatment is best understood as a three-compartment model. The drug is administered directly into the tumour-containing cavity (first compartment). It is then absorbed into the bloodstream (second compartment) and distributed to all other tissues in contact with the blood circulation (third compartment). After injection of a drug into the peritoneal cavity, its concentration gradually falls as the drug leaks into the systemic circulation and is distributed to other tissues as well as being metabolised or excreted from the body. The rate at which the concentration falls is a function of the volume of the peritoneal cavity (V), the surface of the peritoneal membrane through which the drug will diffuse (A), the permeability of the membrane (P) and the difference in free drug concentration between the cavity (Cpc) and the plasma (Cpl). This is defined quantitatively by the following equation.

This means that the greater the clearance from the systemic circulation and the smaller the clearance from the peritoneal cavity, the greater the concentration difference and the greater the advantage of intraperitoneal over intravenous chemotherapy will be. As illustrated in, this advantage depends directly on the drug used for cytostatic therapy.
Table 2. Pharmacokinetics of intraperitoneal chemotherapy

* Ratios of the free drug concentration

Intra-arterial administration

The greatest advantage of intra-arterial infusion is that the drug is distributed at high concentrations into the regional artery supplying the tumour during the initial circulation after administration. When the drug is circulated to the heart, it will be detoxified or excreted before recirculation by being passed through the liver and the kidneys. As a result, its blood concentrations during the next circulation cycle are no different from the levels achieved after intravenous infusion.
This advantage of intra-arterial infusion should be maximised to gain a better response than that obtained after i. v. therapy. To achieve this objective, the following criteria should be met:

• The drug should flow at high concentrations during the initial circulation and adhere to the tumour tissue.
• It should permeate the tissue and be taken up by the tumour cells, or be rapidly activated in the cells to exert its cytotoxic effect.
• It should be degraded, inactivated or excreted almost completely in a short period of time to reduce systemic toxicity.

Fig. 4. Difference in concentration of the agent in regional tissue after i. a. administration (broken lines) as compared with i. v. administration (continuous lines), assuming that one part in four of the agent passing through the tumor circulation is biologically active against tumour cells (x= 1/ 4). The upper scale indicates the difference in systemic blood concentration, which is minimal. Even though regional uptake is some 3 times as great after i. a. administration, in absolute values this is unlikely to be sufficient to make a significant difference in systemic blood concentration. The abscissa indicates the time scale, measured by the number of circulations. In the clinical situation the slope of the curves will be similar to those shown here, but the time scale will be more protracted, as some of the agent will move temporarily into the tissue fluid and subsequently return unchanged into the circulation, resulting in delayed detoxification or excretion.

The pharmacokinetics of intra-arterially administered cytostatics has been studied in detail by Stephens et al.

Based on cardiovascular physiology, they describe an excellent model to account for the advantage of intra-arterial over intravenous application. In this model, the i. v.-applied dose of a drug should be 100%. If it is assumed that 10% of the dose is distributed in the tumour and the regional artery supplying the blood to the tumour, x% of the dose may show its biological activity when it passes through the tumour. In contrast, if 100% of the dose is given into the regional artery supplying the tumour, then 10 x% of the dose should show biological activity against the tumour

during the initial circulation. After the first circulation, the drug is distributed almost equally irrespective of whether it is given intra-arterially (i. a.) or i. v. High concentrations that are achieved during the initial circulation after i. a. infusion are decreased depending on the rate of excretion and detoxification. When different modes of application, such as bolus i. v., continuous i. v. or i. a., are compared, the advantage of the last becomes very evident.

Fig. 5. Serum levels of Adriamycin in rabbits following bolus iv., bolus ia. and continuous ia. administration of 2 mg/ kg.

Fig. 6. Tumour levels of Adriamycin (ADR) in rabbits following bolus i. v., bolus i. a. and continuous i. a. administration of 2 mg/ kg.

Isolated perfusion techniques

To illustrate the benefit of isolated perfusion techniques, we have made a direct comparison between the pharmacokinetic data attained with high-dose i. v. chemotherapy and those attained with a regional mode of administration, especially isolated abdominal perfusion. This special technique has shown promising efficacy in patients with heavily pretreated, recurrent peritoneally metastasised ovarian cancer. Treosulfan was used at a dosage of 2,600 mg/ m 2 for an abdominal perfusion of 30 min. In contrast to this, the maximum tolerated dose for the conventional systemic treatment was defined as 8,000 mg/ m 2.
High-dose chemotherapy with 10,000-56,000 mg/ m 2 . Treosulfan needs the support of autologous peripheral blood stem cell (PBSC) transplantation.
Pharmacokinetic data demonstrated a peak plasma concentration (C max ) after the i. p. administration of 2,600 mg/ m 2
treosulfan that was two-fold that after the i. v. administration of 10,000 mg/ m 2 treosulfan. In addition, when the i. p. route was used the area under the concentration-versus-time curves (AUC) reached exactly the same values with only a quarter of the dose delivered by the i. v. route, with no need of PBSC support.

Fig. 7. Pharmacokinetic data under isolated abdominal perfusion over 30 min by use of 2,600 mg/ m 2 treosulfan

Fig. 8. Pharmacokinetic data of isolated abdominal perfusion using 2,600 mg/ m 2 treosulfan versus high-dose systemic chemotherapy using 10,000 mg/ m 2

Results of regional chemotherapy plus surgery

Advanced non-small-cell lung cancer

Non-small-cell lung cancer (NSCLC) remains the leading cause of cancer deaths in both men and women. In the US in 1999, approximately 171,600 people were diagnosed with lung carcinoma, and 158,900 died as a consequence of lung cancer.
NSCLC accounts for approximately 75% of all lung carcinomas, and 35% of patients with NSCLC will present with stage IIIA or IIIB disease.
In the majority of these patients with mediastinal involvement the disease is not amenable to surgical resection, and primary radiation therapy alone results in 5-year actuarial survival of only 3-7% and a median survival time of 6-11 months.
The majority of these patients ultimately die of distant metastases. Recent efforts to improve their intermediate and long-term survival have therefore focused on neoadjuvant chemotherapy with or without radiotherapy as an induction regimen followed by surgical resection.

The theoretical advantages of the neoadjuvant approach include both systemic and local effects, such as:

• Early control of distant micrometastatic disease
• Prevention of visible tumour seeding at surgery
• An increase in resectability of neoblastic lesions technically unresectable at diagnosis
• A reduction of tumour mass before definitive radiotherapy
• Decreased incidence of positive margins at surgery
• Possibly the use of less radical surgery with organ preservation

Moreover, early chemotherapy has been associated with greater efficacy and improved drug delivery to tumour cells via the intact vascular system. Disadvantages of the neoadjuvant approach include morbidity and mortality related to the side effects of induction chemotherapy and an increase in surgical morbidity and mortality, as well as a delay in time to definitive surgery.
Since local control remains a substantial problem in patients with unresectable stage III NSCLC, strategies designed to enhance local control are likely to enable further improvements in the long-term outcome of patients with this disease. The purpose of a pilot study was to incorporate an isolated thoracic perfusion treatment into a strategy of induction therapy followed by surgery in the treatment of patients with primarily unresectable stage IIIA/ B NSCLC (Fig. 3).

The treatment plan consisted of a combination of regional plus continuous systemic chemotherapy. For regional chemotherapy, the cytostatics mitomycin 10 mg/ m 2 , navelbine 25 mg/ m 2 and cisplatin 30 mg/ m 2 were administered during isolated thoracic perfusion (ITP) via a central venous line on day 1. To prevent severe bone marrow dysfunction and activate immune function, 300 µg GM-CSF was administered during ITP. As systemic chemotherapy, 5-fluorouracil 250-mg/ m 2 and cisplatin 20 mg/ m 2 were given as a continuous infusion over 4 days via a central venous line starting after the end of ITP. The treatment-free interval was 4 weeks.
There were 10 patients with bulky stage III A disease and another 11 with stage III B disease, 6 of these 11 having T4 tumours. One patient had stage IV disease because of contralateral pulmonary metastases. In 1 patient, a third cycle of ITP plus systemic chemotherapy was added with the purpose of attaining further shrinkage of a centrally located T4 tumour.
The toxicity profile of this regional plus low-dose continuous systemic chemotherapy was absolutely acceptable, with only a few cases of grade 3 leucopenia (6.7%). Severe side effects, such as neutropenic fever, bleeding episodes, pneumonitis and lung fibrosis, were completely absent in this series.
The overall response rate of the 22 patients was 59%, but another 6 patients reached a minor response, leading to a regression rate of 86.3%. In this special situation, we have to bear in mind that resectability will be not defined by shrinkage of 50% or more as defined in WHO criteria. After improvements in both general condition and lung function parameters due to a response to therapy, resection of the primary tumour plus the mediastinal lymph-node metastases was possible in 16 out of 22 cases (72.7%). Complete pneumonectomy was carried out in 7 cases, whereas a lobectomy or bilobectomy was necessary to resect visible tumour lesions in the remaining 9 cases.
The rate of complications after major thoracic resection was low, with 1 early and 1 late minor bronchial stump insufficiency and no postoperative mortality.
After a median follow-up of 15 months, 5 of the 22 patients have died. In 4 cases, the deaths were related to malignant diagnosis, whereas 1 patient died of a cardiovascular cause not related to therapy.
The median duration of survival has not yet been reached, and the estimated 1-year actuarial survival was 67.3%. For 10 of the 22 patients with bulky stage III A disease the 1-year survival rate was 87.5%, whereas among those with stage III B/ IV disease only 58.3% were still alive after 1 year (Fig. 9).

Fig. 9. Kaplan-Meier curve for patients with regional induction chemotherapy plus tumor resection for non-small-cell lung cancer (NSCLC) stage IIIA, IIIB, IV

Our data have shown a high efficacy of regional chemotherapy as induction therapy for advanced NSCLC. The treatment modality is active in bulky stage N2 disease as well as stage III B/ IV disease confined to the thoracic region. Especially interesting in this study is the improvement in lung function parameters early after the first ITP, which opens the way to curative resection for patients in reduced condition and provides a new diagnostic technique.

Recurrent ovarian carcinoma

Ovarian cancer is the cause of the greatest number of deaths resulting from gynaecological malignant disease in the developed world. It is responsible for approximately 5% of all female cancer deaths.
A meta-analysis of the efficacy of second-line chemotherapy undertaken by McGuire has shown absolutely depressing results with response rates between 5% and 25% and median survival rates of only 5-6 months.

Fig. 10. Kaplan-Meier curve for 133 patients with colorectal liver metastases treated by immuno-chemoembolization with and without surgery

In view of the consequent need for more effective treatment strategies, isolated abdominal perfusion technique has been developed in order to overcome tumour cell resistance by use of high-concentration regional chemotherapy.
During recent years, this technique has shown promising efficacy in heavily pretreated patients with recurrent abdominally metastasised ovarian cancer. It is always very difficult to use a standard and very strict cytostatic protocol in second-line treatment. Response rates are between 60% and 70%, with an absolutely acceptable toxicity profile. In this situation, we have to bear in mind that these patients have often been pretreated with several kinds of cytostatics and are in a poor general condition owing to ascites or small bowel obstruction.
Small-or large-bowel obstruction often results in a need for surgical intervention. Especially in cases with regressive disease, we have the chance to perform a second debulking procedure. In our series, a secondary resection of abdominal or retroperitoneal tumour mass was possible in 55% of all cases, leading to a reduction in tumour volume, in pain and in symptoms such as ureter obstruction.
The combination of an effective regional chemotherapy for the abdomen and an effective surgical strategy with a low complication rate resulted in a prolongation of survival for this group of patients, with a median survival rate of around 20 months and a 1-year survival rate of 67%.

Colorectal liver metastases

Metastases confined to the liver are very common in the course of colorectal disease: 60-70% of all patients with colorectal carcinoma will develop liver metastases during their life, and only around 10% of all cases are resectable.
Right now, systemic chemotherapy is commonly used in the case of disseminated and unresectable metastases, and results have been improved in recent years with the use of new cytostatics, such as irinotecan or oxaliplatin, and new protocols. In contrast to the commonly applied strategy of systemic chemotherapy, several large randomised trials during the last two decades have shown an advantage, in terms of response rate and survival, for regional over systemic chemotherapy (Table 3).
Kemeny and co-workers reported one of the most interesting trials comparing the i. a. and the systemic approach in the adjuvant setting after curative liver resection for colorectal metastases in 1999, showing a dramatic survival benefit.

Table 3. Randomised studies about regional chemotherapy for colorectal liver metastases

To attain resectability in patients with marginally resectable or unresectable liver metastases, a special kind of i. a. induction chemotherapy was administered using a combination of regional immunotherapy and chemoembolisation. Via an angiographically placed hepatic artery catheter, 150 µm granulocyte-macrophage-stimulating factor (GM-CSF) was administered over 60 min, followed by continuous 24-h infusion of 5-fluoruracil administered in a circadian mode of 2,500 mg on days 1 and 2. 5-Fluorouracil application was combined with low-dose systemic infusion of 100 mg leucovorin.
Chemoembolisation using 40 mg melphalan as the cytostatic and lipiodol and Gel-foam as embolising substances was carried out on day 3.

This treatment protocol consists of five different parts:
• Intra-arterial administration of GM-CSF
• Circadian administration of 5-FU in order to minimize regional and systemic side effects
• Modulation of 5-fluorouracil with leucovorin
• Targeting of tumour vessels using lipiodol
• Increase of cytostatic drug concentration by embolisation and flow reduction with Gel-foam

After two cycles of this chemotherapy patients underwent surgery whenever complete resection seemed feasible. A hepatic artery catheter system was implanted and three cycles of adjuvant i. a. chemotherapy were added. In 28 out of 131 patients with colorectal liver metastases treated in the Department of Surgical Oncology at the Carl von Hess Hospital during the past 4.5 years, it was possible to resect all intrahepatic tumours. This corresponds to a resection rate of 21.4% in an unselected series with 57.3% of all patients pretreated with systemic chemotherapy. Resections of the liver consisted of multiple atypical segment resections in 5 cases, bisegment resections in 10, trisegment resection in 6 and extended hemihepatectomy in 7 cases. Resection margins were free in 24 cases, whereas tumour cells were visible microscopically in 4 patients.
There was no postoperative mortality and a low rate of postoperative side effects and complications.
Side effects after the first two cycles of chemotherapy were acceptable and reversible. Interestingly enough, there was a low rate of leucopenia. The histologically verified proven rate of remission in the resected cases was 82%, with complete remission in 17.8%.
In terms of survival, after a follow-up period of more than 3 years the median survival was not reached for the group of patients (28 patients) with complete tumour resection. In the group of 103 patients treated with immuno-chemoembolisation alone the median survival was 30 months. The 2-year survival rate was 54% for the latter group, and 75% for the group of patients treated with chemotherapy plus surgery (Fig. 9).

Conclusions

During the last decade, new forms and techniques of administration for regional and highly concentrated chemotherapy have been established. All of these have shown promising efficacy in different tumour entities, even in the chemonaive as well as when given as second-line treatment. Pharmacokinetic evaluation of these application forms has shown that higher cytostatic drug concentrations can be attained with regional chemotherapy than with systemic high-dose therapy with autologous bone marrow transplantation, but without the need for expensive stem cell support. The rate of complications and side effects could be reduced even more by further standardisation.
There is increasing evidence that regional chemotherapy is one of the best and most efficient measures complementing surgery in a strategy of induction therapy for advanced and reasonably resectable or definitely not completely resectable solid tumours. This strategy is also feasible in advanced metastases with tumours located only in one area of the body or only one organ. As shown for ovarian cancer patients and those with NSCLC and colorectal liver metastases, this strategy makes it possible to achieve a situation and time period when the tumour is no longer detectable. This will be combined with a dramatic pain relief, a reduction in symptoms and an enhancement of the quality of life. The treatment strategy is associated with an acceptable rate of morbidity and a low mortality rate, and with at least a prolongation in survival. In the end, it will restore the patient's hopes for tomorrow.

Acknowledgement.

The support of Medac Pharmaceuticals (Hamburg, Germany) in carrying out the pharmacokinetic studies is gratefully acknowledged.
References

1. Alberts DS, Liu PY, Hannigan EV, et al (1995) Phase III study of intraperitoneal (IP) cisplatin (CDDP)/ intravenous (IV) cyclophosphamide (CPA) vs. IV CDDP/ IV CPA in patients with optimal disease stage III ovarian cancer: a SWOG-GOG-ECOG Intergroup study. (INT 0051) Proc Am Soc Clin Oncol 14: 273

2. Allen-MershTG, Earlam S, Fordy C, et al (1994) Quality of life and survival with continuous hepatic artery floxuridine infusion for colorectal liver metastases Lancet 344: 1255-1260

3. Baxevanis CN, Tsavaris NB, Papadhimitriou SI, et al (1997) Granulocyte-macrophage colony-stimulating factor improves immunological parameters in patients with refractory solid tumours receiving second-line chemotherapy: correlation with clinical responses. Eur J Cancer 33: 1202-1208

4. Berek JS, Bertelsen K, du Bois A, et al (2000) Epithelial ovarian cancer (advanced stage): consensus conference. Gynecol Obstet Fertil 28: 576-583

5. Chang AE, Schneider PD, Sugarbaker PH, et al (1987) A prospective randomised trial of regional versus systemic continuous 5-fluorodeoxyuridine chemotherapy in the treatment of colorectal liver metastases. Ann Surg 206: 985-993

6. Collins JM (1984) Pharmacological rational for regional drug delivery. J Clin Oncol 2: 498-504

7. Collins JM, Dedrick RL (1982) Pharmacokinetics of anticancer drugs. In: Chabner W (ed) Pharmacologic principles of cancer treatment. Saunders, Philadelphia, pp 77-79

8. Creech O, Krementz, ET, Ryan RF, Winblad JN (1958) Chemotherapy of cancer: regional perfusion utilizing an extracorporeal circuit. Ann Surg 148: 616-632

9. Fleming GF, Roth BJ, Baker SD, et al (2000) Phase I trial of paclitaxel and etoposide for recurrent ovarian carcinoma: a Gynecologic Oncology Group Study. Am J Clin Oncol 23: 609-613

10. Grage TB, Vassilopoulos PP, Shingleton WW (1979) Results of a prospective randomised study of hepatic artery infusion with 5-fluorouracil versus intravenous 5-fluorouracil in patients with hepatic metastases from colorectal cancer: a Central Oncology Group study. Surgery 86: 550-555

11. Harstrick A, Wilke H, Eberhardt W, Klaassen U, Strumberg D, Korn M, Scheulen ME, Baumgart J, Seeber S (1996) A phase I dose escalation trial of intravenous treosulfan in refractory cancer. Onkologie 19: 153-156

12. Hazuk MB, Bunn PA Jr (1992) Controversies in the nonsurgical treatment of Stage III non-small cell lung cancer. Am Rev Respir Dis 145: 967-977

13. Hilger RA, Harstrick A, Eberhardt W, Oberhoff C, Skorzec M, Baumgart J, Seeber S, Scheulen ME (1998) Clinical pharmacokinetics of intravenous treosulfan in patients with advanced solid tumors. Cancer Chemother Pharmacol 42: 99-104

14. Hohn DC, Stagg RJ, Freidman MA, et al (1989) A randomised trial of continuous intravenous versus hepatic intraarterial floxuridine in patients with colorectal cancer metastatic to the liver: the Northern California Oncology Group Trial J Clin Oncol 7: 1646-1654

15. Jett JR (1993) Current treatment of unresectable lung cancer. Mayo Clin Proc 68: 603-611

16. Kemeny N, Atiq OT (1999) Non-surgical treatment for liver metastases. Baillieres Best Pract Res Clin Gastroenterol 13: 593-610

17. Kemeny N, Daly J, Reichman B, et al (1987) Intrahepatic or systemic infusion of fluorodeoxyuridine in patients with liver metastases from colorectal carcinoma -a randomised trial. Ann Intern Med 107: 459-465

18. Kemeny N, Seiter K, Niedzwiecki D, et al (1992) A randomised trial of intrahepatic infusion of fluorodeoxyuridine with dexamethasone versus fluorodeoxyuridine alone in the treatment of metastatic colorectal cancer. Cancer 69: 327-334

19. Kemeny N, Cohen A, Seiter K, et al (1993) Randomised trial of hepatic arterial floxuridine, mitomycin, and carmustine versus floxuridine alone in previously treated patients with liver metastases from colorectal cancer. J Clin Oncol 11: 330-335

20. Kemeny N, Huang Y, Cohen AM (1999) Hepatic arterial infusion of chemotherapy after resection of hepatic metastases from colorectal cancer. N Engl J Med 341: 2039-2048

21. Kerr DJ (2000) Introduction: Pharmacokinetic basis of regional chemotherapy. In: Kerr DJ, McArdle CS (eds) Regional chemotherapy -theory and practice. Harwood Academic, Amsterdam, pp 1-7

22. Klopp CT, Alford TC, Bateman J, et al (1950) Fractionated intra-arterial cancer chemotherapy with methyl bis amine hydrochloride: a preliminary report. Ann Surg 132: 811-832

23. Landis SH, Murray T, Bolder S, Wingo PA (1999) Cancer statistics, 1999. CA Cancer J Clin 49: 33-64

24. Los G (2000) Intraperitoneal chemotherapy. In: Kerr DJ, McArdle CS (eds) Regional chemotherapy -theory and practice. Harwood Academic, Amsterdam, pp 9-27

25. Markman M, Bundy BN, Alberts DS, et al (2001) Phase III trial of standard-dose intravenous cisplatin plus paclitaxel versus moderately high-dose carboplatin followed by intravenous paclitaxel and intraperitoneal cisplatin in small-volume stage III ovarian carcinoma: an intergroup study of Gynaecologic Oncology Group, Southwestern Oncology Group, and Eastern Cooperative Oncology Group. J Clin Oncol 19: 1001-1007

26. Martin JK, O'Connell MJ, Wiegand HS, et al (1990) Intra-arterial floxuridine vs systemic fluorouracil for hepatic metastases from colorectal cancer. A randomised trial. Arch Surg 125: 1022-1027

27. McGuire WP (1992) Ovarian cancer. Experimental chemotherapy. Hematol Oncol Clin North Am 6: 927-940

28. Müller H, Guadagni S (2002) Regional plus systemic chemotherapy -an effective treatment in recurrent non small cell lung cancer. Eur J Surg Oncol 27: 190-105

29. Müller H, Nakchbandi W, Chatzissavvidis I, Valek V (2001) Intra-arterial infusion of 5-fluorouracil plus granulocyte-macrophage colony stimulating factor (GM-CSF) and chemoembolization with melphalan in the treatment of disseminated colorectal liver metastases. Eur J Surg Oncol 27: 652-661

30. Porta C, Moroni M, Nastasi G, Bobbio-Pallavicini E, Barazzoni G (1995) Utility of embolization or chemoembolization as second-line treatment in patients with advanced or recurrent colorectal carcinoma. Cancer 75: 2782-2784

31. Pougier P, Milan C, Lazorthes F, et al (1995) Prospective study of prognostic factors in patients with unresected hepatic metastases from colorectal cancer: Foundation Française de Cancérologie Digestive. Br J Surg 82: 1397-1400

32. Rougier P, Laplanche A, Huguier M et al (1992) Hepatic arterial infusion of floxuridine in patients with liver metastases from colorectal carcinoma: long-term results of a prospective randomised trial. J Clin Oncol 10: 1112-1118

33. Ryan RF (1958) Selected perfusion of isolated viscera with chemotherapeutic agents. Surg Forum 8: 158

34. Scheulen ME, Hilger RA, Oberhoff C, Casper J, Freund M, Josten KM, Bornhäuser M, Ehninger G, Berdel WE, Baumgart J, Harstrick A, Bojko P, Wolf HH, Schindler AE, Seeber S (2000) Clinical phase I dose escalation and pharmacokinetic study of high-dose chemotherapy with treosulfan and autologous peripheral blood stem cell transplantation in patients with advanced malignancies. Clin Cancer Res 6: 4209-4216

35. Stephens FO (1983) Pharmacokinetics of intra-arterial chemotherapy. In: Schwemmle K, Aigner K (eds) Vascular perfusion in cancer therapy. (Recent results in cancer research, vol 86) Springer, Berlin Heidelberg New York

36. Taguchi T (1979) Attainability of anticancer drugs into the tumour. Jpn J Cancer Clin 25: 782-788