Phase I Study of ⁹⁰Y-labeled B72.3 Intraperitoneal Administration in Patients with Ovarian Cancer: Effect of Dose and EDTA Coadministration on Pharmacokinetics and Toxicity¹ Free

Radiolabeled monoclonal antibodies have been used in clinical trials to image tumors by external scintigraphy (1, 2, 3) and for radioimmunotherapy (4, 5, 6). Clinical studies have clearly demonstrated the ability of murine monoclonal antibodies to effectively localize within various solid tumors in sufficiently high relative concentrations making them useful for radioimmunoimaging. Previous studies of radiolabeled monoclonal antibodies administered by the i.p. route have demonstrated relatively high local concentrations of the agent with slow absorption into the blood compared with antibodies administered by the i.v. route. This has resulted in superior localization in peritoneal implants and tumor tissues (7, 8, 9). An early Phase I/II trial of i.p. ¹³¹I-labeled HMFG antibody in ovarian cancer (10) as well as several more recent studies of ⁹⁰Y-, ¹⁸⁶Re-, and ¹⁷⁷Lu-labeled antibodies have demonstrated antitumor activity and prolonged survival under conditions of minimal tumor burden (11, 12, 13, 14).

Monoclonal antibody B72.3 (15, 16) has been shown to be immunoreactive with the glycoprotein complex TAG-72,³ a high-molecular-weight antigen (M_r >200,000) with the characteristic of a mucin (17). TAG-72 expression has been described previously in approximately 50% of primary breast carcinomas, in 85% of primary colon carcinomas, and in the majority of ovarian carcinomas tested (18). In normal adult tissues, the antigen is either not expressed or expressed in only trace amounts.

Antibody B72.3 has been used by several investigators for immunohistochemical staging of ovarian carcinoma, and radiolabeled B72.3 has been used for external scintigraphy of various human tumors (19, 20, 21, 22). Using a novel coupling chemistry allows B72.3 to be radiolabeled with ¹¹¹In for external imaging (23, 24) and with ⁹⁰Y for clinical radioimmunotherapy studies. This report describes our Phase I/II studies of the clinical pharmacology, toxicity, and tumor responses of escalating doses of ⁹⁰Y-labeled B72.3 administered i.p. to patients with epithelial ovarian carcinoma. In addition, this study examined the ability of EDTA coadministration to suppress the bone distribution of the ⁹⁰Y label and, thus, prevent myelosuppression, thereby allowing administration of higher doses of the radiolabel.

Yttrium-90 chloride was obtained from either the Oak Ridge National Laboratory (Oak Ridge, TN) or General Atomics (San Diego, CA). Ready Safe Scintillation cocktail was purchased from Beckman Instruments (Fullerton, CA). Protosol Tissue Solubilizer was purchased from NEN/DuPont (Boston, MA). The TSK 3000 SW column (7.5 mm × 60 cm) was purchased from Varian Associates (Sunnyvale, CA). Clinical grade B72.3-GYK-DTPA monoclonal antibody-chelate was provided free of charge by Cytogen, Inc. (Princeton, NJ)

After being buffered with sodium acetate, the ⁹⁰Y was site-specifically attached to the B72.3 through chelation to the linker-chelation complex, GYK-DTPA, described previously by Rodwell et al. (24). After a 2-h incubation, the percentage of incorporation of ⁹⁰Y label into the antibody was determined by TLC and was required to be at least 90%. All of the samples were sterile-filtered and pyrogen-tested before injection.

Duplicate 100-μl aliquots of plasma, urine, or peritoneal fluid were added to scintillation vials (10-ml) containing 2 ml of Protosol. Tissue samples were weighed into scintillation vials, and 2 ml of Protosol were added. Duplicate 100-μl aliquots of ⁹⁰Y-labeled B72.3 standard were also added to scintillation vials containing Protosol. The samples were incubated at 65°C with intermittent shaking for 18 h. Bone samples were mechanically disaggregated after the incubation period to insure adequate counting of the sample. The samples were then cooled, and 10 ml of Ready Safe Cocktail (Beckman) were added, and the samples were counted in a Packard scintillation counter (model 1500) to determine ⁹⁰Y content. Prior analysis demonstrated that color quenching of ⁹⁰Y was minimal because of the energetic nature of the isotope; therefore, quench correction was not applied to these samples. Analysis of a sample of the ⁹⁰Y-labeled B72.3 infusate served as a monitor of total cpm administered to the patient and as a decay correction standard.

Patients over 21 years of age with a histologically proven diagnosis of refractory ovarian, fallopian tube, or peritoneal papillary serous carcinoma were eligible for this trial provided that they met the following criteria: (a) a performance status <3 on the Zubrod scale and a life expectancy greater than 2 months; (b) symptomatic ascites with at least 5% of cells staining positive for TAG-72; (c) measurable disease; (d) a neutrophil count of >1500 cells/μl, a platelet count of >150,000 cells/μl, a serum creatinine level of <2 mg/dl, and a serum bilirubin level of ≤2 mg/dl; (e) adequate peritoneal distribution; and (f) signed informed consent. All of the patients had to have recovered from any toxic effects from previous treatments. Exclusion criteria included: (a) prior radiotherapy or chemotherapy within the last 4 weeks before treatment; (b) prior treatment with pelvic or abdominal radiotherapy; (c) intestinal obstruction; (d) prior exposure to murine antigens with development of murine antibody; (e) prior treatment with corticosteroid within 4 weeks before study drug; (f) symptomatic brain metastases; (g) previous or current malignancies (except in situ carcinoma of the cervix uteri that underwent cone biopsy or adequately treated basal or squamous cell carcinoma of the skin); and (h) significant neuropathy, heart condition, or poor medical risk because of nonmalignant systemic disease.

Before entry, a complete medical history was recorded, and a physical examination was done. Performance status was noted and lesions measured. A complete blood cell count and relevant blood chemistries, including a CA-125, were analyzed. Chest X-ray and other indicated studies were done as needed and repeated every 8 weeks. Radiological placement of the peritoneal catheter was performed 24 h before drug administration. The peritoneal distribution was evaluated either by contrast radiography or by a nuclear medicine technique with ⁹⁹technetium-sulfur colloid just before ⁹⁰Y-labeled B72.3 administration. Before each course, a clinical examination and a blood chemistry survey were required.

Treatment consisted of escalating radioactive doses of ⁹⁰Yttrium chelated to 2 or 10 mg of B72.3 monoclonal antibody. The starting radioactive dose was 1 mCi, and the drug complex was administered i.p. over 5 min, followed by 1.5 liter of normal saline. Calcium disodium versenate (EDTA) administration started after the observation of toxicity with ⁹⁰Y-labeled B72.3 given as a single agent and consisted of a 12-h continuous i.v. infusion for six doses. EDTA was given at a dose of 25 mg/kg in 500 ml of 0.9% saline every 12 h for six doses. The infusion started 30 min before ⁹⁰Y-labeled B72.3 administration. After i.p. administration of the study drug, the patient’s position was shifted approximately every 4 h, and movement was encouraged. Patients could be treated twice no less than 8 weeks after the first dose if they benefitted from the first treatment and had no major side effects. The second dose was chosen according to the dose level in effect at the time of the second treatment.

DLT was defined as a grade ≥4 hematological or nonhematological toxicity observed in two of three patients. MTD was defined as the dose level that preceded the dose at which two of three patients experienced a DLT. Clinical responses were evaluated at 4 and 8 weeks of treatment. A complete response was the disappearance of all measurable disease for at least 4 weeks. Partial responses represented a 50% or greater decrease in the sum of the products of the perpendicular diameters of the measurable lesions also for a minimum of 4 weeks duration.

The major objective of this study was to determine the MTD of ⁹⁰Y-labeled B72.3 when this agent is administered i.p. using the schedule defined above. Further objectives were to evaluate the toxicity and efficacy of this novel agent. At any dose level, if none of three patients experienced DLT, another cohort of three patients was treated at the next higher level. If one of the three patients experienced DLT, an additional three patients were treated at the same dose level. If one of six patients experienced DLT at a level, the dose escalated for the next cohort of patients. If two or more of three or six patients experienced DLT, then another three patients were treated at the previous dose level. The MTD was then defined at the highest dose level in which six patients were treated with less than two instances of DLT. Because patient numbers with each dose level were small, no actual statistical analysis were performed on the data, but all of the results were examined for potential indications of toxicity relative to the dose level. Efficacy data were also tabulated. Pharmacokinetic and pharmacodynamic data were analyzed at each dose level. Plasma half-life, plasma:ascitic ratios, AUC, and renal clearance were calculated, and the retention of yttrium in bone and marrow was also measured.

Heparinized blood samples were withdrawn at 0 (infusion end), 5, 15, 30, 45, and 60 min and at 2, 4, 24, 48, 72, and 96 h after antibody administration. Peritoneal fluid samples (2-ml) were withdrawn at 5, 15, 30, 45, and 60 min and at 2, 4, 24, and 48 h after administration. The total urinary output was collected from 0–4 h, 4–12 h, 12–24 h, 24–48 h, and 48–72 h. After adjustment was made for isotopic decay, pharmacokinetic parameters were then calculated by conventional techniques (25) using a polyexponential curve-fitting program (PK Analyst, Micromath, Inc.).

Patient characteristics are described in Table 1. Sixty-one patients were registered, of whom 58 received 63 doses of ⁹⁰Y-labeled B72.3. Fifty-seven patients were evaluable for toxicity, and 57 were evaluable for response. Ages ranged from 33 to 74 years (median, 55 years). Zubrod performance status ranged from 0 to 2 (median, 0). All of the patients had TAG-72-positive tumors (>10% of tumor cells immunostained). Six patients received 1 mCi of ⁹⁰Y-labeled B72.3 before the initial debulking surgery, to assess the distribution of B72.3 into the tumor.

Except for the starting dose (1 mCi), patients were treated in cohorts of three with dosage escalations of 5 mCi increments (Table 2). Three patients treated at the 5-mCi dose level had no significant hematological (>grade 2) or other toxicities noted. At the 10-mCi level, two patients had no hematological toxicity (grade 0), and two patients developed a grade 4 thrombocytopenia. Platelet counts decreased on day 29 and 32 and lasted until day 59 and 39, respectively, after which these patients received another anticancer treatment. One of these two patients had already received a 5-mCi dose 6 weeks earlier, after which she developed a grade 2 thrombocytopenia. The other patient also presented on day 47 with a grade 3 neutropenia, which lasted 11 days. Six patients were treated at the 15-mCi level, and all of the patients developed a thrombocytopenia, but only 2 had a grade 3 and 4, respectively. Four of these patients were neutropenic (three grade 3, and one grade 4). The neutropenia onset was observed on day 40 (29–49 days), and lasted for up to 49 days in one patient (14–49 days) until full recovery. This dose was considered to be too high as a single agent, and 10 mCi was considered to be the MTD.

Patients with the same eligibility criteria underwent dosage escalation, again consisting of 5-mCi increments but starting at the 10-mCi dose level with i.v. infusion of EDTA (Table 2). The total number of patients treated in this arm of the study was 38 with a median age of 58 years, and each patient received only one dose. With EDTA, no major hematological toxicity was observed until a dose of 25 mCi was reached. At this level, of six patients treated, one had a grade 3 thrombocytopenia, and another patient experienced a grade 3 neutropenia. At 30 mCi, one patient of three had a grade 3 neutropenia. At 35 mCi, seven patients were treated. One grade 4 neutropenia was observed in one patient, as well as one grade 3 neutropenia, and thrombocytopenia was observed in two different patients. At 40 mCi, there were no grade 4 toxic effects, but one grade 3 thrombocytopenia and one grade 3 neutropenia were observed in two patients of six. At 45 mCi, of seven patients treated, one patient developed a grade 4 neutropenia, one developed a grade 3 neutropenia, and another patient developed a grade 3 neutropenia and thrombocytopenia. Anemia was never more severe than grade 2 when EDTA was used. The nadir for these cytopenias was usually delayed to day 40 for anemia, day 39 for thrombocytopenia, and day 50 for neutropenia. The duration of cytopenia was also prolonged, and from the onset of the side effect to full recovery, lasted for 18 days for anemia, more than 28 days for thrombocytopenia, and more than 17 days for neutropenia. The true duration to full recovery could not be evaluated in most patients because they needed further treatment for disease control.

Nonhematological toxicities were mild. No grade 3 events were observed. Within hours of administration, five patients had mild abdominal pain (two grade 1 and three grade 2) that was self-resolving. There were no cases of peritonitis or subsequent episodes of intestinal obstruction or abdominal symptomatology thought secondary to the radioisotope. Three patients had mild nausea (grade 2), and three patients experienced a grade 1 or 2 emesis, which were considered secondary to the compound. Three patients each had an episode of diarrhea or of constipation. Two patients complained of bone pain, and one patient of fatigue. No allergic reactions or hypertension were noted.

Of the 58 patients treated in this part of the trial, 2 objective complete responses were noted, occurring in masses smaller than 3 cm in diameter. One response was achieved after two courses of ⁹⁰Y-labeled B72.3 in one patient who received 15 mCi first without, and then with, EDTA. This response lasted for 1 year. The second complete response was observed in one patient treated at 30 mCi with EDTA and lasted for 9 months. Two minor responses were also noted in two patients treated at 25 mCi with EDTA (Table 3). Thirty patients had stable disease for a median of 6 months (range, 4–24 months).

As shown in Fig. 1,A, the concentration of ⁹⁰Y in peritoneal fluid remained relatively constant over the 4-h monitoring period for both the 1- and 15-mCi dose levels. On the other hand, concentrations of ⁹⁰Y in plasma over 96 h increased monophasically over time (Fig. 1,B and Table 4) and reached peak concentrations at between 31 and 46 h after i.p. administration. For the 1-, 10-, and 15-mCi dose levels, the absorbance kinetics fit a model of one compartment with first-order absorbance and first-order clearance kinetics. The half-lives for absorbance ranged from 12 to 24 h, and the clearance phase half-life of ⁹⁰Y from plasma ranged between 76 and 116 h. Although the numbers in each dose group were small for the 10- and 15-mCi dose levels, there seemed to be a trend toward decreasing clearance-phase half-lives, decreasing RTs, and a shorter time to maximal concentration of ⁹⁰Y in plasma with increasing dose of the radiolabel. Cumulative urinary excretion of the radiolabel over time (Fig. 2) demonstrated that patients at the 1-mCi dose level excreted approximately 20% of the injected dose of radiolabel over 72 h, whereas patients at the 10- and 15-mCi dose levels excreted almost 3-fold higher (approximately 60%) levels over the same period.

Analysis of bone and marrow biopsy specimens obtained from patients at 72 h after administration (Fig. 3) demonstrated that a significant proportion of the ⁹⁰Y radiolabel was present in bone ranging from 0.08 to 0.014%ID/g at the 1- and 15-mCi dose levels respectively. Marrow content of the ⁹⁰Y label also increased with increasing ⁹⁰Y dose and ranged from 0.001 to 0.005%ID/g at the 1- and 15-mCi dose levels respectively.

Although peritoneal fluid specimens were obtained at various times after administration and were analyzed for ⁹⁰Y content, there was insufficient sampling of peritoneal fluid contents over time to allow for adequate pharmacokinetic analysis.

A previous clinical study by Hird et al. (14) had suggested that the administration of EDTA could reduce the body burden of ⁹⁰Y after the administration of ⁹⁰Y-labeled HMFG1 by provoking increased urinary excretion of the label and could thereby prevent the radiation-induced myelosuppression observed with this reagent. Therefore, the study was modified to include the administration of EDTA before ⁹⁰Y-labeled B72.3 administration. Patients were treated at doses escalating from 15 to 45 mCi.

As shown in Fig. 4 A, the concentration of ⁹⁰Y in peritoneal fluid was monitored over 4 h in patients receiving doses of 15–35 mCi of ⁹⁰Y-labeled B72.3 antibody. Over time, the radioisotope concentration increased in general in all of the patients suggesting that the absorbance of the saline solution occurred somewhat faster than the absorption of the ⁹⁰Y label. Comparison of the peritoneal fluid ⁹⁰Y concentration in two patient groups, both receiving 15 mCi but either with or without EDTA pretreatment, was examined but we found no demonstrable difference in the peritoneal fluid ⁹⁰Y pharmacokinetic behavior between these two groups.

The pharmacokinetic behavior of the ⁹⁰Y label in plasma after i.p. administration is shown in Fig. 4,B and Table 1. The appearance and clearance of the radiolabel in plasma was found to closely fit a model of one compartment with first-order absorbance and first-order clearance kinetics. Absorbance of the radiolabel into plasma occurred with half-lives ranging between 12.4 and 51 h. The calculated absorbance half-lives seemed to decrease with increasing dose of radiolabel. Trends in the calculated clearance-phase half-lives, the RT, and the time to maximal concentration in plasma all appeared to decrease with the increasing dose of the radiolabel. There were insufficient numbers of patients to compare pharmacokinetic parameters of patients receiving 15 mCi with or without EDTA pretreatment to determine whether EDTA affected the pharmacokinetics of ⁹⁰Y-labeled B72.3. However, our data from these two groups suggest that these groups seem to be comparable in terms of plasma ⁹⁰Y content and pharmacokinetics.

The cumulative urinary excretion of the ⁹⁰Y label was also studied in these patient groups. As shown in Fig. 5, urinary excretion ranged between 40 and 60% of the total injected dose of ⁹⁰Y. There appeared to be no differences in the urinary excretion rates between dose groups. Comparison of the two patient groups at 15 mCi with and without EDTA pretreatment also revealed no significant differences between these two groups.

The bone and marrow content of ⁹⁰Y was also measured in patients receiving various doses of ⁹⁰Y-labeled B72.3 after EDTA pretreatment (Fig. 3). Content of the radiolabel in bone was highest at the 15-mCi dose (0.002%ID/g) and declined at higher doses to between 0.001 and 0.0005%ID/g. Content of the ⁹⁰Y label in marrow was highest at the 15-mCi dose (0.005%ID/g) and declined to levels approximating bone at higher dose levels (between 0.001 and 0.0005%ID/g). Calculation of the total ⁹⁰Y radiolabel in bone demonstrated that doses of up to 40 mCi resulted in a bone level of 0.1 μCi/g. By comparison, in patients without EDTA pretreatment, after administration of 15 mCi (MTD), the ⁹⁰Y levels in bone were up to 20-fold higher (2.0 μCi/g). Fig. 3 clearly demonstrates that EDTA pretreatment significantly reduced both the bone and the marrow content of ⁹⁰Y and seemed to be the mechanism by which EDTA administration confers protection of bone and marrow during administration of ⁹⁰Y-containing therapeutic agents.

The intracavity route of administration has been used for therapeutic intervention in a variety of locally spreading cancers (26, 27, 28, 29). Our study generally found that ⁹⁰Y-labeled B72.3 was a very well tolerated radioactive monoclonal antibody and well suited for i.p. administration in patients whose carcinoma expresses the TAG-72 antigen. We found that the major side effect was hematological, particularly prolonged thrombocytopenia and neutropenia. However, the anemia observed was mild and the granulocytopenia was moderate and dose-related. Using the antibody HMFG-1 radiolabeled with ⁹⁰Y-CITC-DTPA, Maraveyas et al. (30) studied 19 patients in an ascending dose Phase I study. Hematological toxicity consisting of Grade III granulocyte and platelet suppression was observed at doses of 19.3 mCi/m². This compared favorably to the current study, which found the MTD of ⁹⁰Y-labeled B72.3 to be approximately 15 mCi without EDTA pretreatment.

The use of EDTA in our study clearly demonstrated the myeloprotective effect of this agent by reduction of both bone and marrow content of ⁹⁰Y without appreciably affecting the pharmacokinetics of the intact ⁹⁰Y-antibody complex. In addition, the concomitant administration of EDTA allowed dosage escalation up to 45 mCi but at the expense of prolonged granulocytopenia. Our study demonstrated that the MTD of this radioimmunotherapeutic agent should be considered to be 40 mCi when concurrently administered with EDTA in a total dosage of 150 mg/kg given as a continuous administration over 72 h. Moderate thrombocytopenia and neutropenia is to be anticipated with otherwise good tolerance to this agent.

Although studies by Stewart et al. (31) used EDTA pretreatment in clinical trials of ⁹⁰Y-labeled HMFG1, there have been no other studies performed that have examined the effects of EDTA on plasma or peritoneal fluid pharmacokinetics in detail. Previous studies (14), however, did suggest that EDTA administration caused a significant increase in the cumulative urinary excretion (from 11 to 32.3% of the injected dose). Although our studies demonstrated a much higher cumulative urinary excretion of ⁹⁰Y in patients not treated with EDTA (∼60%), we were not able to demonstrate the effects of EDTA on urinary excretion. These studies are consistent with independent observations suggesting no significant effects of EDTA on plasma or peritoneal fluid content, absorbance, or clearance of ⁹⁰Y. Therefore, our studies clearly suggest that the myeloprotective effects of EDTA administration are primarily a result of a reduction in bone absorption of the ⁹⁰Y label leading to a reduced irradiation of the contained marrow and not an effect on the total body burden of ⁹⁰Y as had been previously suspected.

On the basis of our previous studies of ⁹⁰Y-B72.3 in patients undergoing laparotomy (32), radiation dose estimates suggested that the mean dose of radiation to tumors was approximately 83 RADS/mCi. These data agree well with studies by Hnatowich et al. (33) of ⁹⁰Y-labeled antibody OC-125 administered i.p. These studies found that tumor sites received approximately 50 RADS/mCi. On the basis of our data, the administration of the MTD dose of 45 mCi of ⁹⁰Y-labeled B72.3 with EDTA pretreatment should result in a mean dose to tumor sites of approximately 3,700 RADS. This projected dose is in the range of therapeutic doses administered by external beam irradiation (34).

Although the current study was designed principally as a dose escalation study, four patients with minimal disease were found to have objective responses lasting 12, 9, 7, and 1 months. We were unable to determine a correlation between TAG-72 expression in the responder patients compared with the nonresponders. Stewart et al. (31) noted tumor regression in 1 of 25 patients treated with i.p. ⁹⁰Y-HMFG. A second patient had palliation of ascites. In a continuing study by the same investigators (26), radiolabeled antibody was administered i.p. in 52 patients after conventional surgery and chemotherapy. Twenty-one patients had no evidence of residual disease. At a median follow up of 35 months, only two patients had died from relapse of their cancer, which suggests that patients may benefit from radiolabeled antibody treatment in an adjuvant setting.

Meredith et al. (35) treated 27 patients with CC49, a high-affinity antibody recognizing TAG labeled with Lutetium-177 (¹⁷⁷Lu). One patient with measurable disease had a partial response to treatment. Seven of nine patients with less than l-cm nodules progressed at 22 months, with the other two patients progressing at 4–5 months. Four patients with microscopic disease remained disease-free at 6, 22, 32, and 35 months. Studies of a ¹⁸⁶Re-labeled pancarcinoma antibody, NR-LU-10, have also demonstrated antitumor activity. Seventeen patients, who had persistent or recurrent disease after platinum-based chemotherapy, received 25–150 mCi/m² of ¹⁸⁶Re-NR-LU-10 i.p. In four of seven patients with tumors of less than 1 cm, decreases in tumor size were noted at repeat laparotomy (36, 37, 38).

Whereas there have been numerous attempts to improve radioimmunotherapy of ovarian cancer by administration into the peritoneum and by the use of various antibodies coupled to different radioisotopes, further strategies for improvement in this area may come from attempts to use agents to up-regulate the TAG-72 antigen on tumor cells (39, 40, 41, 42) or from the use of concurrently administered agents to sensitize tumor cells to the radiolabeled agent (43, 44). The dosimetry characteristics and safety profile of this targeted therapeutic agent seem to make the complex suitable for trial as a consolidation therapy for patients with suitable peritoneal distribution and residual disease after primary surgery and chemotherapy as well as for patients with lesions no greater than 0.5 cm in diameter. On the basis of the results of our trial, these data suggest that further therapy trials with this agent using the i.p. route of administration seem warranted.