Phase 1 Clinical and Pharmacokinetics Evaluation of Oral CI-1033 in Patients with Refractory Cancer

Purpose: To determine the tolerability and pharmacokinetics of oral CI-1033, a pan-erbB tyrosine kinase inhibitor, administered over 14 consecutive days of a 21-day cycle.

Design: Phase 1, multicenter trial involving patients with solid tumors that are refractory to standard therapy. CI-1033 was administered initially at 300 mg/day to a minimum cohort of three patients. Dose escalation proceeded at ≤40% increments. Patients were evaluated for toxicity, pharmacokinetic profile, and evidence of response.

Results: Thirty-two patients entered the trial and were evaluable for safety assessment. Dose-limiting toxicity (diarrhea, rash, and/or anorexia) occurred at the 560 mg dose level; the maximum tolerated dose was 450 mg. No patients achieved objective responses and six patients achieved stable disease. Plasma CI-1033 concentrations increased with increasing dose. CI-1033 was not eliminated in urine to any appreciable extent.

Conclusions: CI-1033 is suitable for phase 2 testing at the 450 mg/day dose level when administered for 14 days in a 21-day cycle. The pharmacokinetic profile is consistent with biologically relevant plasma concentrations over the dosing interval.

CI-1033, an orally available 4-anilinoquinazoline derivative, is a highly potent and selective pan-erbB tyrosine kinase inhibitor that irreversibly blocks intracellular signaling through all members of the erbB receptor family [erbB-1 (epidermal growth factor receptor, EGFR), erbB-2 (HER-2/neu), erbB-3, erbB-4, and the erbB-1 truncate, EGFR vIII; refs. (1–3)]. The erbB family of tyrosine-specific protein kinases is involved in the regulation of many malignant cell functions, including proliferation, differentiation, survival, induction of angiogenesis, metastatic spread, and resistance to chemotherapy and radiotherapy (4, 5). Over 90% of all solid tumors express at least one member of the erbB receptor family (6–8) and erbB receptor/ligand overexpression correlates with poor prognosis and reduced survival in patients with a variety of cancers including breast cancer, melanoma, and head and neck cancer (3, 9–15). In vitro work also shows that overexpression of two or more erbB family members has been correlated with a further worsening of prognosis (16–20).

Numerous anticancer therapeutic approaches to blocking signal transduction through erbB-mediated pathways have been attempted, including antibodies targeting the erbB extracellular domain, small molecule tyrosine kinase inhibitors specifically targeting one or more erbB family members, antisense oligonucleotides, ligand-linked toxins, and immunotoxin conjugates (21–25). Trastuzumab (Herceptin), a monoclonal antibody that specifically targets the erbB-2 extracellular domain, was the first effective clinical translation of this approach and is approved for the treatment of erbB-2 overexpressing metastatic breast cancer—as second-line monotherapy and as first-line therapy when combined with paclitaxel (26–31). More recently, an erbB-1-specific monoclonal antibody, cetuximab (Erbitux, IMC-C225), was approved for treatment of irinotecan-refractory colorectal cancer. To date, the only marketed erbB tyrosine kinase inhibitor is gefitinib (Iressa; ZD1839), specifically targeting erbB-1 and indicated as third-line therapy for locally advanced or metastatic non–small cell lung cancer (NSCLC; ref. 32). Taken together, the efficacy observed with these compounds strongly supports the hypothesis that the erbB signaling pathway is a viable target for antitumor therapy.

The irreversible inhibition of tyrosine kinase activity by CI-1033, via covalent bonding to the ATP binding site of the tyrosine kinase domains of the erbB receptors, differentiates it from most of the other tyrosine kinase inhibitors currently in development, most of which are reversible, competitive inhibitors. Whereas CI-1033 permanently inactivates the tyrosine kinase domain of the erbB receptors, requiring synthesis of new erbB receptor molecules to restore erbB signaling, there is a requirement of the other reversible tyrosine kinase inhibitors to maintain a continued presence at the receptor target for maintenance of signaling inhibition. Irreversible inhibitors have been shown to induce a more prolonged suppression of receptor kinase activity than reversible inhibitors, with a resultant improvement in both in vitro and in vivo antitumor activity (1, 2, 25, 33–35). Additionally, the irreversible receptor binding of CI-1033 is key in its ability to induce ubiquitylation and degradation of both erbB-1 and erbB-2 receptors, a property not shared by reversible erbB tyrosine kinase inhibitors (36).

Preclinical studies of CI-1033 showed significant growth delays and tumor stasis when using a variety of in vitro cell lines and in vivo human xenografts, including those derived from breast, lung, colon, pancreatic, ovarian, brain, and epidermoid cancers. Marked CI-1033-induced tumor regressions were also observed in multiple xenografts, including SF-767 glioblastoma and the MDA-MB-468 human breast cancer model (37, 38). Enhanced chemosensitivity was reported in erbB-1-overexpressing cell lines when cisplatin was combined with CI-1033 (39). A synergistic interaction was also shown when combined with other cytotoxic agents such as topotecan, gemcitabine, cisplatin, and SN-38 (the active metabolite of irinotecan; refs. 39–41). Likewise, significant synergistic radiosensitizing effects were shown when CI-1033 was used in concert with fractionated radiotherapy (42). Studies in animals revealed rapid absorption of CI-1033 from the gastrointestinal tract with good distribution into all tissues except the brain (37). The drug is highly protein-bound and primarily cleared through the biliary tract. The most common side effects of CI-1033 in animals were diarrhea and skin lesions, especially at higher doses (37).

Preliminary results of a phase 1 investigation with CI-1033 administered for 7 days, followed by 2 weeks of no treatment in a 3-week cycle, revealed an objective response in a patient with squamous cell skin cancer and stable disease in 9 of 18 NSCLC patients (>12 weeks in study). In fact, one of these patients showed stable disease for 93 weeks.⁴

Common toxicities included grades 1 to 2 diarrhea, acneiform rash, and nausea in nearly half of the patients. Mucositis and vomiting occurred in one quarter of patients as did transient thrombocytopenia (28%; ref. 38). Biopsy specimens from treated patients showed that total erbB-1 levels were decreased with a proportionate decrease in phosphorylated erbB-1 after CI-1033 had been administered for 7 days. This correlated with decreased tumor cell proliferation as assessed by Ki-67 immunohistochemistry studies from the same biopsy sample. However, 1 week post-treatment, recovery of some of the total EGFR was observed (38). Based on pharmacokinetic data, there was no indication that CI-1033 accumulated following 7 days of once daily dosing, nor was there any correlation between plasma concentrations of CI-1033 and any particular adverse events (43). Previously, we conducted studies on erbB kinase inhibition with a 7-day schedule (38). erbB-1 phosphorylation inhibition was shown to be 40% to 50%; higher doses of CI-1033 failed to show greater inhibition. The current study was designed to test the hypothesis that an intermittent 14 days on/7 days off schedule would allow for greater drug exposure per 3-week treatment cycle compared with a 7 days on/14 days off schedule without a significant increase in toxicity.

The principal objectives of this phase 1 dose-escalation study were to determine the safety profile, maximum tolerated dose, and pharmacokinetics of CI-1033 when given daily for 14 days followed by a 7-day drug-free period (14 days on/7 days off) in patients with advanced stage nonhematologic malignancies. In addition, clinical evidence of antitumor activity would be noted.

This was an open-label, noncomparative, dose-finding, multicenter phase 1 study. Following a baseline period during which protocol-mandated evaluations were obtained, CI-1033 was administered daily as a single p.o. dose for 14 consecutive days beginning on day 1 of a 21-day treatment cycle. Adverse events were evaluated according to National Cancer Institute Common Toxicity Criteria, version 2.0 guidelines (44). Courses were repeated every 21 days, contingent on return of all treatment-related adverse events, except tolerable skin rash or alopecia, to grade ≤1 or baseline toxicity levels. Patients were allowed to continue treatment in the absence of disease progression or intolerable toxicity. For safety reasons, all patients were followed for a minimum of 30 days after the last dose of CI-1033.

Treatment was discontinued if any of the following occurred: (a) protocol criteria for progressive disease were met, (b) adverse events remained intolerable despite dose adjustments, (c) patient refused further study treatment, or (d) the investigator concluded that it was in the patient's best interest to discontinue therapy.

CI-1033 was formulated in gelatin capsules containing 5, 25, or 250 mg of study medication, plus inactive ingredients of lactose, cornstarch, talc, and magnesium stearate. Site personnel stored CI-1033 capsules under refrigerated conditions (2-8°C). Patients stored CI-1033 capsules at room temperature. Patients were told to refrain from consuming food and liquid, other than water, for 2 hours before and 2 hours after each dose and were encouraged to take their dose at approximately the same time each day. Patients were instructed not to double up or replace missed doses. A study medication diary was completed by the patient during each course of treatment.

The starting dose for CI-1033 was 300 mg. This was less than half the maximum tolerated dose (650 mg) determined in a prior phase 1 investigation using a 7 days on/14 days off dosing schedule. CI-1033 dose escalations could proceed at ≤40% increments with a minimum cohort of three patients if the required number of treated patients met the following criteria during cycle 1: (a) received a 14-day course of CI-1033 within a 15-day period, (b) did not vomit within 2 hours after receiving at least 13 doses of CI-1033 during the treatment cycle, (c) completed follow-up through the first cycle of study treatment (day 21). Patients who did not meet the above criteria for reasons other than dose-limiting toxicity (DLT) were replaced in order to achieve a full cohort. Patients within the same dose cohort received the same starting dose regardless of weight or body surface area. Intrapatient dose escalation was not permitted. If a DLT was observed at a given dose level, the cohort was expanded up to six patients. No dose escalation was permitted beyond the dose at which ≥2 patients experienced the same DLT during the first treatment cycle. DLT was defined as any adverse event that met any of the following criteria (National Cancer Institute Common Toxicity Criteria, version 2.0): any grade 4 hematologic toxicity, grade ≥3 neutropenia associated with a documented infection or fever, platelets <25 × 10⁹/L; grade ≥3 nonhematologic toxicity or the inability to begin a subsequent treatment course within 21 days of the scheduled start date due to treatment-related toxicity. Medically manageable nausea, vomiting, or diarrhea, and tolerable skin rash were not considered a DLT.

For patients that interrupted dosing due to DLT, treatment resumption was permitted at the next lower dose level tested, beginning with the next scheduled course, contingent on return of neutrophils to ≥1.5 × 10⁹/L, platelets to ≥75 × 10⁹/L, and all other toxicities (except tolerable skin rash or alopecia) to grade ≤1 or baseline toxicity levels.

Men and women ≥18 years old with nonhematologic malignancies refractory to standard therapy, or for which no effective therapy existed, were eligible for this study. Patient requirements included Karnofsky performance status index of ≥60, ability to swallow intact CI-1033 capsules, and expected survival ≥12 weeks. Tumor erbB overexpression was not a requirement for study entry. Written informed consent was obtained from each subject, and an Institutional Review Board reviewed and approved the study protocol at each institution. This study complied with ethical standards outlined in the Declaration of Helsinki. Patients were treated in the outpatient clinics of either the Mary Crowley Medical Research Center, Baylor University, Dallas, TX, or University of Texas, M.D. Anderson Cancer Center, Houston, TX.

Patients were excluded if they had: creatinine clearance <45 mL/minute (calculated by Cockcroft and Gault formula), total bilirubin >1.5 times the upper limit of normal, alanine aminotransferase or aspartate aminotransferase >3 times the upper limit of normal, absolute neutrophil count <1.5 × 10⁹/L or platelets <100 × 10⁹/L, active gastrointestinal bleeding, acute gastrointestinal ulcers, history of major gastrointestinal surgery that could affect intestinal absorption, concurrent serious infection, or coexisting medical problem of sufficient severity to limit full compliance with the protocol. Patients were not allowed to have had antineoplastic chemotherapy or investigational agents within 4 weeks prior to first study dose (6 weeks for nitrosoureas or mitomycin), hormonal therapy (except luteinizing hormone releasing hormone analogues for advanced stage prostate cancer), immunotherapy, or other biological therapy within 2 weeks prior to first dose, or concurrent radiotherapy or radiotherapy within 2 weeks prior to first dose. Women who were pregnant or nursing, or of childbearing potential and not using an adequate barrier method of birth control were also excluded from the study.

Peripheral blood sampling. Patients underwent pharmacokinetic profiling on day 14 of cycle 1. Blood samples were collected using an EDTA/ascorbic acid anticoagulant just prior to and then 1, 3, and 6 hours after drug administration. Plasma samples were stored at −20°C prior to assay. CI-1033 was isolated from a 0.2 mL aliquot of plasma using an acetonitrile protein precipitation reaction. Separation from supernatant and quantitation was done by liquid chromatography and mass spectroscopy. Resulting range of quantitation for CI-1033 was 1.00 to 500 ng/mL. Overall precision expressed as the relative SD (%CV) of quality control samples averaged 9.9% with overall accuracy expressed as the percentage of relative error (%RE) ranged from −7.7% to −1.6%.

Urine sampling. A 20 mL urine sample was collected prior to CI-1033 administration and from 0 to 8 hours on study days 1, 7, 22, and 28. Total volume was recorded and an aliquot was retained frozen until pharmacokinetic analysis. Analytic methodology was identical to that used for quantitation in plasma with the exception of using a 0.05 mL sample. A working quantitation range of 10.0 to 1,000 ng/mL was established. Overall precision (%CV) of quality control samples averaged 13% with overall accuracy (%RE) ranging from −5.0% to 6.56%.

Antitumor activity. Tumor assessments were completed at baseline and at the end of every two cycles if followed by chest radiographs or CT/MRI, or monthly if followed by physical examination. Objective response (complete or partial) was assessed using the sum of the products of bidimensional measurements of from one to five target lesions. A complete response was defined as disappearance of all measurable disease confirmed by a second evaluation at least 28 days later, with resolution of clinical symptoms of disease and no appearance of new lesions. A partial response was defined as a decrease of ≥50% in the sum of the bidimensional products of all target lesions confirmed no less than 28 days later and without the appearance of new lesions or worsening of tumor-related signs and symptoms. Progressive disease was defined as either an increase of ≥25% in any bidimensional lesion measurement, an increase of ≥25% in the sum of the products of the bidimensional lesion measurements, appearance of new lesions, or progression of clinical signs and symptoms. Patients who remained in the study for ≥12 weeks without objective response or progressive disease were considered to have had stable disease.

Monitoring timetables of visits and procedures. At baseline, patients underwent a medical history, physical exam, electrocardiogram, ophthalmologic exam, clinical laboratory testing including hematology, coagulation profiles, blood chemistry, urine analysis, and, for appropriate women, serum pregnancy testing. Clinical laboratory testing was done weekly during the first two treatment cycles, at the beginning and end of the dosing interval for subsequent cycles, and at the end of the study. Pharmacokinetic samples were collected on day 15 following 14 days of dosing. All observed or volunteered adverse events, regardless of causal relationship to study drug, were recorded during the study. Patients that withdrew from the study for any reason were followed for 30 days to monitor recovery or occurrence of any treatment-related adverse events.

Patient population. Thirty-two patients with advanced stage nonhematologic malignancy were enrolled in the trial; each received at least one dose of study drug and was assessable for safety. Demographics of the patients are shown in Table 1. Most patients (75%) had been treated with two or more prior chemotherapy regimens. Head and neck cancer patients comprised one third of the study population.

Dose escalation. Five dose levels were examined, 300 mg (nine patients), 350 mg (three patients), 450 mg (six patients), 500 mg (eight patients), and 560 mg (six patients). Three of the nine patients entered into cohort 1, and two of the eight patients entered into cohort 4 were not fully evaluable due to poor compliance on medication intake. Grade 3 DLTs were observed in three of six patients at the 560 mg level (Table 2): diarrhea/nausea/vomiting (one patient), rash (one patient), and anorexia (one patient). Two additional patients in the same cohort had non-DLT grade 3 events: diarrhea (one patient) and pruritic erythema (one patient). After discussion with the investigators, it was decided that despite not having the same DLT occur in two patients in the cohort, dose escalation would stop at the 560 mg dose level. Dose finding proceeded at the next lower dose of 500 mg at which three of eight patients experienced DLTs: diarrhea/dehydration/stomatitis (one patient), elevated bilirubin (one patient), and vasculitis/rash (one patient). In addition, three patients experienced grade 3 diarrhea while receiving loperamide. The maximum tolerated dose was declared at the 450 mg dose level at which one of six patients experienced grade 3 dehydration associated with grade 2 stomatitis. There were no DLTs at the 350 mg dose level; however, of the nine patients dosed at 300 mg, there was a grade 3 hypersensitivity reaction which manifest as tongue swelling after an initial CI-1033 dose (one patient), and one patient had dose-limiting grade 3 dehydration immediately preceding a grade 4 cerebral infarct (one patient).

Safety. The most frequently reported adverse events occurred within the gastrointestinal system and skin and were typically grade 1/2 in intensity. Diarrhea, nausea, stomatitis, rash, vomiting, and acne were most common (Table 3). The frequency of diarrhea, nausea, stomatitis, and vomiting seemed to increase with increasing dose, whereas the incidences of rash and acne seemed to be relatively dose-independent within the range of doses tested. One patient (pancreatic cancer) experienced a grade 3 elevation in liver enzymes (aspartate aminotransferase/alanine aminotransferase). Four additional patients had grade 1/2 increases in liver enzymes, although three of these patients had disease involvement of the liver.

The onset of gastrointestinal toxicity usually occurred within the first week of treatment (median 1-6 days) and was typically manageable with standard medical therapy. At the higher dose groups (450, 500, 560 mg), treatment modifications due to intolerable gastrointestinal toxicity occurred in 4 of 12 patients (33%); three patients required treatment discontinuation (one patient per dose group) and one patient at the 560 mg cohort required dose reduction to 450 mg, a dose that he continued for 11 subsequent cycles.

Skin toxicity was described with a variety of adverse event terms including rash, acne, and maculopapular rash. The onset of cutaneous toxicity typically occurred within a week of therapy. Most were grade 1 to 2 in intensity and occurred with similar frequency across dose groups. At the higher dose levels (500-560 mg) 2 of 14 patients experienced grade 3 rash during their first treatment cycle requiring dose reduction (one patient) and discontinuation from the study (one patient). There was no consistent trend in changes in rash intensity (improvement or exacerbation) over multiple treatment courses, as such treatment was related to support management. Most common rash-related symptoms necessitate treatment (“itchiness”) was managed with Benadryl (25-50 mg p.o. 6 hours p.r.n.) and Zantech (75-150 mg p.o. q. 12 hours).

Although thrombocytopenia was commonly reported as a laboratory finding (35% of patients) it was rarely considered a clinically significant adverse event. Decreases in platelet values (grades 1-2) were detectable a week after starting the study drug and in some cases as early as after 4 dosing days. Platelet values typically returned to grade ≤1 within a week off the study drug, regardless of platelet nadir and severity did not seem to worsen as additional cycles of CI-1033 were received. Of those three patients who reported thrombocytopenia as a grade 3 adverse event, there was one concurrent event of mild epistaxis reported.

Allergic reactions to CI-1033 were rare. One patient who received an initial 300 mg dose experienced a grade 3 hypersensitivity reaction (tongue swelling, pruritis, and redness of the head and neck) that was promptly alleviated with oral antihistamine and steroid treatment. Two additional cases of grade 1 urticaria were also reported.

Response. In this end-stage patient population, a high number of patients (75%) discontinued study participation within the first 3 months. The primary reason for treatment discontinuation was disease progression (Table 4).

Fifteen patients with measurable disease were evaluable for objective response. No confirmed partial or complete responses were observed. Six patients fulfilled criteria for stable disease; three patients with head and neck cancer, one patient with NSCLC, one patient with colorectal cancer, and one patient with parotid cancer (median, 24 weeks; range, 13-41 weeks). Two patients with head and neck cancer showed minor responses with tumor reductions of 30% to 40%.

Pharmacokinetics. Twenty-two of 32 patients provided evaluable pharmacokinetic profiles (Table 5). Day 14, pre-dose (trough) concentrations ranged from below the assay limit of detection to 83 ng/mL. There was no apparent dose dependence in trough concentration. One, 3, and 6 hours post-dose concentrations tended to increase with increasing dose with peak concentrations observed within 1 to 3 hours of dosing. Less than 1% of the administered dose was recovered intact in urine with the majority of samples providing levels at or below the limit of quantitation (10 ng/mL). There was no clear pattern in urinary CI-1033 excretion compared with dose or dosing duration.

In the refractory cancer patient population studied, the maximum tolerated dose of oral CI-1033 is 450 mg when administered once-daily for 14 consecutive days, every 3 weeks (14 days on/7 days off). The most common adverse events were gastrointestinal and dermatologic in nature, predicted by preclinical CI-1033 studies and consistent with the phase 1 safety profiles of other small molecule erbB-targeted tyrosine kinase inhibitors in development, such as gefitinib and erlotinib (45, 46). The predominant DLTs (grade 3 diarrhea, dehydration, and rashes) and the frequent occurrence of grade 1-2 diarrhea, nausea, and rashes are also not unexpected for a compound like CI-1033 that targets the erbB family of receptors. Thrombocytopenia has been reported in other CI-1033 phase 1 studies as well as in this study (38). Most thrombocytopenia is of grade 1-2 intensity and is often noted as a laboratory observation that did not require medical intervention; however, clinically significant thrombocytopenia has been reported, particularly at the higher ranges of dose escalation. The etiology of thrombocytopenia is unclear and has proved difficult to assess due to a limited opportunity for bone marrow sampling and the employment of more tolerable doses in phase 2 studies. The potential clinical significance of thrombocytopenia when CI-1033 is combined with radiotherapy and/or myelosuppressive chemotherapy in future studies remains to be evaluated. CI-1033-associated allergic reactions have occurred in a small number of patients. Their infrequency and prompt reversibility with antihistamines and glucocorticoids suggest that this type of adverse event will not likely prevent future studies from using initial p.o. CI-1033 doses up to the maximum tolerated dose (450 mg).

During the phase 1 investigation of CI-1033, it became apparent that as the daily dosing schedule became more continuous, the corresponding maximum tolerated doses became noticeably lower. Whereas 650 mg of p.o. CI-1033 was tolerated when given daily for the first 7 days of a 21-day dosing cycle, no more than 175 mg was tolerated when CI-1033 was taken continuously (47). If one compares the total CI-1033 dose exposures that could potentially be received over 12 weeks of treatment, it becomes evident that >60% more CI-1033 could be given using the intermittent 7 days on/14 days off schedule (18.2 gm) compared with the continuous daily schedule (11.1 gm). Because the key drivers of antitumor efficacy for an irreversible, pan-erbB tyrosine kinase inhibitor such as CI-1033 are not known, the current study was conducted, in part, to test the hypothesis that an intermittent 14 days on/7 days off schedule would allow for maximal CI-1033 exposure, albeit at the possible expense of a lower maximum tolerated daily dose than was achievable on a 7 days on/14 days off schedule. As shown by this current study, maximum drug exposures (25.2 gm per 12-week interval) have been achieved on an intermittent 14 days on/7 days off schedule at the maximum tolerated dose of 450 mg. Whether or not this 14 days on/7 days off schedule will translate to enhanced clinical antitumor efficacy compared with continuous daily dosing of CI-1033 will be determined in phase 2 CI-1033 testing. In contrast to this ability to increase drug exposure of CI-1033 when given intermittently rather than by continuous daily dosing, phase 1 studies of gefitinib do not show such a differential in tolerability when a 14 days on/14 days off schedule (maximum tolerated dose, 525 mg) is compared with continuous daily dosing (maximum tolerated dose, 600 mg; refs. 45, 48). The reason for this discrepancy is not clear, although the distinctive biochemical irreversibility and the broader pan-erbB activity of CI-1033 compared with gefitinib may be contributing factors.

Plasma samples collected in this study were intended for population pharmacokinetic analysis as part of a much larger data set. Although sample collection was limited to no more than four samples per patient, there are still important observations that can be drawn from this subset of patients. With the exception of day 14 pre-dose samples, systemic drug exposure in the present study was generally consistent with that observed in the wider population. Previous analyses have suggested CI-1033 to have a half-life of approximately 3 to 5 hours (46). CI-1033 binds an erbB family irreversibly, rendering the plasma half-life a less effective indicator of dosing frequency requirements. Receptor turnover, instead, provides the best indicator of dosing frequency required for erbB signaling suppression. This is in contrast to what is observed with reversibly bound erbB inhibitors. Early assessments of receptor activity suggest turnover to be slow enough that signaling is still suppressed at trough using once daily dosing regimens. Results reported here show some patients to have pre-dose concentrations significantly above the limit of quantitation which would be inconsistent with the previous half-life estimate. Patient-to-patient variability in drug disposition may be larger than first estimated, giving rise to a larger range of elimination rates and a greater potential for once-daily systemic exposure than first thought. Alternatively, the patients comprising this particular subset may differ from the larger population in ways that slow CI-1033 elimination. The significant lack of CI-1033 excreted in urine indicates elimination to be primarily metabolic. Preclinical studies indicate significant metabolism of CI-1033 in animal models and in animal and human liver microsomes, with the primary route being glutathione conjugation at the reactive site (49). Glutathione transferases are ubiquitous enzymes responsible for the detoxification of a number of electrophilic molecules and are unlikely to vary significantly within the patient population (49). This lends support to the notion that previous estimates of the variability in the intrinsic CI-1033 elimination rate in humans may have been underestimated in the past. It is also possible that previous models have addressed an early distribution phase but have not addressed a more prolonged terminal elimination phase. Further work combining this data with that collected over a wider range of patients and dosing regimens will be required to address the source of this apparent discrepancy. In addition, the relationship of pharmacokinetics to clinical outcome has to be considered in light of biopsy data in a previous trial showing sustained inhibition of EGFR in tumor specimens a week after CI-1033 exposure was stopped. This is well beyond the time required to achieve nonmeasurable plasma levels of CI-1033.

Although there were no confirmed objective responses in this study of single-agent CI-1033 in cancer patients with advanced refractory solid tumors, there was evidence for antitumor activity, namely minor response in two head and neck cancer patients. In addition, although difficult to interpret, prolonged stable disease at the higher CI-1033 doses also suggests disease activity, although it is possible that this simply represents indolent disease. In phase 2 trials with erlotinib, rash severity has been shown to be associated with objective tumor response in NSCLC, head and neck, and ovarian cancer patients (50). Although rash was noted in each of the patients achieving stable disease or minor responses in this phase 1 study with CI-1033, a definitive correlation is impractical due to the small patient numbers.

In summary, CI-1033 is a novel, high-potency, orally available pan-erbB tyrosine kinase inhibitor that holds potential as an effective treatment for solid tumors that express any of the multiple erbB receptor family members. Compared with continuous daily administration of CI-1033, intermittent dosing on a 14 days on/7 days off schedule allows notably greater drug exposure over time, consequent to better tolerability. In this phase 1 clinical study, CI-1033 doses ranging from 300 to 450 mg/day given on a 14 days on/7 days off schedule were well tolerated and showed a pharmacokinetic profile compatible with once-daily dosing. A 14 days on/7 days off dosing regimen may prove important in future efficacy studies investigating CI-1033 as a single agent as well as when it is combined with radiotherapy, chemotherapy, or other targeted therapies. The full potential of CI-1033 as an effective anticancer treatment, perhaps non–cross-resistant to that of other erbB targeted therapies, awaits more definitive phase 2 and 3 testing.

We thank Brenda Marr for her competent and knowledgeable assistance in the preparation of this article.