Imatinib inhibits the c-kit tyrosine kinase, which, accounts for its activity in gastrointestinal stromal tumors. The presence of c-kit protein expression in small cell lung carcinoma (SCLC) tumor specimens, as well as in vitro data supporting the role of c-kit in autocrine and paracrine growth stimulation specifically in SCLC, provided a rationale for studying imatinib in this disease. The authors conducted a Phase II single-institution study of imatinib in patients with recurrent SCLC whose tumor specimens expressed c-kit protein.
Patients with progressive SCLC after one or two previous chemotherapy regimens consented to have their tumor specimens screened by immunoperoxidase stain (CD117, Dako Corporation, Carpinteria, CA) for c-kit protein expression. If present, individuals were then eligible for treatment with an imatinib dose of 400 mg orally twice daily (total, 800 mg per day).
The presence of c-kit protein was assessable in 36 of 39 (92%) tumor samples. Twenty-eight (78%) tumor samples had immunohistochemical staining for c-kit protein. Twelve patients were enrolled in the treatment portion of the current study. No responses were observed, and all patients had disease progression by Week 4. Edema, fatigue, nausea, and electrolyte abnormalities were the primary toxicities.
The benefits of combination chemotherapy in newly diagnosed patients with small cell lung carcinoma (SCLC) are clear, with responses expected in > 80% of patients. However, most individuals ultimately succumb to recurrent disease, which generally becomes evident within a few months after the completion of initial therapy. The currently available chemotherapy treatments have limited efficacy in recurrent disease, and the median survival for patients with recurrent SCLC is approximately 6 months.1
To develop novel therapies for SCLC, a number of potential biologic targets have been identified, one of which is c-kit. C-kit is a glycoprotein, transmembrane receptor that requires phosphorylation of tyrosine residues by a constitutive tyrosine kinase. Numerous investigators have reported c-kit protein expression as determined by CD117 immunohistochemical staining, ranging from 28% to 93% of SCLC tumors.2–6 Some have suggested that c-kit expression has negative prognostic implications,4, 5 although others dispute this finding.7 In vitro studies support the role of c-kit and its ligand, stem cell factor (SCF), on SCLC autocrine and paracrine growth stimulation. Transfection of the NCI-H146 cell line (which naturally expresses SCF but not c-kit) with a c-kit expression vector results in increased cell growth.8 In addition, the growth of another SCLC cell line that expresses SCF and c-kit was markedly decreased when a defective c-kit gene was transfected.8
In addition to its well known inhibition of bcr-abl in chronic myelogenous leukemia, imatinib mesylate (STI-571, Gleevec; Novartis, Basel, Switzerland) also blocks c-kit signaling,9 providing a rationale for its study in SCLC.10 Imatinib can induce regressions of gastrointestinal stromal tumors (GIST),11 a disease whose pathogenesis is strongly related to constitutive c-kit activation. Furthermore, imatinib has demonstrated growth inhibition of multiple SCLC cell lines in vitro.12, 13 This growth inhibition correlates both with c-kit expression and the blockade of c-kit phosphorylation. We undertook this Phase II study to determine the efficacy of imatinib in patients with SCLC whose tumor specimens express c-kit protein.
MATERIALS AND METHODS
The primary objective of this single-institution Phase II trial was to determine the radiographic partial and complete response rate of imatinib in patients with SCLC tumor specimens that expressed c-kit and who had disease progression after one or two previous chemotherapy regimens. Patients were first consented for pathologic analysis. The presence of c-kit protein was determined by using CD117 polyclonal rabbit immunostain (Dako Corporation, Carpinteria, CA). Because no standard is available for grading c-kit staining, focal strong staining or diffuse staining was considered positive, whereas focal weak staining was considered negative.
Patients whose tumor specimens expressed c-kit were eligible to participate in the treatment portion of the study. Other eligibility criteria included a Karnofsky performance status ≥ 70% and measurable or evaluable disease. At least 4 weeks must have elapsed since the last chemotherapy treatment. Acceptable hematologic and biochemical parameters included an absolute neutrophil count ≥ 1.5 × 106 cells/μL, a platelet count ≥ 100,000/mm3, a serum creatinine level ≤ 1.2 mg/dL or creatinine clearance ≥ 50 mg/min, a total bilirubin level ≤ 1.5 mg/dL, and an aspartate aminotransferase level < 2.5 × the upper limit of normal. Patients with symptomatic brain metastases or leptomeningeal disease and patients with concurrent active disease were excluded. Because of a potential interaction with drugs metabolized by cytochrome p450 enzymes, patients were not allowed to take warfarin during the study.
Patients were treated with imatinib at a dose of 400 mg twice daily (a 800-mg total daily dose). We selected a higher dose than was tested in a similar SCLC trial14 to ensure that a dose response was not evident. A toxicity evaluation and physical examination were conducted pretreatment, and at Weeks 2 and 4. Patients developing Grade 3 or 4 hematologic toxicity had treatment withheld until resolution to Grade ≤ 1, and then were restarted at a dose of 300 mg twice daily. For patients with nonhematologic toxicity Grade ≥ 2, doses were withheld until resolution to Grade ≤ 1, with a similar dose reduction for Grade 3 or 4 toxicity. A computed tomography scan to evaluate response was performed 4 weeks later.
A Simon two-stage design was used to calculate sample size (P0 = 0.05, P1 = 0.20, α and β errors = 0.10). Twelve patients were enrolled initially, and was then expanded to 37 patients if a response was observed.
This protocol was reviewed by the institutional review board at Memorial Sloan-Kettering Cancer Center.
Thirty-nine patients agreed to pathologic analysis of their tumor samples. Three tumor samples were inadequate. Of the remaining 36 tumor samples, 28 (78%) had positive immunohistochemical staining for c-kit protein.
Twelve patients were enrolled in the treatment portion of this study. Table 1 details their baseline characteristics. Approximately two-thirds of the patients had been treated with one previous chemotherapy regimen, and one-third had been treated with two previous regimens. Approximately half of the patients had sensitive disease (i.e., disease progression > 3 months after their first-line chemotherapy regimen).
Table 1. Baseline Characteristics of Patients in the Treatment Portion of the Study (n = 12)
No. of patients (%)
Karnofsky performance status
No. of previous regimens
Response to first-line therapy:
Sensitive (≥ 3 mo response)
No complete or partial responses were observed (0% observed rate, 95% confidence interval, 0–27%). All patients developed evidence of disease progression by Week 4. The median survival period was 2 months (Fig. 1).
The most common toxicities are shown in Table 2. Edema, either of the lower extremities or face, was the most common severe adverse event with two patients developing Grade 3 edema requiring dose reduction. Other side effects included fatigue and nausea. Electrolyte abnormalities also were quite prevalent. One patient with extensive liver metastases developed a Grade 3 elevation in alanine aminotransferase that did not resolve with drug discontinuation and was believed to be due most likely to disease progression.
The frequent expression of c-kit protein in SCLC tumor specimens provided a strong rationale for evaluating the effects of imatinib in this disease. The rate of c-kit protein expression in the current study was 78% which fits among the broad range from previous reports. However, with regards to treatment, even in patients selected for having tumors expressing c-kit protein, no responses were observed. A challenging group of patients was enrolled, with one-third progressing after two previous chemotherapy regimens and one-half having refractory disease. Nonetheless, the lack of efficacy of imatinib in the current study is evident.
These results concur with those of another recently reported Phase II study.14 In that trial, 9 patients with untreated advanced-stage SCLC and 10 patients with sensitive disease recurrence were treated with imatinib at a dose of 600 mg daily. No responses were observed, and the median time to disease progression was 0.8 months. C-kit expression was not determined before the initiation of imatinib, although tumor specimens were collected. When analyzed, only 4 of 19 (21%) tumor specimens stained with CD117. As in the current study, the tumor samples were not screened for mutations in c-kit.
The basis for the lack of activity likely rests in the absence of activating c-kit mutations in SCLC. Patients with GIST who demonstrate impressive and durable disease remissions have evidence of mutations in the c-kit or PDGRFA genes that cause constitutive activation of tyrosine kinases that are inhibited by imatinib. In a Phase II study of imatinib in GIST, patients whose tumors had exon 11 or exon 9 kit mutations had partial response rates of 84% and 48%, respectively, whereas those with no kit or PDGFRA mutation demonstrated no response.15 The mechanism underlying c-kit protein expression in SCLC is uncertain. Activating mutations in c-kit have been reported, but at a much lower rate than c-kit protein expression. In one series, 14 of 22 (64%) SCLC tumor samples expressed c-kit protein as assessed by immunohistochemistry, yet no activating mutations in c-kit exon 11 were detected.16 Another group identified kit mutations in 5 (2 in exon 9, 2 in exon 11) of 60 SCLC tumor samples.7 Whether imatinib would be effective in patients whose SCLC tumor specimens harbor kit mutations is unknown. A xenograft model using SCLC lines known to constitutively express c-kit shows no effect of imatinib on tumor growth, despite achieving therapeutic concentrations of drug in the plasma and the xenograft tumors of the mice, casting doubt on this approach.17 Some investigators have suggested that if imatinib has cytostatic rather than cytotoxic effects, its use in conjunction with chemotherapy could potentially improve time to disease progression.18 Studies combining imatinib with standard chemotherapy have been conducted, including at our institution, and preliminary results indicate no apparent improvement in outcomes; however, there may be an increase in toxicity.19
The rapidity of disease progression in this population of persons with SCLC demonstrates a compelling “unmet need” for therapy that can also provide a unique opportunity for the testing of new therapies in lung carcinoma. In the current series, disease progression was documented in all patients by the first imaging study scheduled for just 4 weeks after starting imatinib. The median time to tumor progression was 1 month and the median survival was just 2 months. Organizations developing new therapies should consider these patients for early inclusion because of the desperate need in a substantial number of affected patients (> 10,000 subjects annually in the U.S. alone).