Search

Preclinical and clinical data suggest that the combination of hepatic arterial bromodeoxyuridine (BrdUrd), a thymidine analogue radiation sensitizer, and high-dose three-dimensional conformal radiation therapy offer a high potential for improving the local control of intrahepatic cancers. A key step in the design of a successful protocol is to determine in patients the conditions for BrdUrd administration that would be expected to produce selective radiosensitization of the tumor. Therefore, we designed a clinical trial to assess BrdUrd incorporation into the DNA of hepatic colorectal metastases and normal liver after a 7-day continuous BrdUrd infusion at a dose rate of 25 mg/kg/day (the maximal tolerated dose for a 14-day infusion) for patients undergoing laparotomy for either resection of liver metastases or hepatic arterial catheter and pump placement. Thirteen patients were entered into this study. We found that the average replacement of thymidine by BrdUrd in the tumor and normal liver were 11.6 +/- 1.2% and 1.1 +/- 0.2%, respectively. This extent of incorporation would be expected to produce a single fraction radiation enhancement of 1.5 in the tumor without detectable sensitization of the normal liver. Immunohistochemical staining for BrdUrd revealed heterogeneity of incorporation with a range of approximately 60-80% of the cells labeled in different regions of the specimens. These findings suggest that hepatic arterial BrdUrd given at this dose and schedule has a high likelihood of producing clinically significant radiosensitization for patients with hepatic metastases from colorectal cancer. Furthermore, the demonstrated selectivity of tumor perfusion that can be obtained with hepatic arterial infusion combined with the high proliferative rate of colorectal metastases (versus normal liver) suggests that these patients may be good candidates for tumor-directed gene transfer therapy by using regionally delivered retroviral vectors.

The treatment of hepatic tumors remains unsatisfactory. These lesions receive most of their blood supply from the hepatic artery, therefore the hepatic artery administration of beta-emitting particulate radiopharmaceuticals is an attractive approach to deliver therapeutic irradiation to the liver and differentially to tumors within the liver.A Phase I dose escalation study of the hepatic tolerance to radiation delivered by 90Y containing glass microspheres was carried out in 24 patients with hepatic malignancy. Doses of 90Y microspheres to achieve an estimated whole-liver nominal absorbed radiation dose of 5000 cGy (two patients), 7500 cGy (six patients), 10,000 cGy (seven patients), 12,500 cGy (six patients), and 15,000 cGy (three patients) were administered via the hepatic artery. The administered nominal absorbed radiation dose (NARD) was estimated based on liver volume determined from CT scans and the assumption of uniform distribution of microspheres throughout the liver.No hematologic, hepatic or pulmonary toxicity was encountered in the dose range examined during a mean follow-up period of up to 53 mo. Reversible gastritis or duodenitis was encountered in four patients without imaging or biopsy evidence for extrahepatic deposition of microspheres. Response data, based on CT scans obtained 16 wk after treatment, showed progressive disease in eight patients, stable disease in seven patients, minimal response in four patients and partial response in five patients. Subsequent follow-up revealed three long-term survivors at 204, 216 and 228 wk.These preliminary data demonstrate that in the examined dose range, radiation may be safely delivered to liver tumors by means of 90Y glass microspheres with encouraging response data.

We have previously shown that the thymidine analogue radiation sensitizer bromodeoxyuridine (BrdUrd) is incorporated into human tumors to a greater extent than into the livers of athymic mice bearing these tumors as xenografts. However, incorporation into the intestine and bone marrow exceeds that of the tumor (T. S. Lawrence, M. A. Davis, J. Maybaum, S. K. Mukhopadhyay, P. L. Stetson, D. P. Normolle, P. E. McKeever, and W. D. Ensminger, Cancer Res., 52: 3698-3704, 1992). We hypothesized that the ratio of tumor incorporation to intestinal or bone marrow incorporation might increase during a period of drug elimination following the termination of an infusion. To test this hypothesis, we infused athymic mice bearing HT29 human colon cancer xenografts with BrdUrd and measured incorporation in the tumor and normal tissues up to 7 days after the infusion was discontinued. In addition, we assessed the effect of exposure to BrdUrd on subsequent incorporation in vitro and in vivo through the use of a stable isotope of BrdUrd ("isotopic BrdUrd"), which could be differentiated from normotopic BrdUrd using the gas chromatographic-mass spectrometric assay. We found a significant increase in the ratio of BrdUrd in the tumor compared to bone marrow and intestine during the drug elimination period. We also found that BrdUrd incorporation slowed the kinetics of subsequent BrdUrd incorporation and elimination. These findings suggest that when the radiation dose-limiting organ is rapidly proliferative, such as the intestine or bone marrow, delivering radiation during a drug elimination period may improve the therapeutic index.

This investigation was undertaken to determine the pharmacokinetic parameters relevant to hepatic arterial (HA) infusion of 5-bromo-2'-deoxyuridine (BrdUrd) and to ascertain the maximum tolerated dose and related toxicities of BrdUrd administered as a 14-day HA infusion. In the pharmacokinetic study, 6 patients received a 2-h i.v. infusion (to steady-state) of BrdUrd at each of 5 escalating dose rates (10 to 160 mg/kg/day) with simultaneous blood sampling for BrdUrd levels from HA and hepatic venous catheters. Dose dependent HA and hepatic venous drug levels, total body clearance, hepatic extraction, and estimated regional exposure advantage were determined. The total body clearance of BrdUrd was high and dose rate dependent, falling from 3340 ml/min with a 10-mg/kg/day infusion to 2180 ml/min at a 160-mg/kg/day dose rate. Hepatic extraction was high and dose rate dependent as well, declining from 80% extraction at 10 mg/kg/day to 68% at 160 mg/kg/day. The calculated estimate for the exposure advantage achievable with HA as compared with i.v. infusion reflects the dose rate dependence of total body clearance and hepatic extraction and decreases from a 70-fold advantage at 10 mg/kg/day to a 30-fold advantage at 160 mg/kg/day. In the Phase I study aimed at determining the maximum tolerated dose, successive groups of 3 patients were administered continuous HA infusions for 14 days at escalated BrdUrd dose rates (5, 10, 15, 25, and 35 mg/kg/day) in order to ascertain dose-limiting toxicity. The maximum tolerated dose of BrdUrd for a 14-day continuous HA infusion was found to be 35 mg/kg/day with reversible thrombocytopenia as the sole dose-limiting toxicity. Skin and other toxicities were infrequent, minor, reversible, and non-dose dependent. No hepatic toxicity was detected despite direct drug infusion into the liver. The high total body clearance and hepatic extraction of BrdUrd substantiate its administration via the hepatic artery as a means to achieve higher exposure with intrahepatic tumors than can be obtained by systemic administration. Despite higher hepatic exposures, no hepatic toxicity was noted, and readily reversible systemic toxicity (thrombocytopenia) was dose limiting for the 14-day continuous HA infusion. Thus, HA infusion of the potent radiosensitizer BrdUrd is both pharmacokinetically rational and well tolerated.

Although the thymidine analogues 5-bromo-2'-deoxyuridine (BrdUrd) and 5-iodo-2'-deoxyuridine (IdUrd) have been used successfully as radiation sensitizers in clinical trials, it is not clear which of these agents is the more promising to pursue. To begin to assess this question with regard to colorectal cancer metastatic to the liver, a study was carried out using HT29 human colon cancer cells in culture and implanted in nude mice as xenografts. Cells and animals were treated with BrdUrd +/- the thymidylate synthase inhibitor 5-fluoro-2'-deoxyuridine (FdUrd), and the results compared to our previous studies with IdUrd +/- FdUrd (T. S. Lawrence, M. A. Davis, P. E. McKeever, J. Maybaum, P. L. Stetson, D. P. Normolle, and W. D. Ensminger. Cancer Res., 51: 3900-3905, 1991). Using cultured cells, it was found that FdUrd (at concentrations of greater than 10 nM) increased: (a) the incorporation of BrdUrd into the DNA of cultured tumor cells; (b) BrdUrd-mediated radiosensitization; (c) BrdUrd-mediated increase in radiation-induced DNA damage; and (d) BrdUrd-mediated decrease in the repair of radiation-induced damage. The incorporation of BrdUrd was greater than or equal to the incorporation of IdUrd previously determined under the same exposure conditions. Studies using nude mice bearing HT29 xenografts showed that FdUrd increased BrdUrd incorporation more into tumors than into the normal liver. Most tumor cells incorporated BrdUrd (labeling index after a 4-day infusion = 87 +/- 2%; SE); in the liver, labeling was confined chiefly to nonparenchymal cells. In both the presence and absence of FdUrd, the incorporation of BrdUrd into tumors was significantly and consistently greater than the incorporation of IdUrd measured under the same conditions of drug administration (by a factor of 1.2-3.6). Furthermore, the administration of BrdUrd +/- FdUrd tended to produce less weight loss and hematological toxicity than IdUrd +/- FdUrd. These findings suggest that BrdUrd may be superior to IdUrd as a radiation sensitizer in the treatment of colorectal cancer metastatic to the liver.

A study was conducted to assess the potential of 5-fluoro-2'-deoxyuridine (FdUrd) to increase the incorporation and radiosensitizing properties of 5-iodo-2'-deoxyuridine (IdUrd) using HT29 human colon cancer cells both in vitro and in nude mice bearing these tumors as xenografts. The purpose of this study was to assess (a) whether FdUrd could increase IdUrd efficacy using clinically achievable concentrations of drugs; (b) the relationships among radiosensitization, DNA damage and repair, and analogue incorporation; and (c) whether FdUrd improved the selectivity of IdUrd incorporation into tumor cells compared to normal tissues. It was found that FdUrd, at clinically achievable concentrations (1-100 nM), significantly increased IdUrd incorporation under all conditions but particularly when the IdUrd concentration was less than or equal to 10 microM. FdUrd increased IdUrd-mediated radiosensitization in proportion to the increase in IdUrd incorporation. FdUrd potentiated the ability of IdUrd to increase radiation-induced DNA double-strand breaks and to slow their repair. When IdUrd alone (100 and 200 mg/kg/day) was infused into nude mice bearing tumors, the extent of thymidine replaced in the tumor was 1.6 +/- 0.4 (mean +/- SE) and 2.5 +/- 0.4%, respectively. The combination of FdUrd (0.1 mg/kg/day) and IdUrd (100 mg/kg/day) increased the incorporation in the tumor to 5.3 +/- 0.9% with less toxicity than resulted from the use of 200 mg/kg/day of IdUrd alone. These data show that FdUrd is an effective biomodulator, because, for the same extent of normal tissue incorporation, the combination of IdUrd and FdUrd produces significantly greater incorporation into the tumor compared to the use of IdUrd alone. Furthermore, they suggest that the regional application of FdUrd with IdUrd, either through the use of regional infusions or in combination with focused irradiation, could potentially improve the outcome of treatment of localized gastrointestinal cancer.

Human glioma-derived cell lines were found to vary in their ability to incorporate the radiosensitizer 5-bromo-2'-deoxyuridine (BrdUrd) into DNA after one cell doubling. The U-251 cell line was the best incorporator of BrdUrd, whereas U-118 and D-54 demonstrated poor incorporation with respective C50 (BrdUrd concentration required for 50% of the maximum amount of BrdUrd incorporation into DNA) values of 2.8- and 6-fold greater than that of U-251 (P less than 0.001). Modulation of radiosensitizer uptake into DNA could be achieved using the thymidylate synthase inhibitors 5-fluorouracil or 5-fluoro-2'-deoxyuridine (FdUrd). Incorporation into U-251 cells increased only slightly in the presence of the fluoropyrimidines. The BrdUrd concentration required for 50% of the maximum amount of BrdUrd incorporation into DNA changed (P less than 0.001) from 1.8 +/- 0.11 microM (SD) in the absence of a modulator to 1.1 +/- 0.09 or 1.1 +/- 0.16 microM in the presence of 10 microM 5-fluorouracil or 5 nM FdUrd, respectively. The D-54 cell line, which was the worst incorporator of BrdUrd, was found to have an extensive amount of BrdUrd into DNA following biomodulation. The C50 in the absence of modulation was 7.3 +/- 1.3 microM, which was reduced (P less than 0.001) to 0.62 +/- 0.04 and 0.32 +/- 0.13 microM, respectively, in the presence of 10 microM 5-fluorouracil and 5 nM FdUrd. This represents a 12- to 22-fold reduction in the concentration of radiosensitizer required to achieve the same level of BrdUrd incorporation into DNA. Furthermore, this enhancement of BrdUrd DNA incorporation seen in the presence of the fluoropyrimidines is observed at clinically achievable concentrations. The degree of radiosensitization was solely dependent upon the amount of BrdUrd incorporated into DNA. D-54 cells grown in the presence of 0.18 microM BrdUrd plus 5 nM FdUrd or 2.8 microM BrdUrd alone yielded a similar level of BrdUrd incorporation into DNA and radiosensitization, though a 15-fold lower BrdUrd concentration was used in the presence of FdUrd. The combined use of a radiosensitizer with a fluoropyrimidine may overcome poor incorporation of BrdUrd into DNA that may exist among resistant subpopulations of cells within malignant glioma.

A mechanical infusion pump was used to introduce 99mTc-sulfur colloid at a rate of 1 ml/minute into the hepatic arteries of six patients. The radiotracer distribution patterns were compared to those obtained with "bolus" radionuclide angiography and those seen with standard contrast angiography. In all cases, the slow flow rate radionuclide studies evidenced dramatic differences from the patterns obtained with rapid flow rate injections. The results indicate that standard, rapid-injection contrast techniques may not produce the most reliable information regarding potential flow distribution. It is concluded that radiotracers introduced at flow rates approximating those attained with infusion pumps will offer the best estimates of both initial catheter placement and subsequent patterns of hepatic distribution of chemotherapeutic agents.

BCNU at a dose of 200 mg/m2 was administered in a 60-minute hepatic arterial (HA) infusion to three patients with liver-predominant neoplastic disease. Nitrosourea blood levels were measured during the 60-minute infusion and for 30 minutes after completion of the infusion in samples obtained simultaneously from a hepatic venous (HV) catheter and a peripheral venous (PV) site. Direct extraction of blood with diethyl ether was used in order to remove quantitatively the nitrosourea from blood and thereby stabilize it against in vitro breakdown in blood prior to colorimetric analysis. As determined by blood levels, steady-state was achieved within 50 minutes of infusion. Mean HV levels at steady-state were 2.5-fold higher in the HV compared to PV samples. The higher HV compared to PV levels must reflect higher tumor exposure to nitrosourea with HA compared to the usual iv route of administration. No drug-related hepatic toxicity was noted. Myelosuppression was noted in only one patient in whom reversible leukopenia (granulocyte nadir, 500/mm3) and thrombocytopenia (platelet nadir, 20,000/mm3) developed.

We evaluated a totally implanted system consisting of a subcutaneous injection port connected to a silicone elastomer central venous catheter for vascular access, including blood drawing, in 35 patients with cancer. All patients lacked peripheral venous access sites and were undergoing aggressive chemotherapy programs. The cumulative duration of successful access exceeded 2,900 days (for individual patients: range, five to 203 days; median, 61 days). In no instance were infusions or injections unsuccessful. Blood-sampling attempts were successful 90% of the time. The system did not require flushing between uses, being filled with heparinized saline after each entry. There were no instances of irreversible catheter occlusion or shear and no system-related infections. Thus, this device appears to have advantages over other central venous catheters in terms of patient acceptance and lack of maintenance between uses.

Regional cancer chemotherapy is a means of exploiting dose-response effects by delivering more drug to regionally confined tumors. Regional chemotherapy can be divided into two major categories, third-space (cerebrospinal, peritoneal, pleural, and pericardial fluid) and intra-arterial treatments. Interest in regional chemotherapy has risen recently for a variety of reasons, including evidence for significantly higher response rates compared to systemic therapies. The applicable pharmacokinetic principles for drug selection have been defined and indicate that regional drug exposure advantage is directly proportional to total-body drug clearance and is inversely proportional to rate of egress from a third-space or to regional blood flow. Reliable drug delivery to the entire region in question is important. The recently introduced totally implanted devices (pumps and ports) make regional therapy more reliable, safe, and convenient. In addition, nuclear medicine scanning techniques have proven useful in assessing drug distribution with regional delivery. Regional therapy may control regional tumor, but extraregional failure may ensue. However, when regional therapy is regionally selective in its toxic effects, systemic therapy may be combined in full doses. Future efforts will need to focus on the development of more potent regional therapies, alone and in combination with systemic therapies, as well as on the validation of such treatments through controlled clinical trials.

Subcutaneous infusion ports (SIP) were inserted for chronic venous access during 329 procedures in 300 patients over the past five years at the University of Michigan Medical Center, with a total follow-up experience of 318 patient years. Seventy-four per cent of the SIP were surgically implanted while patients were hospitalized. The SIP were used for chemotherapeutic agents (83.0 per cent), blood products (29.0 per cent) or hyperosmolar total parenteral nutrition (8.5 per cent) and accessed a median of three occasions. Eighty-four per cent were used in an outpatient setting at least part of the time. Thirty-nine per cent of SIP were associated with complications, including local infection or sepsis (16.4 per cent), thrombosis of the catheter or central vein (9.7 per cent) and extravasation from the port secondary to needle dislodgement (6.4 per cent). The risk of complication was slightly higher in those SIP first used ten to 14 days after placement as compared with those used earlier or later (p less than 0.05). In 23 of 32 episodes, clinically diagnosed local infection unassociated with systemic sepsis or skin necrosis was successfully treated without removal of the port using aggressive intravenous and oral antibiotics. Treatment of thrombosis of the catheter with either urokinase or streptokinase infusion was successful in ten of 15 attempts. Seventy-two (22 per cent) of SIP were eventually removed, either after completion of the chemotherapy (20) or because of a complication (52) with 29 SIP being replaced. There was no correlation between the risk of infection or thrombosis and the perioperative use of antibiotics, frequency of SIP use or preoperative white blood cell count, platelet count, coagulation profile, blood urea nitrogen or albumen concentration. SIP provide an excellent method of chronic venous access, having a lower rate of infection and thrombosis in historical comparison with external vascular access devices.

Hepatic arterial infusion chemotherapy results in a higher concentration of drug delivered to the tumor with less systemic exposure than is possible with intravenous therapy. However, extrahepatic blood flow and/or shunting to the lung can impose a limitation. This study describes a quantitative method for calculating the extrahepatic component and monitoring its changes due to a new adjunctive therapy, degradable starch microspheres (DSM). DSM temporarily occlude the hepatic arterial circulation, thereby increasing the uptake of therapeutic drugs. Twenty patients with metastatic liver cancer underwent hepatic arterial perfusion scintigraphy (HAPS) using Tc-99m MAA to determine blood-flow distribution and to quantitate extrahepatic uptake. The percent shunt index (PSI) was determined at baseline and after each incremental dose of DSM. The baseline PSI ranged from 6-26% (mean 12.3 +/- 5.8 s.d.) and changed progressively after each injection of DSM/Tc-MAA suspension. The patterns of change in shunting are described. Quantitative HAPS provides a means of measuring the extrahepatic component, warns of potential side effects, and helps guide chemotherapeutic decisions.

Intraarterial administration of 40-microns degradable starch microspheres (DSM) in a drug solution can temporarily retard flow of the drug-blood column through the arteriolar-capillary bed and lead to increased local drug deposition. Premonitory to Phase II-III efficacy studies applying this concept to regional therapy, it was necessary to determine the DSM dose to use. Patients with hepatic cancers were treated with varying doses of DSM with mitomycin C coadministered into the hepatic artery to define a dose of DSM which produces acceptable toxicity with maximal hepatic drug deposition as determined by a reduction in systemic mitomycin C exposure. Comparison of six patients receiving 6 ml of DSM (6 X 10(6) particles/ml) with ten patients receiving 15 ml showed a lower incidence and decreased severity of acute toxicity in terms of nausea/vomiting (16% versus 50%) and right upper quadrant hepatic pain (none versus 40%) with 6 ml of DSM. Reduction in systemic mitomycin C exposure evaluated by decrements in the area under the concentration curve in peripheral blood with time due to DSM was similar in both groups. Another seven patients were treated with escalating doses of DSM concurrently with 5 mg of mitomycin C. Although all seven patients tolerated 6 ml of DSM, higher doses (9 ml, 12 ml, 15 ml) led to incremental patient drop-out due to severe, acute right upper quadrant pain with only two patients able to receive 15 ml of DSM. In these patients, 6 ml of DSM appeared nearly equivalent to higher doses in terms of systemic exposure to mitomycin C. Eleven additional patients were evaluated for tolerance to repeated 6-ml dosing of DSM. Four patients had epigastric pain correlating with flow to the stomach demonstrated by nuclide angiography. The seven patients with no pain and no flow to stomach were treated with good tolerance for three-plus courses. Thus, 6 ml of DSM appear to be appropriate for Phase II-III studies.

We have conducted a Phase 1 study of aminopterin (AMT) with leucovorin (LV) in 17 patients. AMT was administered by bolus injection every 7 to 14 days in dosages from 25 to 425 mg/sq m. LV rescue was instituted at 24 hr and continued for 48 to 72 hr. At dosages above 50 mg/sq m, we observed nephrotoxicity defined as greater than or equal to a 25% increase in serum creatinine 24 hr after AMT administration, but its incidence was not strictly dose related. Urinary alkalinization and volume expansion appeared to reduce the incidence of nephrotoxicity. Nephrotoxic drug courses were associated with 24-hr plasma AMT levels [3.6 +/- 2.0 (S.D.) X 10(-6) M] which were significantly higher than nonnephrotoxic courses (1.6 +/- 1.0 x 10(-6) M) (p less than 0.05). In nonnephrotoxic courses, serum elimination pharmacokinetics appeared to be biphasic with a t1/2 alpha of 1.08 +/- 0.01 hr and t1/2 beta of 12.31 +/- 0.06 hr. Systemic toxicity (myelosuppression and mucositis) could be prevented in patients with impaired AMT clearance by the administration of LV at an increased dose rate. In several courses, systemic toxicity occurred in spite of apparently normal plasma clearance, suggesting that 24-hr plasma levels may not accurately reflect intracellular drug effects. Cytokinetic studies on bone marrow aspirates allowed determination of the rescue effect of LV and may prove useful in predicting marrow protection.

Forty-two patients with metastatic squamous cell carcinoma were treated with bleomycin, vincristine, and mitomycin C with or without methotrexate (BOM +/- M). The overall response rate of 64% (complete response [CR] rate, 19%) included 19 responses among 26 patients (seven CRs) with head and neck cancer, three responses among eight patients with cervical cancer, and three responses among five patients (one CR) with lung cancer. Six of 12 patients (two CRs) responded to BOM and 21 of 30 patients (six CRs) responded to BOMM. The median duration of response was 16 weeks. Toxic effects included nausea or vomiting in 33% of the patients, fever of101 degrees C in 26%, stomatitis in 29% and pulmonary toxicity in 19%. Four of eight cases of pulmonary toxicity were fatal and the incidence was related to the amount of both bleomycin and mitomycin C administered. The occurrence of pulmonary toxicity could not be predicted by serial determination of pulmonary function or blood gases. A wbc count nadir of2500/mm3 occurred in 15 of 42 patients. There were two episodes of sepsis with one death. A platelet count nadir of75,000/mm3 occurred in eight of 42 patients with no episodes of hemorrhage. BOMM produces a high objective response rate in patients with squamous cell cancer. However, the duration of remission is brief, and use of the regimen carries an increased risk of fatal pulmonary toxicity.

Diaziquone (AZQ) is a lypophilic alkylating agent that crosses the blood-brain barrier and has shown broad activity in animal tumor models. Five of 12 patients with malignant astrocytoma treated with iv AZQ had clinical and/or radiographic improvement (Schold, Neurology 34:615, 1984). Intra-arterial administration of AZQ to patients with brain tumors should produce higher peak levels of drug in the tumor and should reduce systemic toxicity. Twenty-one patients with astrocytoma (grade II, four; grade III, 11; and grade IV, six), in all of whom irradiation and intra-arterial carmustine chemotherapy failed, received intra-arterial AZQ as a single dose every 28 days. Two of 20 evaluable patients experienced partial responses of 5 and 8+ months, respectively. Four patients had disease stabilization of 3, 4, 5, and 8 months' duration, respectively, and one of these patients had tumor shrinkage (partial response) after seven courses of AZQ. The initial dose in the first three patients was 10 mg/m2, and doses in subsequent groups of three patients were begun at increases of 5 mg/m2. The within-group dose escalation was 5 mg/m2 per course if there was no hematologic toxicity. Dose-limiting toxicity was myelosuppression, which occurred at doses greater than 15 mg/m2. The maximum tolerated dose was 25 mg/m2. Intra-arterial AZQ appears to be of marginal effectiveness in patients refractory to carmustine and offers no advantage over iv AZQ in efficacy or toxicity.

The techniques now exist to deliver drug infusions reliably through the hepatic artery to infuse the entire liver and tumor within the liver. Drug selection for use in these systems should be rational and include agents with short half-lives (high total body clearance) and some evidence of activity against the tumor type in question. Hepatic extraction and metabolism of the drug will in turn decrease systemic exposure or allow more drug to be given per set amount of systemic exposure. The dose-limiting toxicity of appropriate drug programs may well be regional and not systemic. Further pharmacokinetic studies and controlled clinical trials are needed to evaluate existing regimens and to design new regimens. Future advances are likely to involve the combination of effective drugs and the use of therapeutic microspheres to improve selectivity based on tumor microcirculation.

In 17 patients receiving regional hepatic chemotherapy, Tc-99m macroaggregated albumin imaging was used to aid arterial catheter placement and to assess perfusion patterns. Intraoperative imaging with a portable gamma camera allowed immediate monitoring of hepatic and extrahepatic perfusion patterns and assisted catheter manipulation when necessary to achieve optimal flow distribution. In all 12 patients with standard hepatic arterial anatomy, complete perfusion of both lobes of the liver was achieved, although three of them required intraoperative catheter manipulation and repeat imaging after initial placement. The remaining five patients had aberrant hepatic arterial anatomy, and complete perfusion was more difficult to achieve; they exemplified the need for dual catheters, ligation of accessory hepatic branches, and repeated imaging.

Hepatic arterial catheters were placed for therapy in 8 patients with primary or metastatic liver cancer. Temporary hepatic venous catheters allowed direct sampling of blood for hepatic venous drug concentrations. Patients were administered from three to six infusions at rates of 10, 30, 90, 135, 180, 210, and 270 mg/kg/day (0.053 to 1.43 microM/kg/min), given over 2 h, of 5-fluorouracil (FUra). In Method 1, FUra was infused i.v., and FUra was measured in plasma from hepatic arterial and hepatic venous blood. In Method 2, FUra was given i.v. at one time and infused into hepatic arterial blood at another time, and FUra was measured in plasma from peripheral blood at the same site in both cases. Steady-state FUra plasma concentrations were measured by a sensitive and specific high-performance liquid chromatography method. Data were computer fitted to the equations appropriate for a physiological two-compartment flow model with Michaelis-Menten elimination from the peripheral compartment and blood flow rate, Q, between the central and peripheral compartment. Methods 1 and 2 gave mean Vmax and Km values which did not differ significantly; the overall mean Vmax was 2.02 microM/kg/min, and the overall mean Km was 10.9 microM. For Method 1 the mean Q1 value was 0.0803 liters/(kg X min) or 5.26 liters/min, which is the same as cardiac output, but for Method 2 the mean Q2 value was higher, namely 0.189 liters/(kg X min) or 13.0 liters/min. Steady-state systemic and intrinsic clearances and extraction ratios decreased progressively as the dose rate increased. Intra- and inter-subject variation of both Vmax and Km were of the same order of magnitude. As a result, dose rate escalation should be conservative for dose rates above 135 mg/kg/day. The results support hepatic arterial infusion as a means of improving the therapeutic index of FUra in the treatment of cancer of the liver.

The four-drug combination of bleomycin, vincristine, mitomycin C and methotrexate produces a high response rate in patients with squamous cell carcinoma. In this study we have examined the cytokinetic effects of this drug combination in vitro and in human bone marrow in vivo. The in vitro analysis revealed that mitomycin C produces a concentration dependent slowing of S-phase transit with partial G2/M and G1/S blocks in cell cycle progression. A partially synchronized S-wave occurs 4-8 and 16-20 hours following a two-hour drug exposure. Bleomycin produces a dose dependent G2/M block during a 14-hour drug exposure. Drug removal did not result in appreciable cell cycle synchrony. In vitro exposure to the two drug combination resulted in loss of both the marked G2/M accumulation seen with bleomycin, and the partially synchronized S-phase waves seen following mitomycin C exposure. The sequential changes in bone marrow cytokinetics were determined in eight patients during and following administration of vincristine, mitomycin C and a continuous four-day infusion of bleomycin. The major cytokinetic changes observed were an increase in G2/M phase cells 18 hours following vincristine, and an increase in thymidine incorporation and S-phase cells 24 and 48 hours following the end of the bleomycin infusion. The clinical course of 24 patients was reviewed. Eleven patients who received methotrexate 36 to 42 hours following bleomycin had a subsequent median leukocyte nadir of 1.8 x 10(3)/mm3. Thirteen patients receiving methotrexate 60 to 72 hours after bleomycin had median leukocyte nadir of 3.9 x 10(3)/mm3. The objective response rates in the two groups was 75% and 80%, respectively. This study demonstrates that bone marrow cytokinetic analysis may allow schedule modification to avoid myelosuppression without loss of therapeutic activity.

Thymidine rescue was administered following 63 courses of high-dose methotrexate in 20 patients. In the first part of this study, the methotrexate was given as a 24-hr infusion and the dose was escalated from 0.14 to 8.54 g/sq m; in the second part, methotrexate was infused to maintain a serum concentration of 15 micrometer for 30, 36, or 40 hr. Thymidine rescue was started immediately after the end of the methotrexate infusions, and consisted of 8 g/sq m/day for 3 days or until serum methotrexate was below a toxic level. Mucositis and myelosuppression were the major toxicities. Neither was dose related. Serum methotrexate levels were proportional to the logarithm of the methotrexate dose. There was a mean 6-fold increase in thymidine concentration during rescue. However, thymidine levels prior to and during rescue were not related to the incidence of subsequent toxicity. Recovery of DNA synthesis in bone marrow cells was evident by nucleoside precursor incorporation at 24 hr after the start of rescue. Two of 16 evaluable patients achieved partial responses. This study indicates that thymidine is an effective rescue agent for high-dose methotrexate in humans.

We treated six patients with unilateral malignant astrocytomas with pulse doses of intra-arterial BCNU (200 mg/m2 initially, with escalating doses every 6-8 weeks) via transfemoral selective internal carotid catheterization. Four patients had had partial resections without prior radiation therapy and received no steroids. They had decreased tumor stain and surrounding edema after two to four cycles of chemotherapy. For one patient, results of the neurologic examination returned to normal, with total disappearance of her tumor as assessed by computerized tomographic scan. Objective tumor response continued for 7 months in three patients and for 3 months in one. In two of the four patients, chemotherapy was discontinued because of retinal toxicity and not because of treatment failure. Two patients had had partial resections and radiation therapy. One patient had stable disease for 4 months, and the second had progressive disease with gradual visual loss beginning in the infused eye 3 weeks after the second treatment. The catheterization procedure is safe; it was without immediate complication in 17 BCNU infusions. In summary, high-dose intra-arterial BCNU is well-tolerated and is an effective initial chemotherapeutic modality.

Twenty-seven patients with advanced solid tumors (ten with lung cancer, eight with sarcoma, five with melanoma, four with miscellaneous tumors) were treated with high-dose doxorubicin (120 mg/m2) every 3 weeks for a total of 75 courses. As expected, marked myelosuppression was observed, with all patients having a granulocyte count nadir of less than 500 cells/mm3. The median platelet count nadir was 91 X 10(3)/mm3 following the first cycle and 50 X 10(3)/mm3 following the fourth cycle. Moderate or severe stomatitis was seen with 65% of the courses, and 46% of the courses were complicated by fever greater than 101 degrees F. Congestive heart failure was observed in only one patient after five cycles (600 mg/m2) of high-dose therapy. This patient has received 540 mg/m2 of doxorubicin 2 years previously. Radionuclide ventriculography (MUGA) performed serially in 12 patients (eight of whom received 480 mg/m2) and clinical evaluation did not suggest an increased risk of cardiotoxicity at this dose rate. Overall, 11 of 24 (46%) evaluable patients responded (58% if patients with malignant melanoma are excluded). Responses included one complete and three partial responses in eight evaluable patients with lung cancer, one complete and two partial responses in seven evaluable patients with sarcoma, and no objective responses in five patients with malignant melanoma. With appropriate supportive care, repetitive courses of doxorubicin at a dose of 120 mg/m2 can be given to patients of good performance status without a major increase in cardiotoxicity. However, the low complete remission rate, considering the observed toxicity, does not justify routine use of this regimen in patients with solid tumors.

Continuous i.v. thymidine (TdR) was given to 12 patients with metastatic cancer in an attempt to prevent methotrexate (MTX) toxicity. MTX was infused in 27 courses with progressive dose increase from 80 mg/sq m for 24 hr to 6 g/sq m for 72 hr. TdR at 8 g/sq m/day was infused concurrently and continued 24 to 48 hr beyond MTX infusion. The median pretreatment serum TdR level was 0.19 micron. With TdR infusion, the median level was 1.5 micronM. Serum TdR fell with a half-time of 8 to 10 min after a pulse dose or cessation of infusion. Spinal fluid TdR equaled serum TdR levels after 2 hr of infusion. Less than 2% of administered TdR appeared in urine. MTX serum levels were proportional to dose infused, ranging from 80 to 100 micronM with 2 g/sq m/day. The half-time for MTX clearance from serum was 4 to 8 hr. Spinal fluid MTX reached equilibrium at 3 to 12% of serum levels by 4 hr. Bone marrow dysfunction during MTX infusion was prevented by TdR as determined by labeling indices and cytofluorographic analyses. Toxicity was not seen in patients with normal MTX clearance using 48-hr infusions of MTX where TdR was continued for an additional 48 hr after the MTX infusion had ended. However, 3 of 6 courses of MTX at 6 g/sq m over 72 hr led to toxicity. Toxicity was reversible in 2 patients, 1 of whom was retreated with a similar dose duration of MTX without toxicity when TdR was continued beyond the end of the MTX infusion for 48 hr instead of the usual 24 hr. The 3rd patient with toxicity died of progressive disease and thrombocytopenia 19 days after treatment. No TdR-related toxicity or unusual MTX toxicity was detected. Antitumor effects were noted in 4 patients. TdR offers significant protection against MTX toxicity and deserves further clinical study.