The epidermal growth factor receptor (EGFR) has become an important therapeutic target in the treatment of nonsmall cell lung cancer (NSCLC). Erlotinib (Tarceva®; Genentech, Inc., South San Francisco, CA) and gefitinib (Iressa™; AstraZeneca Pharmaceuticals, Wilmington, DE) inhibit the tyrosine kinase located within the intracytoplasmic domain of EGFR and have demonstrated objective response rates of approximately 10% in unselected individuals with advanced NSCLC who received previous chemotherapy.1, 2 More recently, it has been shown that erlotinib prolongs survival in previously treated patients with NSCLC.2 Many patients who are treated with these agents, however, fail to benefit from their use. Mutations within the EGFR tyrosine kinase domain have been identified as predictors of sensitivity to the EGFR tyrosine kinase inhibitors (EGFR-TKI) gefitinib and erlotinib.3–5 Other characteristics, such as EGFR and HER2 gene amplification and EGFR protein expression, also play a role.6–9 Mechanisms of intrinsic resistance remain incompletely understood but may include elevated levels of downstream kinases within the EGFR signaling pathway, such as Akt and mitogen-activated protein kinase (MAPK).10–12
In this regard, preclinical work has demonstrated that higher concentrations of erlotinib not only inhibit EGFR but also inhibit Akt and MAPK, implying that higher doses of erlotinib had the potential to overcome resistance to the drug in patients who had EGFR-driven tumors accompanied by the up-regulation of phosphatidylinositol 3′-kinase and/or p-AKT or by the down-regulation of phosphatase and tensin homolog (PTEN).13, 14 Most studies of erlotinib have focused on daily administration at the maximum tolerated dose (MTD), 150 mg. The principal toxicities of erlotinib administered on this dose and schedule are rash and diarrhea.1, 15 A weekly dosing schedule potentially would permit the delivery of higher doses of erlotinib with resulting higher peak concentrations. To determine whether higher doses may be delivered and whether antitumor efficacy is enhanced, we designed the current study of weekly high-dose erlotinib.
A prior Phase I trial attempted to study a similar strategy in patients with advanced cancers by administering erlotinib on Days 1, 8, and 15 of a 28-day cycle. Those investigators reported 2 episodes of “severe skin toxicity” in patients who received erlotinib at doses from 1400 mg to 1600 mg on this schedule, and the study subsequently was closed without defining the MTD.16 Because of those results, we assumed that patients would be unlikely to tolerate doses in excess of 2000 mg weekly; thus, 2000 mg erlotinib was the highest dose chosen for study in the trial reported here. Prior to closure, the previous Phase I study demonstrated adequate absorption of oral high-dose erlotinib, resulting in serum concentrations high enough to inhibit Akt and MAPK based on preclinical observations. The Phase I/II trial reported here was designed to evaluate the tolerability and efficacy of weekly erlotinib in patients with NSCLC.
MATERIALS AND METHODS
All patients had pathologically confirmed NSCLC, either Stage IIIB (based on the presence of pleural or pericardial effusion), Stage IV, or recurrent/medically inoperable disease.17 Eligibility requirements included at least 1 prior chemotherapy regimen for NSCLC, the completion of any prior chemotherapy and radiation therapy directed to major bone marrow-containing areas ≥3 weeks previous to study entry, measurable or evaluable disease in a nonirradiated field, and a Karnofsky performance status ≥60%. Laboratory parameters included a white blood cell count ≥3000/μL, hemoglobin ≥9 g/dL, platelet count ≥100,000/μL, total bilirubin ≤1.5 × the upper limit of normal (ULN), aspartate aminotransferase ≤1.5 × ULN, and creatinine ≤1.5 mg/dL or creatinine clearance ≥55 mL/minute. Patients were excluded if they had known brain metastases that were unstable radiographically or that required escalating doses of corticosteroids; if they had a concurrent, active cancer or other unstable serious illness; if they had received prior treatment with an EGFR-TKI; or if they currently were pregnant or lactating. Fertile men or women who were not using not using effective contraception also were ineligible for participation in the trial. This trial was reviewed and approved by the Institutional Review Board of the Memorial Sloan-Kettering Cancer Center. All patients signed informed consent.
Erlotinib was supplied by Genentech, Inc. in 100-mg and 150-mg tablets.
The Phase I portion of the trial consisted of 3 planned dose levels: 1200 mg, 1600 mg, and 2000 mg. Patients were enrolled in 3 patient cohorts with no intrapatient dose escalation. The final dose level was predefined to include 6 patients who were evaluable for toxicity after completing 4 weeks of therapy.
Dose escalation was based on the toxicities encountered through Day 28 of therapy. Toxicity for erlotinib was graded by using Common Toxicity Criteria, version 2.0, and had to be related possibly or probably to treatment. Dose-limiting toxicity (DLT) was defined as any of the following: Grade 3 diarrhea that lasted longer than 48 hours (despite intensive loperamide therapy), Grade 4 diarrhea, Grade 3 rash, Grade 3 fatigue that lasted more than 1 week, Grade 4 hematologic toxicity, or other Grade 3 nonhematologic toxicity (excluding alopecia, nausea, or emesis). If DLT occurred in 1 of the first 3 patients at a given dose level, then 3 additional patients were treated at that dose level. If no other patient experienced unacceptable toxicity, then dose escalation continued. If DLT occurred in 2 of 6 patients, then dose escalation was stopped. The MTD was defined as 1 dose level lower than the dose level at which at least 2 of 6 patients experienced DLT. The Phase II portion of the study commenced after completion of the Phase I portion using the dose defined as the MTD in the Phase I portion or the final dose level if no MTD was identified.
Management of Toxicity, Dose Reductions, and Interruptions
Patients were educated on the appropriate use of loperamide with the development of any treatment-related diarrhea.18 Doses of erlotinib were attenuated for Grade 2 or 3 diarrhea that persisted despite optimal medical management, including intensive loperamide therapy. Management of Grade 1 or 2 rash included symptomatic management and continuation of study drug at the current dose. Patients with troublesome Grade 1 or greater severe rash were treated with topical tetracycline, topical clindamycin, topical silver sulfadiazine, diphenhydramine, oral prednisone (short course), or oral minocycline. Doses of erlotinib were attenuated for Grade 3 rash and for intolerable or refractory Grade 2 rash.
Reduction/interruption of erlotinib dosing for adverse events was allowed at any time during the study. Two dose reductions per patient were permitted in the event of unacceptable toxicity. First and second dose reductions were to doses of 900 mg and 700 mg, respectively, at the 1200-mg dose level; 1200 mg and 900 mg, respectively, at the 1600-mg dose level; and 1600 mg and 1200 mg, respectively, at the 2000-mg dose level. Dose interruptions for a maximum of 2 weeks were allowed if clinically indicated and if the toxicity was not controlled by optimal supportive medication.
Within 2 weeks after a dose interruption or reduction, study drug-related toxicity must have improved by at least 1 grade (to Grade 1 or less), or further dose reduction by 1 level was required. No more than 2 dose reductions per patient were allowed.
Pretreatment and Follow-Up Evaluations
At baseline, all patients provided a history and underwent a physical examination and laboratory evaluations, including complete blood count, serum chemistry (including total bilirubin, aspartate aminotransferase, alkaline phosphatase, sodium, potassium, chloride, bicarbonate, creatinine and blood urea nitrogen), and electrocardiogram within 2 weeks of study entry; serum pregnancy test for women of childbearing potential; and baseline medical imaging that included computed tomography (CT) scanning of all relevant disease sites within 4 weeks of study entry. Patients then provided an interval history and underwent a physical examination weekly for the first 5 weeks and monthly thereafter. Laboratory evaluations, including complete blood counts and serum chemistry, were obtained every 2 weeks for the first month and monthly thereafter. Imaging with CT scans was obtained monthly for the first 2 months and every 2 months thereafter.
Objectives, Biostatistics, and Definitions
All patients who received erlotinib were monitored for toxicity and tumor response. The primary endpoint of the Phase I portion of this trial was to determine the MTD or to establish the safety and tolerability of the final, predefined dose level. Patients who received ≥4 weeks of study drug or who completed <4 weeks of study drug because of toxicity were included in the analysis of drug tolerability. All patients in this phase of the trial who developed disease progression prior to completing 4 weeks of study drug were replaced.
The primary endpoint of the Phase II portion of this trial was to determine the major objective response rate of weekly erlotinib administered at the dose identified in the Phase I portion of the trial. By employing a Simon 2-stage design, with the ineffective and desirable response rates set at 10% and 25%, respectively, both error rates (accepting a poor drug and rejecting a promising drug) were set at 10%. Enrollment of 50 patients was planned for the Phase II portion of the trial. To ensure sufficient activity of the study drug prior to proceeding with further enrollment, an interim analysis was planned after the 21st patient, with ≥3 objective responses required in the first 21 patients to continue enrollment. Responses were defined by using the Response Criteria to Treatment in Solid Tumors (RECIST) for unidimensional measurements,19 and all responses were to be confirmed with a follow-up scan ≥4 weeks later. All patients from the Phase I portion of the trial who were treated at the final dose level and for whom response assessments were available were included in the efficacy analysis of the Phase II portion of the trial.
Additional endpoints of the Phase II portion were to measure the survival and time to progression of patients who received the 2000-mg dose of weekly erlotinib. Overall survival in this study was estimated by using the method of Kaplan and Meier.20 The time to progression was measured from the start of treatment until the criteria for disease progression were met.
After the discovery that EGFR mutations were predictive of responsiveness to EGFR-TKIs,3–5 this trial was amended in 2004 to enable EGFR sequencing of available tumor tissue. Pretreatment tumor samples from selected patients were analyzed for EGFR mutations. DNA was extracted from tumor specimens, EGFR exons 19 and 21 were amplified, and uncloned polymerase chain reaction (PCR) fragments were sequenced in both sense and antisense directions for the presence of heterozygous mutations. All sequence variants were confirmed by independent PCR amplification, as reported previously.5 Patient selection for this analysis was based on their status as never-smokers.
Twenty-seven patients were enrolled between September 2003 and August 2004. Patient characteristics are outlined in Table 1. The median age was 63 (range, 39–82 years), and the median Karnofsky performance status was 80%. There were 10 women, 15 patients with adenocarcinoma pathology, and 5 never-smokers (defined as having smoked less than 100 cigarettes in their lifetime),21 all factors that are considered clinical predictors of responsiveness to erlotinib.22, 23
Table 1. Patient Characteristics
|No. of patients enrolled||27|
|Median age (range), y||63 (39–82)|
|No. of women/men||10/17|
|No. of ever-smokers||5|
|Karnofsky performance status (no. of patients)|
|Histologic cell type of NSCLC (no. of patients)|
|No. who received ≥2 prior regimens of chemotherapy||7|
|No. who received ≥3 prior regimens of chemotherapy||0|
|No. who received previous irradiation||6|
Fourteen patients were enrolled in the Phase I portion of this study. Three patients received erlotinib 1200 mg once weekly, and another 3 patients received erlotinib 1600 mg once weekly. Six patients initially were enrolled to receive 2000 mg once weekly. Two patients in this 2000-mg dose level developed progressive disease prior to completing 4 weeks of treatment and were replaced according to the protocol.
Selected toxicities are presented in Table 2. No Grade 3 or 4 toxicities were observed in the Phase I portion of the trial, although 1 patient was removed from the trial for intolerable Grade 2 fatigue. The MTD was not identified. Rash and diarrhea were the most frequent toxicities. Eleven of 12 evaluable Phase I patients experienced Grade 1 or 2 rash. The rash generally persisted but improved with continued weekly dosing. Of the 5 patients who developed Grade 2 rash in the Phase I portion of the trial, most did so within 2 weeks of starting erlotinib. The rash lasted 1 or 2 weeks in 3 patients and >4 weeks in 2 patients. These 5 patients were given topical clindamycin and/or oral minocycline. Similarly, 11 of 12 patients experienced Grade 1 or 2 diarrhea. Diarrhea generally was experienced within 24 to 48 hours of drug administration and was controlled well with loperamide therapy given on an as-needed basis.
Table 2. Selected Toxicities Possibly or Probably Related to Erlotinib*
|Nausea and/or emesis||2||0||1||1||2||1||0||0|
|AST and/ or ALT||0||0||1||0||3||0||0||0|
Other drug-related toxicities included fatigue, nausea, emesis, and stomatitis. Similar to the timing of the diarrhea, the nausea and emesis generally were short-lived and were experienced within 24 to 48 hours of drug administration. Laboratory studies generally improved from baseline on this study. None of the patients in the Phase I portion of the trial required interruption or dose reduction of erlotinib.
Neither of the 2 patients who developed progressive disease prior to completing 4 weeks of treatment and who were replaced in the Phase I portion of this trial experienced any unusual or Grade 2 toxicity. Both patients developed Grade 1 rash.
All 8 patients enrolled in the 2000-mg dose level of the Phase I portion, including the 2 patients who were replaced because of early disease progression, were evaluable for tumor response and were included in the efficacy analysis of the Phase II portion. Thirteen additional patients who received 2000 mg of erlotinib weekly were enrolled. Among the 20 patients for whom response assessments were obtained, 1 patient experienced a partial radiographic response, 11 patients had stable disease, and 8 patients had disease progression. The 21st patient was removed from the study after 1 week of treatment (see below) and had no response assessment. The patient who had a response, a male never-smoker with squamous cell carcinoma, had a near complete radiographic response at all sites of disease but 1 and subsequently developed progressive disease after 11.1 months of therapy. An intention-to-treat analysis of all 21 patients who received erlotinib 2000 mg weekly yielded a 4.7% objective response rate (95% confidence interval [95% CI], 0.2–22%).
The median time to progression for the 21 patients who were treated at the Phase II dose of erlotinib was 1.7 months (95% CI, 0.9–5.8 months). The median survival for all 21 patients was 9.5 months (95% CI, 7.3 months to not estimable). The 1-year survival rate was 40.3% (95% CI, 22.9–70.9%).
Toxicity in the Phase II portion of the study was similar to that identified in Phase I. Grade 2 rash with weekly erlotinib was more frequent numerically with dose escalation and developed in 42% of patients who received 2000 mg and in 17% of patients who received 1200 mg or 1600 mg (P = .36; Fisher exact test). Most patients on this trial also experienced drug-related diarrhea, although no patient experienced Grade 3 or 4 diarrhea. The frequency of Grade 2 diarrhea was 26% in the 2000-mg weekly group and 33% in the 1200-mg weekly and 1600-mg weekly groups (P = 1.0; Fisher exact test). Three patients experienced Grade 3 toxicity that was considered possibly or probably related to weekly erlotinib. One patient experienced Grade 3 dehydration and a Grade 2 creatinine elevation in the setting of Grade 1 nausea, emesis, and diarrhea. One patient developed Grade 3 fatigue that lasted <1 week. The third patient developed erlotinib-related, Grade 3 pneumonitis after a single dose. This patient presented on Day 8 complaining of cough and dyspnea on exertion. His oxygen saturation was 93% on room air. His chest CT scan that day revealed new, bilateral pulmonary infiltrates consistent with pneumonitis. He was hospitalized and treated with supplemental oxygen and glucocorticoids. His respiratory status improved with a return of his dyspnea to baseline. There was improvement but not resolution of bilateral infiltrates on his CT scan. Prednisone was discontinued after 4 weeks.
Only 2 patients required dose reductions on study. One patient had a dose reduction for persistent Grade 2 stomatitis. The other patient had a dose reduction for Grade 3 fatigue. In each patient, the erlotinib dose was reduced from 2000 mg weekly to 1600 mg weekly.
Analysis of EGFR mutation status in selected patients
Pretreatment tumor specimens from 4 patients who were never-smokers were analyzed for mutations within EGFR exons 19 and 21. The 5th never-smoker had no tumor tissue available. Of the 4 patients whose tumors were analyzed, 2 patients experienced disease progression, 1 patient had stable disease, and the other patient experienced a partial response. The patient who had a partial response had a tumor that was identified with an 18 base-pair deletion within EGFR exon 19 that encompassed the conserved leucine, arginine, glutamic acid, and alanine (LREA) motif. The 3 other patients' tumors had no EGFR mutations identified in exons 19 and 21.
This trial was undertaken to determine the tolerability and efficacy of high-dose erlotinib administered on a weekly schedule to patients with advanced NSCLC. The results demonstrated that oral erlotinib administered at doses of 1200 mg, 1600 mg, or 2000 mg once weekly was associated with modest toxicity in most patients. Although these doses amounted to a sizable increase compared with 150 mg erlotinib daily, the toxicity was similar.1, 2 In the BR-21 trial, 55% of patients who received erlotinib experienced diarrhea, 76% experienced rash, and 9% experienced Grade 3 or 4 rash.2 Although patients in the current trial experienced rash more frequently than with daily erlotinib, none had Grade 3 or 4 rash.
One patient developed Grade 3 pneumonitis after a single dose of erlotinib 2000 mg. Pneumonitis has been described in 0.3% to 2.0% of patients who received gefitinib daily.24, 25 Among 4900 erlotinib-treated patients, 0.7% experienced pneumonitis (Tarceva Investigator Brochure, 9th edition, April 2005). The rate of pneumonitis among 731 patients in the BR-21 trial was similar to that among erlotinib-treated and placebo-treated patients.2
Erlotinib 150 mg daily reportedly has an objective response rate of 8.9% to 12.3%.1, 2 The current trial was designed in part to determine whether weekly high-dose erlotinib could improve on that response rate. The intention-to-treat analysis, however, identified a 4.7% response rate (95% CI, 0.2–22%) among patients who received erlotinib 2000 mg weekly. Because we documented only 1 partial response in the first 21 patients who received erlotinib 2000 mg weekly, we did not meet the response threshold to continue the study according to the Simon 2-stage design we employed. Although >50% of patients who were treated experienced disease stabilization, this was short-lived, as demonstrated by the median time to progression of 1.7 months, which is not substantially different from the 2.2 months reported in the BR-21 trial.2 Although the median survival of patients who received weekly high-dose erlotinib on the current study was 9.5 months and compared favorably with the 6.7 months reported for patients who received daily erlotinib on the BR-21 trial, the number of patients treated on the current study was small, and the difference in survival may reflect improved patient selection in a single-institution, Phase II study. We do not believe that further testing of single-agent erlotinib administered at 2000 mg weekly is warranted.
Of particular interest is the limited activity that this treatment had in never-smokers. We chose to perform an exploratory analysis of EGFR mutational status in never-smokers, because these patients have a nearly 50% incidence of EGFR-TK domain mutations,26, 27 which many investigators believe is a surrogate for sensitivity to erlotinib and gefitinib.3–5, 27 Within the small group of patients who had their tumors sequenced, only the patient who had a radiographic response had a tumor that harbored an EGFR exon 19 deletion. The time to progression for this patient (11.1 months) was similar to that reported previously for erlotinib-responsive or gefitinib-responsive patients with NSCLC who received daily therapy (median, 5–8 months).1, 2, 28 The other 3 patients who had their tumors analyzed had no EGFR mutations identified. Two of those patients experienced disease progression within 8 weeks of starting treatment, whereas the 3rd patient experienced a prolonged period of stable disease (54 weeks). Although the small number of specimens analyzed precludes definitive conclusions with regard to the correlation between EGFR status and responsiveness to this schedule of erlotinib administration, the clinical courses of the patients with and without EGFR mutations in this study were not dissimilar from what has been reported for patients who received EGFR-TKIs administered on a daily basis. Potential drawbacks in the design of this trial included the lack of pharmacokinetic and pharmacodynamic studies, the limited number of dose escalations, and the absence of correlative studies.
Incomplete absorption and inadequate drugs levels are unlikely explanations for the limited efficacy we observed. In a prior study by Karp et al., in which patients received high-dose erlotinib on Days 1, 8, and 15 on a 28-day cycle, the bioavailability of erlotinib was approximately 60%. In addition, data from that study demonstrated that once weekly erlotinib at a does of 1600 mg resulted in peak serum concentrations (Cmax) of 8960 ng/mL (range, 5720–9430 ng/mL). Data from Akita et al., in which EGFR-overexpressing cell lines were stimulated with transforming growth factor α, demonstrated not only a dose-dependent inhibition of EGFR but also a similar relation for Akt and MAPK. Furthermore, those dose response curves were steep, such that a small increase in the concentration of erlotinib could cause near total inhibition of both of these downstream kinases. Based on the available in vitro data, the Cmax associated with erlotinib 1600 mg would have inhibitory effects on Akt and MAPK in vivo (unpublished results).13, 14, 16 It is plausible that the clinical activity of erlotinib is enhanced by consistent and prolonged exposure, which would be observed with daily administration, and that the fluctuations in plasma erlotinib concentrations that are possible with weekly dosing may reduce effectiveness.
We chose not to assess whether the biologic targets of the high-dose erlotinib, such as Akt and MAPK, were affected by this dosing schedule. Prior studies of EGFR-directed therapies used skin biopsies as surrogates to assess whether signaling through the EGFR and its downstream pathways were affected by the targeted therapy.29–31 Such biologic correlative studies have not been validated as representing intratumoral expression of the same kinases, which led us exclude those studies in design of the current trial. If promising activity had been seen, then the determination of intratumoral levels would have been justified. Our study did not address several issues fully, including whether Akt and MAPK were inhibited by erlotinib administered at this dose and schedule, whether higher intermittent dosing would cause better target inhibition, how long such inhibition may last, and whether there may even be a rebound phenomenon associated with intermittent inhibition.
Targeting Akt and MAPK in conjunction with EGFR-TK inhibition remains a valid objective for further translational research. Means of such dual inhibition may include the use of an intermediate dose of erlotinib, administered more frequently than weekly and at a higher dose than the daily MTD, to reach higher peak concentrations and, at the same time, to avoid significant fluctuations in plasma concentration and low trough levels. Alternatively, the simultaneous inhibition of EGFR, Akt, and MAPK could be tested clinically with the concurrent use of multiple agents. For example, mammalian target of rapamycin (mTOR) is a serine-threonine kinase located downstream of Akt that serves as a critical regulator of cellular growth and proliferation. Ongoing studies are testing the concurrent use of gefitinib or erlotinib and an mTOR inhibitor, such as Everolimus (Novartis Pharmaceuticals USA, East Hanover, NJ), in patients with advanced NSCLC.32
In summary, weekly high-dose erlotinib yielded a partial response rate of 5% in this unselected group of patients with NSCLC. Future studies should focus on understanding primary erlotinib resistance and on means by which resistance can be assessed pretreatment. Trials combining erlotinib with agents that target sites downstream of EGFR provide another approach to overcome resistance.