Phase I studies of imatinib mesylate combined with cisplatin and irinotecan in patients with small cell lung carcinoma




Small cell lung carcinoma (SCLC) cell lines commonly express KIT and its ligand, stem cell factor, suggesting an autocrine loop promoting cell growth. Imatinib inhibits KIT kinase activity. SCLC cells treated with imatinib in vitro undergo cell cycle arrest. Imatinib reduces resistance to irinotecan in vitro. Common metabolic pathways suggest there may be drug interactions between imatinib and irinotecan or cisplatin. In the current study, the authors investigated the feasibility of combining these drugs in the treatment of patients with SCLC.


Two Phase I studies were conducted independently at two institutions. Patients with extensive-disease SCLC underwent therapy with cisplatin, irinotecan, and imatinib using two similar regimens. In one study, immunohistochemical analysis of the expression of potential imatinib targets was performed on pretreatment biopsy specimens, and blood specimens were collected and analyzed for imatinib, irinotecan, and cisplatin pharmacokinetic parameters.


Nine patients were enrolled and were evaluable for toxicity. A high incidence of neutropenia, diarrhea, and thrombosis was observed that precluded dose escalation. Six patients were evaluable for response after four cycles; five patients experienced a partial response and the other patient had developed progressive disease. Four of six tumor specimens tested expressed platelet-derived growth factor receptor-α and two expressed KIT. Irinotecan clearance was found to be significantly decreased by imatinib (P < 0.04). No significant alteration in the disposition of cisplatin was observed.


The maximum tolerated dose for this combination with granulocyte–colony-stimulating factor support was identified as imatinib at a dose of 300 mg/day with irinotecan (at a dose of 65 mg/m2) and cisplatin (at a dose of 30 mg/m2) given on Days 1 and 8, every 21 days. The decreased irinotecan clearance may explain the high incidence of diarrhea and neutropenia noted in the current study. Cancer 2006. © 2005 American Cancer Society.

Advances in the understanding of the molecular mechanisms of cancer have led to novel biologic agents with antitumor effects. Imatinib mesylate (STI571) is a potent tyrosine kinase inhibitor that selectively inhibits the ABL family (ABL, BCR-ABL, ABL-related gene [ARG]), platelet-derived growth factor receptor (PDGFR), and KIT kinases1 Imatinib has antitumor activity against chronic myelogenous leukemia (CML)2 and gastrointestinal stromal tumors3 through its inhibition of BCR-ABL and KIT kinases, respectively.

Small cell lung carcinoma (SCLC) cell lines and tumor specimens commonly coexpress KIT and its ligand, stem cell factor (SCF).4–6 Treatment of SCLC cell lines with imatinib in vitro has been reported to inhibit SCF-mediated KIT activation and signal transduction. Imatinib also induces growth inhibition in SCLC cell lines and apoptosis when those cells are grown in the presence of exogenous SCF.7, 8 Imatinib was found to have modest antitumor activity in one of three human SCLC cell lines injected subcutaneously into mice.9

The efficacy of single-agent imatinib for SCLC was evaluated in three Phase II studies and was not associated with objective tumor regression.10–12 However, imatinib may affect tumor response when combined with traditional cytotoxic agents by preventing tumor regrowth between treatment cycles. Imatinib also may prevent resistance to irinotecan by inhibiting the ABCG2 transporter or increasing topoisomerase I activity.13, 14

In the current study, we report on the evaluation of imatinib combined with irinotecan and cisplatin in two trials for patients with SCLC. A variety of different chemotherapy regimens have been used for SCLC and none have been proven clearly superior.15 The combination of cisplatin and irinotecan was reported to be superior to cisplatin and etoposide in terms of survival for patients with extensive SCLC in a single Phase III study.16 The Memorial Sloan-Kettering Cancer Center (MSKCC) trial used the same schedule of irinotecan and cisplatin as was used in this Phase III study. In the University of Texas M. D. Anderson Cancer Center (MDACC) trial, we used a schedule based on a Phase II study of weekly irinotecan and cisplatin in patients with advanced nonsmall cell lung carcinoma.17 In this study, patients received weekly cisplatin at a dose of 30 mg/m2 and irinotecan at a dose of 65 mg/m2 for 4 weeks and then had 2 weeks with no chemotherapy (“4 on–2 off”). Because many patients missed Weeks 3 and 4 of therapy, we gave patients weekly cisplatin at a dose of 30 mg/m2 and irinotecan at a dose of 65 mg/m2 for 2 weeks followed by 1 week without chemotherapy (“2 on–1 off”) to increase dose intensity.

There are potential pharmacokinetic interactions between imatinib and chemotherapy. Imatinib is principally metabolized by CYP3A4 to the N-demethyl derivative; other cytochrome p450 enzymes play minor roles in its metabolism.18 Imatinib also is a potent competitive inhibitor of CYP2C9, CYP2D6, and CYP3A4/5. The coadministration of imatinib and agents that are metabolized by cytochrome 450 enzymes may result in increased exposure to imatinib and the coadministered agents. Irinotecan is a camptothecin derivative that is converted by carboxyesterases to an active metabolite, 7-ethyl-10-hydroxycamptothecin (SN-38).19, 20 SN-38 is conjugated further in the liver and then excreted in the bile and urine. In addition, irinotecan undergoes oxidation mediated by CYP3A4/5 to various metabolites with varying degrees of activity.19, 20 Cisplatin is a platinating agent with an unclear metabolic pathway.21 To the best of our knowledge, no pharmacokinetic drug interactions have been observed between irinotecan and cisplatin.22 Given imatinib's shared metabolic pathway with irinotecan, combination therapy with these agents may lead to increased irinotecan or imatinib exposure and toxicity.

In the current study, we conducted two independent Phase I studies to establish the toxicity and maximum tolerated dose (MTD) of imatinib when combined with irinotecan and cisplatin for the treatment of patients with extensive SCLC. In the study conducted at MDACC, tumors were evaluated for the expression of ABL, ARG, PDGFR-α, PDGFR-β, and KIT. We collected pharmacokinetic data to assess the interactions of these three agents when given concurrently. Because of the potential interaction between irinotecan and imatinib, our starting dose of imatinib was reduced to 300 mg from the standard 400 mg dose used in treating CML patients. The study conducted at MSKCC used the same drugs with a different schedule.


Patient Eligibility

Patients with a histologic/cytologic diagnosis of untreated SCLC, extensive disease, and asymptomatic central nervous system metastases were eligible for these studies. The inclusion criteria included measurable or evaluable disease; adequate performance status (Zubrod performance status of 0-2 required at MDACC and a Karnofsky score ≥ 70% required at MSKCC); no prior invasive tumor within 5 years; and adequate hematologic, hepatic, and renal function. The studies were approved by the two institutional review boards and patients provided written informed consent.. Pretreatment tissue for immunohistochemistry was required for the MDACC study, but results did not affect eligibility.

Study Design

The doses of irinotecan and cisplatin were fixed, and the dose of imatinib varied. A standard “3+3” design was used, with the first cohort treated at Dose Level 0. If no patients experienced a dose-limiting toxicity (DLT), the next cohort would be treated at the next higher dose level. The MTD was defined as that dose producing a DLT in two of six patients or that dose level immediately below the one producing a DLT in at least three of six patients.

In the MDACC trial, patients received irinotecan at a dose of 65 mg/m2 intravenously [i.v.] and cisplatin at a dose of 30 mg/m2 i.v. on Days 1 and 8, every 21 days. Oral imatinib was administered Day 2 of the first cycle (to allow for baseline pharmacokinetics on Day 1). There were 3 predefined dose levels of imatinib: Dose Level 0 (300 mg daily), Level 1 (400 mg daily), and Level 2 (600 mg daily). In the MSKCC trial, patients received irinotecan (at a dose of 60 mg/m2 i.v.) on Days 1, 8, and 15 and cisplatin (at a dose of 60 mg/m2 i.v.) on Day 1 only of a 28-day cycle. Treatment with oral imatinib was initiated at Dose Level 0 on Day 22 of Cycle 1 (1 week before Cycle 2). There were 3 predefined oral dose levels of imatinib: Dose Level 0 (400 mg daily), Level 1 (600 mg daily), and Level 2 (800 mg daily). Imatinib was continued after the completion of four cycles of chemotherapy until disease progression or unacceptable toxicity developed.

The starting dose level of imatinib was based on the safety, tolerability, and serum levels of imatinib in patients with CML. Higher doses may be beneficial in patients with SCLC given the relatively high 50% inhibitory concentration (IC50) of imatinib (5 μM) reported in SCLC cell lines in vitro.

Using the National Cancer Institute (NCI) Common Toxicity Criteria, the DLT in the MDACC study was defined as a Grade 3 or higher nonhematologic toxicity (excluding nausea and emesis), Grade 4 neutropenia lasting > 7 days, febrile neutropenia, or Grade 4 thrombocytopenia. The MSKCC study also used NCI Common Toxicity Criteria with similar parameters to define the DLT: any Grade 4 hematologic toxicity, neutropenic fever, a dose delay of > 2 weeks, creatinine ≥ 2 mg/L for > 2 weeks, Grade 3 diarrhea lasting > 48 hours, any Grade 4 diarrhea, Grade 3 fatigue lasting > 1 week, or any other Grade 3 or 4 nonhematologic toxicity.

Chemotherapy Administration

The cisplatin infusions were given first, followed by irinotecan. Imatinib was given daily beginning on Day 2 of Cycle 1 (at MDACC) or Day 22 of Cycle 1 (at MSKCC).

Weekly chemotherapy was delayed for 1 week if the absolute neutrophil count was ≤ 1200 k/μL or the platelet count was ≤ 100,000 k/μL. Nonhematologic toxicity must have resolved to Grade 1 before chemotherapy could continue. If the creatinine clearance was 40–59 mL/minute, cisplatin was given at a 50% reduced dose. Cisplatin was withheld for a creatinine clearance < 40 mL/minute.

Evaluation of Response to Treatment

Pretreatment laboratory tests included a complete blood count with differential and platelet count; alanine aminotransferase (ALT), aspartate aminotransferase (AST), total bilirubin, alkaline phosphatase, total protein, albumin, calcium, electrolytes, blood urea nitrogen (BUN), creatinine, and magnesium; and urinalysis. A chest X-ray, electrocardiogram, computed tomography (CT) scan of the chest and abdomen, magnetic resonance imaging of the brain, and radionuclide bone scans were obtained within 4 weeks of the initiation of treatment.

Weekly laboratory tests included complete blood count with differential and platelet count and electrolytes, magnesium, BUN, and creatinine. Alkaline phosphatase, ALT, total bilirubin, calcium, and albumin were measured before each cycle. A medical oncologist evaluated each patient before every cycle.. Chest X-rays were obtained during Weeks 4, 7, and 13 in the MDACC study. Patients were restaged with CT after Cycle 4 (at MDACC) and after Cycles 2 and 4 (at MSKCC).

World Health Organization criteria23 (at MDACC) and Response Evaluation Criteria in Solid Tumors (RECIST) criteria (at MSKCC) were used to define response.24

Pharmacokinetic Studies

Pharmacokinetic studies were conducted in the MDACC trial. Samples for determining plasma concentrations of imatinib were collected at the following times: 0 (predose), 2 hours, 4 hours, 6 hours, and 8 hours after oral ingestion on Day 15 of Cycle 1 (imatinib alone) and Day 1 of Cycle 2 (imatinib and chemotherapy). Samples for determining irinotecan and cisplatin concentrations were collected on Day 1 of Cycle 1 (chemotherapy alone) and Cycle 2 (imatinib and chemotherapy) at the following times: 0 (predose), 0.5 hours, 1.7 hours, 4 hours, 5.5 hours, and 8 hours after the initiation of irinotecan infusion and at 0 (predose), 1 hours, 3 hours, 4.5 hours, and 7 hours after the initiation of cisplatin infusion. All samples were collected into a heparinized vacutainer and separation was performed by centrifugation at 1500 revolutions per minute for 10 minutes at 5 °C. For cisplatin, an additional separation step was taken to isolate unbound cisplatin via ultracentrifugation.25 Processed samples were stored at −70 °C.

The concentrations of imatinib were analyzed using a validated liquid chromatography/tandem mass spectrometry method with a lower limit of quantitation of 0.100 ng/mL.26 We were unable to obtain SN-38 to use as a standard to quantitate SN-38 levels. The concentration of irinotecan was analyzed using a validated high-performance liquid chromatography method with fluorescence detection.27 Concentrations of cisplatin in plasma ultrafiltrate samples were analyzed using validated flameless atomic absorption spectroscopy (Varian, Palo Alto, CA) with a graphite tube atomizer assay.25 The dynamic range for platinum was 50–400 ng/mL. All values calculated from the calibration curve were converted to an equimolar amount of cisplatin. For all analytic methods, the interday and intraday coefficients of variation were < 15%. Pharmacokinetic parameters for imatinib, irinotecan, and cisplatin were estimated using standard noncompartmental methods (WinNonlin [version 3.1]; Pharsight Corporation, Mountain View, CA). For irinotecan and cisplatin, the following parameters were determined: elimination half-life (t½), maximum plasma concentration (Cmax), time to reach Cmax (t max), volume of distribution (V), and area under the concentration time curve from zero to the last measured time point (AUC(0-t)). Statistical analysis was completed using a Student t test for paired data (SigmaPlot 2002 [version 8.0], SPSS Inc., Chicago, IL).

Immunohistochemical Analysis

Immunostaining was performed on pretreatment biopsies from patients at MDACC. Routine immunohistochemical protocols similar to a previous study28 in melanocytic lesions were used to detect KIT (Dako Corporation, Carpinteria, CA), c-ABL, ARG, PDGFR-α, and PDGFR-β (Santa Cruz Biotechnology, Santa Cruz, CA). Antibody labeling was optimized using antigen retrieval (treatment with pepsin at 37 °C for 10 minutes for PDGFR-α and PDGFR-β, microwave treatment for 5 minutes for KIT) and no antigen retrieval for c-ABL or ARG. Diaminobenzidine was used as a chromogen. The percentage of positive cells and the intensity of staining were semiquantitatively recorded by 2 observers (V.G.P. and P.T.): < 5%, 5% to < 25%, ≥ 25% to < 75%, or ≥ 75% of positive cells. Labeling intensity was categorized into weak, moderate, or strong. For a marker to be considered positive, the intensity had to be moderate in ≥ 25% of the tumor cells.


Patient Characteristics and Treatment Administration

Six patients were enrolled in the MDACC trial in two separate cohorts (Table 1). Five underwent four cycles each of therapy at Dose Level 0. One patient discontinued therapy after two cycles because of preexisting periodontal disease that required full-mouth dental extractions before further chemotherapy. Four patients received maintenance imatinib for 3–5 months after completing 4 cycles of chemotherapy. One patient required a dose reduction of irinotecan for Grade 3 diarrhea, nausea and emesis, and hypokalemia. Granulocyte–colony-stimulating factor (G-CSF) was added prophylactically for all patients in the second treatment cohort because of neutropenia that was observed in the first cohort. Because of toxicity, no additional patients were enrolled at any other dose levels.

Table 1. Patient Characteristics of the MDACC and MSKCC Studies
No. of evaluable patients
Patient characteristicMDACC (n = 6)MSKCC (n = 3)
  • MDACC: University of Texas M. D. Anderson Cancer Center; MSKCC: Memorial Sloan-Kettering Cancer Center.

  • a

    A Zubrod performance status of 0-2 was required at the University of Texas M. D. Anderson Cancer Center and a Karnofsky score ≥ 70% was required at Memorial Sloan-Kettering Cancer Center.

Mean age in yrs (range)59 (44–73)72 (70–74)
Performance statusa  

Three patients were enrolled at MSKCC (Table 1). Only one patient remained in the study and underwent all four cycles of chemotherapy. One patient discontinued therapy after one cycle and never received any imatinib because of toxicity (neutropenia and diarrhea) from treatment with cisplatin and irinotecan alone. One patient completed three full cycles of chemotherapy but was unable to complete Cycle 4 (Days 8 and 15) because of prolonged neutropenia. Although maintenance imatinib was planned, none of the patients received it because of early withdrawal for toxicity and disease progression.

Toxicity (the MDACC Study)

The first three patients at MDACC experienced substantial neutropenia that required G-CSF support. Patient 1 experienced neutropenic fever during Course 3 and Patient 2 experienced a delayed granulocyte nadir after Course 1. G-CSF was added for both patients, and neither experienced significant neutropenia in subsequent courses. Patient 3 had delayed Grade 2 neutropenia after Course 2 and therefore G-CSF was added to Courses 3 and 4. However, this patient died with neutropenic fever after Course 4. Because the only DLT noted in the first cohort was neutropenia, G-CSF was added prophylactically to the subsequent cohort of three patients. The second cohort did not experience significant neutropenia. There were no other Grade 4 hematologic toxicities noted (Table 2).

Table 2. MDACC and MSKCC Drug Toxicities
ToxicityaMDACC (n = 6 patients)MSKCC (n = 3 patients)
No. of patients (n = 6)No. of courses (n = 22)No. of patients (n = 3)No. of courses (n = 9)
  • MDACC: University of Texas M. D. Anderson Cancer Center; MSKCC: Memorial Sloan-Kettering Cancer Center.

  • a

    Toxicity was graded according to the National Cancer Institute Common Toxicity Criteria.

Anemia, Grade 32311
Thrombocytopenia, Grade 31100
 Grade 21400
 Grade 32222
 Grade 42211
 Grade 51100
Neutropenic fever2222
 Grade 251900
 Grade 31111
 Grade 40011
Thrombosis, Grade 32311
Electrolyte disorder, Grade 3 or 41233

The most common nonhematologic toxicity was diarrhea (Table 2). Only 1 patient experienced Grade 3 diarrhea, but 5 patients experienced Grade 2 diarrhea in 19 of 22 chemotherapy courses (86%). Two patients experienced thrombosis; one patient developed a deep vein thrombosis (DVT) during Cycle 1 and both had asymptomatic pulmonary emboli (PE) found on a restaging chest CT scan after Cycle 4. The patient with Grade 3 diarrhea also experienced Grade 3 hypokalemia. There were no other Grade 3 or 4 nonhematologic toxicities reported.

The MTD was based on two of six patients, who experienced nonhematologic DLTs: one patient with PE and DVT and one patient with Grade 3 diarrhea and Grade 3 hypokalemia. Although each of the first three patients in the original cohort experienced substantial neutropenia, this was subsequently prevented in all patients with the addition of G-CSF support.

The four patients who underwent imatinib maintenance therapy tolerated it well, with no serious toxicity reported. One patient with known liver metastases experienced pancreatitis after 3 months of maintenance imatinib and died 3 weeks later. The cause of the pancreatitis was not apparent on imaging.

Toxicity (the MSKCC Study)

All three patients at MSKCC experienced Grade 3 or 4 neutropenia, and two patients developed neutropenic fever. There were no other Grade 4 hematologic toxicities reported. Diarrhea was the most common and severe nonhematologic toxicity. One patient each experienced Grade 3 and Grade 4 diarrhea with associated electrolyte abnormalities (Table 2). One patient developed a DVT after Cycle 2, was treated with heparin, and subsequently experienced a left thigh hemorrhage. One patient experienced a Grade 3 elevation of transaminases, and one patient experienced Grade 3 fatigue. All three patients experienced DLT, and no additional patients were enrolled at any other dose levels.

Immunohistochemical Analysis

The pretreatment tissue of all six patients at the MDACC was analyzed by immunohistochemistry for KIT, ARG, ABL, PDGFR-α, and PDGFR-β. Those cases with higher numbers of labeled cells tended to demonstrate the strongest labeling intensity. Two patients' tumors expressed KIT; of the other four tumors, two focally expressed KIT (classified as negative) and two were negative. No tumors expressed ABL, and only one tumor expressed ARG. All tumors had some degree of expression of PDGFR-α and PDGFR-β, but only tumors from Patients 3 and 1, respectively, met the criteria for positivity (Fig. 1). At MSKCC, one of two evaluable patients was found positive for KIT by immunohistochemistry.

Figure 1.

Representative photomicrographs of immunohistochemical staining for (B) KIT and (D) platelet-derived growth factor receptor-α (PDGFR-α) in pretreatment core needle biopsy specimens are shown along with corresponding hematoxylin (A) and eosin (C) stains. Original magnification ×400.


Pharmacokinetic analyses of cisplatin and irinotecan were completed in the MDACC study during Cycle 1 before imatinib administration and during Cycle 2 after patients had been receiving imatinib for 20 days (Table 3). In the presence of chronic exposure to imatinib, the mean irinotecan clearance decreased by 36% (P < 0.01), which increased the mean irinotecan exposure from 1818 mg/L × hour to 2925 mg/L × hour. Cisplatin clearance was found to be unaffected by imatinib (P = 0.77).

Table 3. Effect of Imatinib on the Pharmacokinetics of Irinotecan and Cisplatin
Pharmacokinetic parameters of irinotecan) or cisplatinCycle 1Cycle 2Fold changeP value
  1. SD: standard devation; AUC: area under the curve.

Irinotecan clearance (mL/hr/m2)41.80.0226.80.010.65< 0.01
Irinotecan half-life (hrs)2.170.373.391.251.56< 0.04
Irinotecan, AUC (mg/L × hr)1818925292513211.67< 0.04
Cisplatin clearance (mL/hr/m2)2.740.922.630.820.960.77
Cisplatin half-life (hrs)2.590.712.200.790.850.34
Cisplatin, AUC (mg/L × hr)25.412.825.812.01.020.89

The plasma concentrations of imatinib were measured over 8 hours after a dose of imatinib on Day 15 of Course 1 (7 days after chemotherapy) and Day 1 of Course 2 (the day of chemotherapy). There was significant interpatient variability in the serum levels of imatinib (Fig. 2) and limited sampling led to difficulties in evaluating the disposition of imatinib. The degree of intrapatient variability was found to be lower.

Figure 2.

In the M. D. Anderson Cancer Center study, blood samples for determining plasma concentrations of imatinib were collected at 0 (predose), 2 hours, 4 hours, 6 hours, and 8 hours after oral ingestion on Day 15 of Cycle 1 (solid symbols, without chemotherapy) and on Day 1 of Cycle 2 (open symbols, with chemotherapy). Complete data were available for five patients, and only data regarding Day 1 were available for Patient 4 (-x-) (3 μg/mL equals approximately 5 μM of imatinib).


All six patients enrolled at MDACC were found to have marked tumor regression on chest X-ray after one cycle. Four patients were evaluable for response after all four cycles. All 4 patients experienced partial responses and subsequently developed progressive disease after 3–5 months of imatinib maintenance therapy. Three of these patients had died at the time of last follow-up (Table 4).

Table 4. Patient Survival
Patient no.Time to disease progression (wks)Overall survival (wks)
  • MDACC: University of Texas M. D. Anderson Cancer Center; MSKCC: Memorial Sloan-Kettering Cancer Center; NA: not applicable.

  • a

    The patient died during the study after achieving a partial response as determined by chest X-ray.

  • b

    The patient left the study early for dental extraction, and experienced disease progression.

  • c

    The patient was lost to follow-up at 44 weeks.

  • d

    The patient left the study early because of toxicity.


Two patients enrolled at MSKCC were evaluable for response. One patient developed progressive disease 2 weeks after completing all 4 cycles of therapy. The second achieved a partial response but experienced disease progression 10 weeks after the discontinuation of treatment.


These Phase I studies were designed to determine the MTD of imatinib when combined with irinotecan and cisplatin in patients with extensive-disease SCLC. There was unexpected toxicity at initial dose levels, including a high frequency and degree of neutropenia and diarrhea and the occurrence of thrombosis, that was observed in both studies independently. Although neutropenia was prevented by the addition of G-CSF, we were unable to escalate the dose of imatinib in the MDACC study because two patients experienced nonhematologic DLT. Chronic exposure to imatinib led to an increased half-life and AUC of irinotecan. We hypothesize that inhibition of the oxidative pathways via CYP3A4 of irinotecan by imatinib led to the decreased clearance of irinotecan, leading to increased SN-38 metabolites via the carboxyesterases pathway and therefore increased neutropenia and diarrhea. There were no apparent alterations in the disposition of cisplatin or imatinib.18

To our knowledge two other clinical trials to date have evaluated imatinib and cytotoxic chemotherapy in SCLC patients. A Phase I study of etoposide, cisplatin, and imatinib in patients with extensive-disease SCLC closed after three patients were enrolled because of slow accrual. The regimen was not associated with unexpected hematologic toxicity, significant diarrhea, or thrombosis (unpublished data). Preliminary data from a Phase II trial of carboplatin (AUC = 4, Day 1), imatinib (at a dose of 600 mg daily), and irinotecan (at a dose of 60 mg/m2 on Days 1, 8, and 15) demonstrated reasonable tolerance with a very lose dose of carboplatin and accrual was ongoing at last follow-up.29 Previous experience with imatinib as a single agent or in combination with other types of chemotherapy for patients with leukemia or other solid tumors does not appear to predict the toxicity we observed.30–32

Diarrhea and neutropenia are common toxic effects of the combination of irinotecan and cisplatin. These two studies used two different chemotherapy schedules. In the MSKCC study, there was more Grade 3/4 diarrhea noted compared with the MDACC study. However, with only three patients and six patients, respectively, in each study, comparisons are problematic. The weekly regimen on which the MDACC study was based17 actually had a higher rate of Grade 3/4 diarrhea compared with the regimen on which the MSKCC study was based,16 although the comparison of these two published studies is hindered by the fact that the patients had different malignancies and ethnic backgrounds.

Although the neutropenia and diarrhea could be due to decreased irinotecan clearance or the additive toxicities of this combination, the reason for the thrombosis is not clear. Thrombosis has not been reported in patients with SCLC who were treated with irinotecan and cisplatin,16 but a 31% incidence of thrombosis was reported in a Phase I trial of irinotecan, cisplatin, 5-fluorouracil, and gemcitabine in patients with solid tumors.33 A panel review of irinotecan trials in patients with colorectal carcinoma defined a vascular syndrome that included myocardial infarction, cerebrovascular accidents, and PE.34 Arterial and venous thromboses also have been described in patients receiving cisplatin.35 It has been hypothesized that damage to tumor cells or the vascular endothelium by chemotherapy may lead to the release of procoagulant factors. Cisplatin prolongs the activation of platelets when they are cocultured with peripheral blood monocytes36 and levels of von Willebrand factor are reported to increase after the administration of cisplatin.37 In animal models, imatinib led to peritumoral blood clotting that may be the result of an inhibition of KIT on mast cells.38 Therefore, it is possible that the combination of these agents leads to an increased risk of thrombosis.

Our finding of KIT expression in two of six SCLC tissue samples is consistent with findings in the published literature.4–6 The results of the current study also confirm previous findings demonstrating the frequent expression of PDGFR-α in SCLC tumors.39 To our knowledge, the significance of the inhibition of PDGFR kinases has not been determined for SCLC. It is interesting to note that the cytostatic effects of imatinib are not entirely dependent on KIT expression in vitro7 or in vivo,40 suggesting that another target, such as PDGFR or ABL, may be relevant. A larger sample size is needed to determine the frequency of expression of PDGFR and ARG in patients with SCLC.

There were no objective responses to imatinib reported in three Phase II trials, two of which required KIT-positivity.10–12 The most likely explanation for these negative results is that KIT is not critical for the survival and growth of SCLC. The relatively high IC50 of imatinib in vitro (approximately 5 μM)7 and a paucity of activating mutations in SCLC (8.3%)6 support this hypothesis. This does not preclude an additive effect of imatinib with chemotherapy. Although there was no synergy reported with etoposide or carboplatin in vitro,7 imatinib may reverse resistance to camptothecin analogs by inhibiting the ABCG2 transporter (breast carcinoma resistance protein [BCRP]).13 SCLC cells express BCRP/ABCG2 in vitro, and inhibition via antisense oligonucleotides can increase sensitivity to irinotecan.41

To our knowledge, these are the first studies published to date to examine the combination of cytotoxic chemotherapy with imatinib in patients with SCLC. It appears probable that the increased toxicity observed with the addition of imatinib is because of decreased irinotecan clearance. It might have been possible to escalate the imatinib dose if the irinotecan dose was reduced. However, we believed it was not appropriate to reduce the dose of an effective drug to escalate the dose of an experimental agent. A schedule in which irinotecan is separated from the imatinib dose might decrease the pharmacokinetic interactions. However, this would theoretically also reduce therapeutic interactions as well. Therefore, on the basis of the inability to escalate the imatinib dose in this combination, the problematic drug interaction, and the lack of single-agent activity, we do not plan any further investigation of the combination.