Phase I Study of Cediranib in Combination with Oxaliplatin and Infusional 5-Fluorouracil in Patients with Advanced Colorectal Cancer

Colorectal carcinoma is the second leading cause of cancer death in North American. A majority of the ∼160,000 patients diagnosed with colorectal cancer in 2007 will succumb to this disease (1). In the past decade, there have been tremendous improvements in the treatment of advanced colorectal cancer with the introduction of several new agents. Combination chemotherapy with irinotecan, oxaliplatin, and fluorouracil produced response rates of 40% to 50% and median overall survival of ∼20 months, compared with 20% and 12 to 14 months with fluorouracil alone (2–4).

The ability to initiate and sustain angiogenesis is a hallmark of malignant transformation (5). Targeting tumor angiogenesis is therefore a rational strategy of improving the efficacy of anticancer therapy. Bevacizumab, a humanized monoclonal antibody against vascular endothelial growth factor, has been shown, in combination with chemotherapy, to improve outcomes in patients with metastatic colorectal cancer in both first line and second line settings (6–9).

Cediranib (RECENTIN; AZD2171) binds to the intracellular tyrosine kinase domain of the three vascular endothelial growth factor receptor with high affinity, thereby interfering with signal transduction of vascular endothelial growth factor binding to its receptor (10). It is orally bioavailable and has shown antitumor activity in a single-agent phase I study (11). Common side effects of cediranib include diarrhea, dysphonia, and hypertension. It is currently undergoing evaluation in several tumor types as a single agent or in combination with chemotherapy (12–18).

Oxaliplatin administered with infusional 5-fluorouracil (5-FU; FOLFOX) has become a standard treatment for metastatic colon cancer in both first line and second line settings. Common grade 3/4 side effects of FOLFOX regimens include neutropenia (40-50%) with 1% to 4% being febrile neutropenia, diarrhea (∼10%), and neuropathy (18-34%; ref. 19).

Purposes of this study were to determine the recommended phase II dose of cediranib in conjunction with standard doses of modified FOLFOX-6 (mFOLFOX-6), and the tolerability, safety, pharmacokinetics, and antitumor activity of this combination in patients with previously untreated advanced colorectal cancer.

Eligibility. Patients were required to have histologically or cytologically confirmed locally advanced or metastatic adenocarcinoma of colon or rectum with radiologically or clinically measurable disease; Eastern Cooperative Oncology Group performance status of ≤2; and adequate hematologic, renal, and hepatic function. Prior chemotherapy was not allowed except for neoadjuvant or adjuvant therapy, which must have been completed at least 6 mo before study registration. Prior therapy with antiangiogenic agents was not permitted.

The study was designed and conducted by the National Cancer Institute of Canada Clinical Trials Group. It was approved by the research ethics board at both participating institutions. Patients gave written informed consent before study registration.

Study design and treatment. The mFOLFOX-6 consisted of oxaliplatin 85 mg/m² i.v. over 2 h concurrent with leucovorin 400 mg/m², followed by 5-FU 400 mg/m² as an i.v. bolus followed by 5-FU 2,400 mg/m² over 46 h. Each treatment cycle was 2 wk.

Cediranib was given orally continuously starting on day 2 of the first cycle. Two dose levels were planned: 30 and 45 mg. Three patients were entered initially at each dose level. If one patient experienced a dose-limiting toxicity (DLT), then, at least three more patients were enrolled at the same dose level. For dose escalation to 45 mg, no >2/6 patients must have experienced a DLT. Dose de-escalation to 20 mg was planned if required. If >2/6 patients experienced DLT at a dose level, that dose level would be considered to be higher than the maximum tolerated dose and the next lower dose level would be the recommended phase II dose for cediranib with mFOLFOX-6.

DLT was defined as grade 3 or higher toxicity related to cediranib (including an increase in severity or frequency of expected toxicity of chemotherapy) in cycle 1 and included proteinuria, hematuria, elevated creatinine, fatigue, hand-foot syndrome, uncontrolled hypertension, other toxicities of concern to investigators, and delay of cycle 2 chemotherapy for >14 d.

Assessments. Baseline evaluations included a complete medical history, physical examination, and laboratory evaluations including complete blood count with differential, renal, metabolic, and hepatic functions. Toxicities were assessed and graded every cycle according to Common Terminology Criteria for Adverse Events, version 3.0. Blood pressure was measured weekly for the first 2 cycles, then every cycle. Hypertension was managed according to a standard algorithm with a calcium channel antagonist as the first choice of therapy (12). Response and progression were evaluated using Response Evaluation Criteria in Solid Tumors every 8 wk (20).

Oxaliplatin, 5-FU, and leucovorin doses were adjusted for related toxicities according to manufacturers' guidelines, and no dose re-escalation was permitted. Patients who discontinued chemotherapy could continue cediranib in the absence of disease progression or unacceptable toxicity. Therapy with cediranib was held for grade 2 to 3 toxicity until resolution to grade 1 and then restarted at either the same dose (for grade 2 toxicity) or at 1 dose level lower (for grade 3 toxicity). After any renal toxicity of grade 2 or 3, weekly 24-h urine collections for creatinine clearance and protein were obtained. Patients with grade 4 toxicity, a treatment pause or chemotherapy delay of >2 wk, or >2 dose reductions of cediranib were discontinued from protocol therapy.

Pharmacokinetic evaluations. For oxaliplatin, blood samples were collected from a peripheral vein into heparinized tubes at the end of 2-h oxaliplatin infusion, and 0.25, 0.5, 0.75, 1, 3, 5, 6, and 24 h after completion of oxaliplatin infusion on cycles 1 and 2. Free intact oxaliplatin in plasma ultrafiltrate was measured by a high performance liquid chromatography tandem mass spectrometry method. For tandem mass spectrometry analysis, m/z transitions of 397.1 → 305.0 for oxaliplatin and 371.1 → 247.1 for the internal standard, carboplatin, were monitored. The lower limit of quantitation for oxaliplatin was 10 ng/mL.

For 5-FU, blood samples were drawn at 0 and 1 h after 5-FU infusion. Plasma 5-FU concentrations were measured using a high performance liquid chromatography–UV method (21). For cediranib, blood samples were drawn on day 1 of cycle 2 before drug administration, and at 0.5, 1, 2, 4, 6, 8, and 24 h postcediranib administration. Plasma cediranib concentrations were measured as previously described (11). Pharmacokinetic variables were calculated using noncompartmental methods (WinNonLin; Pharsight).

Sixteen patients were enrolled. Patient demographics are summarized in Table 1. A total of 150 cycles of therapy were administered (median, 6; range, 1-20). All patients were evaluable for toxicity and 14 patients were evaluable for response.

Dose escalation, toxicity, and dose intensity. No protocol-defined DLTs were seen in the first 3 patients enrolled at the 30-mg dose level. However, overall incidences of grade 1 to 2 toxicities seemed to be higher compared with past experiences with mFOLFOX-6. Three additional patients were therefore enrolled at this dose level. Again, no protocol defined DLTs in cycle 1 were encountered; however, all three patients missed some cediranib doses in cycle 2 due to adverse events. Another three patients were enrolled, and one DLT was observed (grade 3 diarrhea). Therefore, cediranib dose was cautiously escalated to 45 mg to further explore dose and toxicity relationships. This decision was also based on the observation of dose-related antitumor effects in the single-agent phase I study (11). No DLTs were observed in the first 3 patients; however, 1 patient was not able to get week 2 of cediranib due to mFOLFOX-6–related toxicities. An additional four patients were enrolled at this dose level, and no DLTs were observed (Table 2).

No grade 4 cediranib-related toxicity was observed in this study. Common grade 3 cediranib-related toxicities included hypertension, diarrhea, fatigue, and anorexia (Table 3). Grade 3/4 hematologic toxicities included neutropenia in 6 patients with 2 being grade 4. During the course of therapy, the majority of patients required dose modification of either cediranib and/or mFOLFOX-6. At the 30-mg dose level, 4 of 9 patients required cediranib dose reduction, whereas 4 of 7 patients required cediranib dose reduction at the 45-mg dose level. Oxaliplatin and 5-FU dose reductions were required for 4 of 9 patients at the 30-mg dose level and 2 of 7 patients at the 45-mg dose level. Unplanned treatment delays occurred in 6 patients for a total of 7 cycles at the 30-mg dose level, and in 4 patients for a total of 12 cycles at the 45-mg dose level. At the 45-mg dose level, only 50% patients received >90% planned cediranib from cycle 2 on. In addition, >50% patients required dose reduction of mFOLFOX-6 beyond cycle 1 (Table 4). At the 30-mg dose level, 66.7% and 77.8% patients received >90% planned oxaliplatin and 5-FU, respectively. Therefore, cediranib at 30 mg daily continuously was considered to be more tolerable and was declared the recommended phase II dose in combination with mFOLFOX-6 for unselected patients in broad phase II and III programs.

Hypertension was seen in 11 of 16 (68.8%) patients, with 6 (37.5%) being grade 3. These patients were easily managed with standard antihypertensive medications. Nose bleeding was the most common bleeding side effect seen and it occurred in 7 of 16 patients (43.8%). All of these episodes were grade 1 and resolved spontaneously. Three patients developed grade 1 proteinuria and 1 patient developed grade 2 proteinuria on study. Thyroid functions were monitored in this study, and thyroid-stimulating hormone levels were elevated in 6 patients. Although these patients were asymptomatic, they were treated with thyroid hormone replacement therapy.

During the course of the study, five patients required hospital admission for adverse events. One patient developed pneumothorax as a complication of central line insertion, and one patient required hospitalization due to salmonella infection while vacationing out of the country. These were not considered to be related to cediranib. One patient developed central line thrombosis resulting in superior vena cava obstruction, and two patients required hospital admission for various symptoms, including grade 3 hypertension and grade 3 anorexia. These were considered to be at least partially related to cediranib. There were no deaths on study.

Pharmacokinetics. Twelve of 16 patients were evaluable for cediranib pharmacokinetics. Cediranib is rapidly absorbed after oral administration achieving the maximal concentration in 3 to 4 hours. There was marked interpatient variability in C_ss,min, C_ss,max, and AUC_ss, with the coefficient of variation ranging from 67.5% to 120.8% (Table 5). Because of this large interpatient variability and the small number of patients enrolled at each dose level, there were no dose-proportional increases in pharmacokinetic variables between 30- and 45-mg dose levels.

Steady-state 5-FU concentrations were available for 8 patients at the 30-mg dose level and 6 patients at the 45-mg dose level. The average 5-FU concentration was 1,645 ± 1,590 ng/mL for cycle 2 versus 745 ± 391 ng/mL for cycle 1 at 30-mg dose level, and 999 ± 512 ng/mL for cycle 2 versus 682 ± 269 ng/mL at 45-mg dose level. Administration of cediranib did not affect steady-state 5-FU concentrations.

Free intact oxaliplatin concentrations in plasma ultrafiltrates were measured in 5 patients with and without cediranib (Fig. 1). Free intact oxaliplatin concentrations were not measured in other patients due to improper sample processing. Free intact oxaliplatin was eliminated rapidly from plasma with a mean half-life of ∼15 minutes. As a result, free intact oxaliplatin concentrations decrease below the limit of quantitation of 10 ng/mL 1 hour after the completion of oxaliplatin infusion. Similar to 5-FU, administration of cediranib did not affect free intact oxaliplatin pharmacokinetics.

Antitumor activity. Fourteen patients were evaluable for tumor response, one patient did not have measurable disease based on independent radiology review and another patient had only one cycle of therapy. There were 6 confirmed partial responses (response rate, 42.9%; 95% confidence interval, 17.7-71.1%), with duration of responses ranging from 3.7 to >13 months. Six patients had stable disease as the best response with a median duration of 4.8 months (range, 2.7-9.2 months). The median progression-free survival for all 16 patients enrolled was 9.3 months (95% confidence interval, 4.1 months-not reached yet). Two patients underwent resection for initially nonresectable hepatic metastasis after 20 and 16 cycles of therapy, respectively. Cediranib was discontinued 6 and 8 weeks before surgical resection. The postoperative course was complicated by hyperbilirubinemia in one patient, which resolved spontaneously. There were no wound-related complications in either patient.

Two doses of cediranib, 30 and 45 mg, were studied in combination with mFOLFOX-6. No grade 4 cediranib-related toxicity was observed at either dose level, and common grade 3 cediranib-related toxicities included hypertension, diarrhea, fatigue, and anorexia, similar to those reported in previous studies. Although no DLTs were observed at the 45-mg dose level, cediranib dose reduction and delay were required in 50% of patients. In addition, >50% patients required dose reduction or delays of mFOLFOX-6 beyond cycle 1. Although there were no statistically significant difference, the actual dose intensity for mFOLFOX-6 seemed to be higher at the 30-mg dose level, with 66.7% and 77.8% patients receiving >90% planned oxaliplatin and 5-FU, respectively. In comparison, Tournigand et al. (4) reported that the oxaliplatin relative dose density was 84.7% in previously untreated patients and 90.1% in patients previously treated with FOLFIRI. Therefore, cediranib at 30 mg orally continuously was considered to be more tolerable and was declared to be the recommended phase II dose in combination with mFOLFOX-6.

In the initial cediranib single-agent phase I study, doses of 45 mg or less seemed to be well-tolerated (11). In several subsequent phase II single-agent studies, cediranib starting dose had to be reduced from 45 to 30 mg because of toxicity (15–17). Furthermore, cediranib at 45 mg has been shown to be associated with higher incidences of toxicities in combination with regular doses of a number of chemotherapeutic agents, such as, pemetrexed at 500 mg/m² every 3 weeks (18), irinotecan at 300 mg/m² every 3 weeks (18), cisplatin/gemcitabine (13), or carboplatin/paclitaxel (12). In a study of cediranib and mFOLFOX-6 in patients with advanced solid malignancies, 2 DLTs were observed in 8 patients at the 30-mg dose level, and cediranib dose was de-escalated to 20 mg (18). In a large randomized phase II study, Cunningham et al. (22) evaluated cediranib at 20 mg daily or 30 mg daily in combination with mFOLFOX6 with better safety results in the 20-mg group. Therefore, cediranib at 30 mg or lower is more likely to be tolerated in combination with other antitumor agents.

Like other agents targeting angiogenesis, hypertension was a common side effect seen in this study with 11 of 16 patients developing hypertension (23). However, it was easily managed with calcium channel antagonists. Only grade 1/2 proteinuria was seen. Although no patients developed clinically significant hypothyroidism, 6 patients developed elevated thyroid-stimulating hormone and required thyroid hormone replacement.

The exposure to cediranib in this study seemed to be higher than those reported in the single-agent phase I study (11). For example, C_ss,max and AUC_ss were 131 ± 104 ng/mL and 1,732 ± 1,169 ng × h/mL, respectively at the 30-mg dose level in this study, whereas it was 69.6 ± 34.5 ng/mL and 741 ± 361 ng × h/mL at the 30-mg dose level from the single-agent phase I study. Similar results were seen at the 45-mg dose level as well. It is not clear whether the higher exposure seen in this study contributed to the overall incidence of toxicities. In another study of cediranib in combination with mFOLFOX-6, the mean cediranib C_ss,max was 83.0 ± 20.5 ng/mL and the mean AUC was 926 ± 423 ng × h/mL at 30-mg dose level. However, these preliminary results were from three patients only (18). In the present study, cediranib pharmacokinetics were evaluated at cycle 2, with concomitant administration of mFOLFOX-6. Therefore, it was not possible to determine whether mFOLFOX-6 and antiemetics affected cediranib pharmacokinetics resulting in higher exposure and larger interpatient variability. In addition, cediranib samples were taken less frequently in the present study compared with the single-agent study, potentially contributing to the larger interpatient variability.

Like other platinum agents, oxaliplatin rapidly forms a variety of reactive intermediates after i.v. administration. These intermediates then irreversibly bind to other macromolecules and are inactivated (24). Total platinum contents in plasma ultrafiltrate that represent all unbound platinum species are generally measured with atomic absorption or mass spectrometry methods. However, these methods could not distinguish intact oxaliplatin from platinum intermediates (24–26). Recently, a high performance liquid chromatography with postcolumn derivation method was reported for measuring free intact oxaliplatin (27). In the current study, free intact oxaliplatin concentrations were measured with a high performance liquid chromatography–tandem mass spectrometry method. A rapid initial decline in free intact oxaliplatin concentrations was observed with a half-life of ∼15 minutes. However, the slower elimination phases described previously were not observed in our study because these phases may represent the slower release of platinum species from platinum-amino acid conjugates after the degradation of cellular macromolecules. Therefore, these phases are unlikely to be of any clinical relevance. Results from our study are compatible to those reported by Ehrsson et al. (27). Cediranib did not affect the free plasma intact oxaliplatin and the steady-state 5-FU concentrations.

As expected, significant and promising antitumor activity was observed in this study, with only 2 of 14 evaluable patients having a best response of progressive disease. The median progression-free survival of 9.3 months in this study is comparable with the median progression-free survival of 9.4 months for patients treated with FOLFOX/XELOX plus bevacizumab in the XELOX-1 study (8). Side effects from the combination of cediranib and mFOLFOX-6 were manageable, although higher than expected for mFOLFOX-6 alone. This combination is currently being evaluated in randomized phase II/III and phase III studies in both previously untreated and treated metastatic colorectal cancer.

E.X. Chen, honoraria, AstraZeneca, Inc. L. Seymour, ownership interest, AstraZeneca, Inc.

We thank Jane Robertson, MD of AstraZeneca, Inc., for her thoughtful comments during the design of this study and preparation of this manuscript.

RECENTIN is a trademark of the AstraZeneca Group of Companies.