A phase II study of the safety and efficacy of the multidrug resistance inhibitor VX-710 combined with doxorubicin and vincristine in patients with recurrent small cell lung cancer

Authors


Abstract

BACKGROUND.

Tumors with multidrug resistance (MDR) frequently up-regulate efflux proteins, including MDR-associated protein (MRP-1) and P-glycoprotein (Pgp). MDR represents an obstacle to successful chemotherapy treatment and is reversible in Pgp- or MRP-1-expressing cells by the inhibitor VX-710. A Phase II study was designed to evaluate VX-710 in combination with doxorubicin and vincristine in patients with sensitive, recurrent small cell lung cancer (SCLC).

METHODS.

Eligible patients had recurrent SCLC after a response to first-line chemotherapy. Stage 1 safety evaluation was completed with planned expansion if 9 responses were confirmed in the first 35 patients. Patients were treated every 21 days until progression or intolerable adverse events (AEs).

RESULTS.

Thirty-six patients were enrolled from 1998 to 2000. Neutropenia was the major toxicity, occurring in 26 of 36 patients (72%). Neutropenia was more severe (30% vs 20% grade 4) and developed earlier (58% vs 38% in Cycle 1) among the 15 patients who were enrolled prior to an amendment that required neutropenia prophylaxis. Four patients died on study: 2 from infections likely related to therapy and 2 from cancer progression. Seven of 36 patients (19%) had partial responses; 6 patients sustained responses through 6 cycles of treatment, with 1 response lasting 3 years. Three additional patients had unconfirmed responses, and 4 patients had stable disease. The median survival was 6 months. No correlative 99mTc-sestamibi uptake in tumor tissue was observed with the addition of VX-710 in this study.

CONCLUSIONS.

The addition of VX-710 to doxorubicin and vincristine therapy did not significantly enhance antitumor activity or survival. Although there were durable responses, criteria were not met to proceed with Stage 2 expansion. Cancer 2007 © 2007 American Cancer Society.

Small cell lung cancer (SCLC) is an aggressive form of lung cancer, and the mainstay of treatment for SCLC is chemotherapy. However, despite initial high susceptibility to antiproliferative therapy (response rates of 60–80%1, 2), recurrences are common, and resistance to subsequent therapies is frequent. Chemotherapy for patients who develop recurrent SCLC is less effective than the initial chemotherapy; the response rate with the approved second-line therapy, topotecan, is only 22%.1 The duration of response in second-line therapy is shorter than in first-line therapy and produces a median survival of 6 months.2 Therefore, alternative strategies are needed to identify treatments that can overcome the development of resistance to improve outcomes for patients with recurrent SCLC.

A number of resistance mechanisms have been identified in SCLC, including up-regulated expression of genes involved in multidrug resistance (MDR). MDR factors include genes encoding for P-glycoprotein (P-gp) and the MDR-associated protein (MRP-1), both of which prevent accumulation of some chemotherapeutic agents through adenosine triphosphate-dependent transport out of cells.

Several studies have evaluated the expression of P-gp and MRP-1 in drug-resistant SCLC cell lines and tumor specimens. The majority of studies of P-gp and MRP-1 expression have demonstrated a correlation between increased expression and resistance to chemotherapy.3–5 It is noteworthy that this correlation appears to be specific to P-gp and MRP-1 expression and not to other efflux transporters. One study analyzed biopsy specimens of 40 patients with SCLC prior to chemotherapy for immunohistochemical evidence of P-gp and MRP-1.6 None of the patients who had a clinical response to 6 cycles of cisplatin and etoposide showed evidence of either P-gp or MRP protein detected by immunohistochemistry, whereas 12 of 20 patients with stable or progressive disease after the same treatment expressed both proteins (P < .01).

Thus, an agent that modifies MDR through P-gp or MRP-1 could provide substantial clinical benefit for patients with SCLC who are treated with chemotherapy. Several different MDR inhibitors have been developed, but most are specific for a particular MDR transport protein (reviewed by Robert, 19986 and Tan et al., 20007). For example, PSC833 (Valspodar) interacts specifically with P-gp. However, multiple proteins have been implicated in MDR, which may limit the efficacy of a highly specific inhibitor. VX-710 (biricodar, INCEL) is an agent that binds directly to both P-gp and MRP-1. The binding of biricodar to P-gp and MRP-1 inhibits their pump efflux activity and increases intracellular accumulation and retention of cytotoxic agents. VX-710 restored drug sensitivity to drug-selected as well as combinational DNA-transfected cell lines at levels from 1 μm to 5 μm for all MDR class cytotoxic drugs tested (doxorubicin, vincristine, vinblastine, etoposide, paclitaxel, and colchicine).9

A Phase I study of VX-710 plus doxorubicin established the safety and pharmacokinetics of VX-710 administered by continuous intravenous infusion for 96 hours in combination with intravenous bolus doxorubicin at 45 mg/m2. VX-710 blood steady-state concentrations of 10 μM were sustained at a dosage of 120 mg/m2 per hour, and VX-710 had no apparent effect on doxorubicin pharmacokinetics or pharmacodynamics.8

Therefore, we conducted a multicenter, open-label, Phase II study to evaluate VX-710 plus doxorubicin and vincristine in patients with SCLC who had developed recurrent disease after achieving a response to first-line therapy. Cyclophosphamide, vincristine, and doxorubicin (CAV) historically was the regimen of choice for SCLC before the advent of etoposide and platinum combinations. In a key randomized trial of patients with previously treated SCLC, CAV was compared with topotecan9 in patients with chemotherapy-sensitive, recurrent SCLC (>60 days from the last chemotherapy to disease progression). The outcome of patients who received CAV was similar in efficacy to that of patients who received topotecan with an 18% response rate, a 12-week time to progression (TTP), and a 6-month median survival. In selecting a regimen to use with VX710, the decision to omit cyclophosphamide was based on the fact that it is unaffected by MDR efflux pumps.

The doses of doxorubicin and vincristine were chosen to be similar to those used in the CAV regimen. The initial cohort planned for a half dose of vincristine to establish the safety of the combination of vincristine and VX-710. Like in previous studies, the VX-710 infusion was initiated prior to administration of the chemotherapy drugs to inhibit P-gp and MDR and to increase cellular retention of the subsequent chemotherapy, and the infusion was continued for 72 hours to sustain inhibition through 2 to 3 half-lives of doxorubicin and vincristine.

A subset of patients underwent imaging with 99mTc-sestamibi before and after their first dose of VX-710 as a potential tool for measuring the impact of MDR inhibition. 99mTc-sestamibi is a transport substrate for P-gp and MRP-1.10, 11 VX-710 and other MDR inhibitors restore 99mTc-sestamibi uptake and retention in P-gp and MRP-1 expressing cells in vitro.12 Studies in patients with breast cancer and lung cancer have reported higher retention of 99mTc-sestamibi in tumors that expressed low levels of P-gp. In contrast, a higher 99mTc-sestamibi wash-out rate and reduced retention were noted in lesions that expressed high levels of P-gp.13–15 Increased wash-out and poor retention appeared to correlate with a poor treatment response.

The objectives of the current study were 1) to establish the safety of VX-710 when used in combination with doxorubicin and vincristine in patients with SCLC; 2) to establish the ability of VX-710, given in combination with doxorubicin and vincristine, to improve the response rate to chemotherapy in patients with SCLC who developed recurrent disease on first-line therapy; and 3) to evaluate the MDR profile of patients by analyzing biopsy specimens for the expression of P-gp and MRP and using functional imaging of tumors with 99mTc-sestamibi.

MATERIALS AND METHODS

Patient Population

Patients who were eligible for this study were aged ≥18 years and had histologically confirmed SCLC and documented progressive disease (new lesions and/or increase in size of existing lesions) after first-line chemotherapy. No more than 1 prior chemotherapy regimen was allowed. Patients were required to have had an objective response to initial chemotherapy treatment and then a recurrence >60 days after the last day of first-line chemotherapy administration. Patients were required to have bidimensionally measurable disease, which was defined by ≥1 lesion(s) that measured ≥1 × 1 cm. Patients needed an Eastern Cooperative Oncology Group performance status ≤2. Prior to entry on the study, required laboratory values included an absolute neutrophil count (ANC) ≥1500 cells/μL, a platelet count ≥100,000 cells/μL, bilirubin and serum creatinine levels within normal limits, an aspartate aminotransferase level ≤2 times the upper limit of normal, and a cardiac ejection fraction >45% determined by multiple-gated acquisition (MUGA) scan or echocardiogram. Patients with brain metastases had to have clinically controlled neurologic symptoms, which we defined as surgical excision and/or radiation therapy followed by 30 days of stable neurologic function and no evidence of central nervous system disease progression.

Exclusion criteria included prior radiation therapy to >50% of bone marrow, concurrent serious infections, any unstable or serious concurrent medical condition, brain or bone metastases as the only measurable site of disease recurrence, radiation therapy within the previous 30 days, a history of prior cancer (except patients who had received adequate treatment for basal cell carcinoma, squamous cell skin cancer, or any in situ carcinoma) unless they were disease-free for ≥5 years, and a first-line chemotherapy regimen that included doxorubicin or vincristine. Informed consent was obtained according to local Institutional Review Board requirements.

Study Design

This was an open-label, nonrandomized, multicenter study that was conducted in 2 stages with planned interim and final analyses. The sample sizes of 35 evaluable patients for the interim analysis and the potential for 92 additional evaluable patients for the final analysis were based on assumed response rates of 21% for the null hypothesis. The sample size to detect a response rate of 36% for the alternative hypothesis (or a 75% relative improvement from standard treatment) was calculated with 90% power at a 5% type I significance level. The null hypothesis rate was selected to be intermediate between the second-line therapy response rate observed for patients who were treated with CAV (17.3%) and the initially reported response rate for topotecan (24.3%).9 The study was to continue to the second stage if the interim analysis identified ≥9 confirmed responses among these first 35 evaluable patients.

VX-710 was given by continuous intravenous infusion (IV) for 72 hours at a dose of 120 mg/m2 per hour, as described previously.16 Doxorubicin (45 mg/m2) and vincristine (0.7 mg/m2 or 1.4 mg/m2) were given as IV bolus injections 4 hours after the start of the VX-710 infusion. The regimen was to be repeated every 21 days until evidence of disease progression, intolerable adverse events (AEs), or withdrawal of patient consent. Patients were to receive up to 6 cycles of treatment.

The first 3 patients were treated with a half dose of vincristine (0.7 mg/m2) to assess toxicity. The incidence of dose-limiting toxicities (DLTs) in these patients within the first cycle was used to determine whether the vincristine dose could be increased to the full dose (1.4 mg/m2 up to a maximum of 2 mg). DLT was defined as Grade 3 or 4 nonhematologic toxicities (according to World Health Organization [WHO] toxicity criteria); ANC <500 cells/μL for >7 days, febrile neutropenia, platelets <20,000 cells/μL, or grade 3 or 4 nausea and/or vomiting despite aggressive antiemetic treatment.

None of the first 3 patients had a DLT; therefore, the next 3 patients were treated with a full dose of vincristine (1.4 mg/m2 or a maximum of 2.0 mg). Subsequently, an additional 33 patients were treated with full-dose vincristine in the first stage of the study. Although the original study plan was to analyze only those patients who received full-dose vincristine, the patients who received half-dose vincristine ultimately were included in the efficacy analysis for the most conservative assessment of antitumor activity.

After 15 patients had been enrolled, the study was modified to include administration of prophylactic granulocyte-colony stimulating factor (G-CSF) at 5 μg/kg per day on Days 6 through 13 of each treatment cycle or prophylactic ciprofloxacin at 500 mg twice daily on Days 7 through 17 of each treatment cycle to prevent serious neutropenic infection. This modification was based on grade 4 neutropenia, which was noted during Cycle 1 in 58% of the first 15 patients enrolled, and 2 patient deaths from febrile neutropenia.

After 36 patients had been enrolled in the first stage of the study, an interim analysis was conducted. Enrollment in the study was stopped prior to Stage 2 based on <9 responses noted in Stage 1.

Pretreatment and Follow-up Evaluations

Seventeen patients had diagnostic biopsy samples that were sent to IMPATH, Inc., where immunostaining was performed. After optimization of assay conditions using 2 different antibodies to each protein, 1 antibody for each was chosen for patient samples in the study. Clone C219, immunoglobulin G2A (IgG2A) (8710-01, Signet Laboratories, Inc.) was used to detect Pgp expression; and clone MRP-r1, IgG (G8760-01; Signet Laboratories, Inc.) was used to detect MRP expression. Expression was analyzed in formalin-fixed, paraffin-embedded specimens for the site of reactivity, the degree of reactivity, and staining intensity (0–3+). Intensity was compared in specimens that expressed either protein in >20% of tumor cells by using human liver tissue as a positive control.

Patient history, physical examination, assessment of baseline symptoms, an electrocardiogram, and routine laboratory studies, including a pregnancy test for women of childbearing potential, were required within 2 weeks before study registration. Baseline tumor measurements were obtained by computerized tomography (CT) within 2 weeks before study entry. At some centers, 99mTc-sestamibi scans and MUGA/echocardiograms were performed within 2 weeks before study entry and on Day 3 of Cycle 1. Patients underwent a history and physical examination with laboratory tests in each cycle. Complete blood counts were obtained before the start of each cycle, at Days 10 thorough 14, and at Day 22. Tumor measurements were obtained after the second, fourth, and sixth cycles.

Response Criteria and Endpoints

Patients were considered evaluable for response assessment using WHO criteria17 if they completed at ≥1 cycle. 99mTc-sestamibi retention was used as a surrogate marker for enhanced retention of cytotoxic agents, as described previously.8 The image analysis was performed during baseline screening (within 2 weeks of starting study therapy) and on Day 3 (from 48 to 72 hours after initiation of the VX-710 infusion) of Cycle 1. The images were evaluated by a blinded reviewer to determine the uptake and retention within the liver, an organ that constitutively expresses both P-gp and MRP-1; in the heart as a reference negative value; and in imageable tumor lesions. A ratio of 99mTc-sestamibi retention in the target tissue compared with background was determined in scans that were acquired during baseline screening and on Day 3 of Cycle 1.

Blood samples were collected from the first 13 patients, and vincristine levels were measured at successive time points after administration of the drug (1 minute, 30 minutes, 2 hours, 4 hours, 6 hours, 8 hours, and 24 hours). However, pharmacokinetic analyses were not performed in this study, because vincristine levels were undetectable after 30 minutes in all of the 13 patients sampled using high-performance liquid chromatography. It is unclear whether this was caused by insensitivity of the assay in humans, but preclinical studies in dogs had shown no effect of VX-710 on the pharmacokinetics of vincristine and no effect on associated hematologic, neurologic, or other toxicities.

RESULTS

Patients

In total, 36 patients were enrolled in this study from December 1998 through December 2000. Patient demographics, prior treatment history, and disease characteristics are summarized in Table 1. Although this study had an unusually greater proportion of women than men, other attributes were consistent with other studies of patients with SCLC. Notably, the majority of patients had received prior radiation in addition to chemotherapy.

Table 1. Patient Demographic and Disease Characteristics
VariableNo. of patients (%)
  1. ECOG indicates Eastern Cooperative Oncology Group.

No. of patients treated36 (100)
Sex
 Men13 (36)
 Women23 (64)
Race
 Caucasian36 (100)
Median age [range], y58 [44–79]
Stage at diagnosis
 Limited stage23 (64)
 Extensive stage13 (36)
Prior radiation therapy24 (75)
ECOG performance status
 013 (36)
 119 (53)
 24 (11)

The 36 patients received a total of 108 treatment cycles of treatment. The mean dose per cycle was 99% of the maximum prescribed dose for VX-710, 97% for doxorubicin, and 95% for vincristine. The median number of cycles per patient was 2 (range, 1–8 cycles).

The most common reasons for treatment discontinuation were progression of SCLC (16 patients) or AEs (6 patients). Four additional patients died on study, and 6 patients completed all planned cycles. The 4 remaining patients were discontinued because of patient or physician preference.

Safety and Tolerability

All 36 patients were assessable for safety. Although all patients reported at ≥1 AE, most were considered mild to moderate in severity (grades 1 and 2). The most frequently reported AE was neutropenia, which occurred in 69% of patients (Table 2). Of 9 patients (25%) with grade 4 neutropenia, 2 patients died of complications related to neutropenic infection. Other hematologic toxicities also were significant. Forty-two percent of patients developed grade 1 to 3 anemia, and 28% of patients experienced thrombocytopenia, with 80% of those experiencing grade 3 or 4 toxicity. However, no significant bleeding events occurred. There were no significant changes from baseline for other laboratory parameters.

Table 2. Most Frequent Hematologic Toxicities by Severity
Adverse eventTotal no. (%)No. of patients, N = 36
Grade 1Grade 2Grade 3Grade 4
Neutropenia26 (72)34712
Anemia15 (42)4830
Thrombocytopenia10 (28)1261

Notable improvements in ANC values and the incidence of severe hematologic toxicity were observed among patients who were enrolled after the protocol was amended to require prophylactic use of G-CSF or ciprofloxacin beginning with Cycle 1 of study treatment. Only 38% of patients experienced grade 4 neutropenia in Cycle 1 after the amendment requiring prophylaxis, compared with 58% of patients who were enrolled prior to the amendment. In subsequent cycles, grade 4 neutropenia was reduced from 30% to 20%. After the amendment, no further deaths occurred from febrile neutropenia.

The most frequently reported nonhematologic AEs (>30% overall incidence) (Table 3) were asthenia (75%), nausea (53%); vomiting (44%); anorexia (42%); alopecia (39%); back pain, dyspnea, and constipation (36% each); arthralgias/myalgias and parasthesia (33% each); and diarrhea (31%). Nonhematologic AEs that were considered likely to be related to treatment included asthenia (53%), nausea (50%), constipation (44%), alopecia (42%), vomiting (39%), and anorexia (33%).

Table 3. Most Frequent Nonhematologic Toxicities by Severity
Adverse eventTotal No. (%)No. of patients, N = 36
Grade 1Grade 2Grade 3Grade 4
Asthenia27 (75)710100
Nausea19 (53)7930
Vomiting16 (44)5740
Anorexia15 (42)8610
Alopecia14 (39)7610
Constipation13 (36)5620
Back Pain13 (36)4540
Dyspnea13 (36)4612
Arthralgias/myalgias12 (33)4530
Paresthesia12 (33)7320
Diarrhea11 (31)5510
Abdominal pain10 (28)2530
Headache10 (28)5410
Insomnia7 (19)4210
Chest pain7 (19)0520
Confusion10 (28)3430

Three patients had >10% absolute decreases from baseline in their MUGA scan. Two of those decreases were considered to be related to study medication, and 1 decrease resulted in discontinuation from the study. Physical examination results and vital signs did not change substantially during the study.

In total, 17 patients experienced 20 life-threatening (grade 4) AEs, with neutropenia being the most predominant (12 patients). Two patients experienced grade 4 dyspnea. Other reported severe AEs that were not related clearly to study drug included 1 patient with pericardial effusion and tamponade, 2 patients with pulmonary emboli, and 1 patient with severe herpes zoster.

In addition to the 2 deaths on study because of neutropenic complications, 2 other patients died on study of disease progression, and 4 patients died within 30 days after study termination. Three of those 4 deaths were caused by disease progression, and the fourth death was caused by an infection that was not related to the study drug.

Efficacy

Of the 36 treated patients, 32 were considered evaluable for response. Four patients had no response data (1 patient withdrew consent during Cycle 1, 1 patient withdrew because of a severe AE during Cycle 1, and 2 patients died before the completion of Cycle 1 from infectious complications).

In total, 7 of 36 total patients (19%) achieved a confirmed partial response (PR) at the time of termination, which translates to 22% of evaluable patients (7 of 32 patients). Four of the PRs were confirmed by independent radiology review, and 3 were determined by radiology review at the study center (Table 4). One of the 7 patients who achieved a PR had a brain metastasis detected by MRI at study termination; the other 6 patients completed study treatment with a continuing response. Because patients' responses were censored at that time, an accurate calculation of response duration was not possible. However, 1 patient who completed 2 additional cycles (optional under study guidelines) had a response that was sustained for 3 years.

Table 4. Summary of Efficacy
VariableNo. of patients (%)
  • *

    Four patients did not have response data recorded because of discontinuation or death during Cycle 1.

  • The second value in parentheses indicates the percentage of 32 evaluable patients.

Total no. of patients36
No. of evaluable patients32*
Partial response7 (19, 22)
Unconfirmed partial response3 (8, 10)
Stable disease4 (11, 16)
Progressive disease18 (50, 56)

Four additional patients had notable tumor shrinkage at 1 measurement time point. Three patients had unconfirmed responses, with tumor measurements decreasing 80% to 90% after the Cycle 2 assessment. These 3 patients discontinued study treatment and did not have a subsequent confirmatory response assessment. One patient had 57% tumor shrinkage at the Cycle 2 assessment but had progressive disease at study termination. Four patients ended treatment with stable disease. Eighteen patients (56%) had disease progression, and the median TTP for those patients was 42 days.

Fifteen patients were evaluated by using 99mTc-sestamibi retention measured with planar and single-photon emission CT imaging. Scans were compared before treatment (during baseline screening) and on Day 3 of Cycle 1. Compared with baseline values, the administration of VX-710 increased retention of 99mTc-sestamibi in the liver by 1.3- to 3.2-fold in 10 of 15 patients. Five patients had no increase in 99mTc-sestamibi liver retention. Within tumor tissue, retention of 99mTc-sestamibi was increased for only 1 patient (by 1.4-fold); 13 patients had no increase, and 1 patient had no imaging data collected on Day 3.

Diagnostic biopsy specimens were obtained for 20 of the 36 patients enrolled on the study. The tumor cells in diagnostic specimens rarely expressed P-gp (3 of 20 patients; 15%). In contrast, 12 of 20 patients (60%) had detectable MRP-1 by immunohistochemistry. One patient expressed both proteins at moderate levels. These expression frequencies are similar to those reported in other studies4, 5, 18; however, unlike those studies, there was no clear correlation with expression of either protein in the current study and response to treatment. Among the 10 responders (7 confirmed and 3 unconfirmed), 2 patients had 1+ expression of P-gp, and 1 patient had 1+ expression of MRP: The remaining patients did not express either protein. Whether these proteins were up-regulated with the development of resistance to first-line treatment in unknown, because no patients consented to repeat biopsies prior to treatment with VX-710.

In total, 21 patients were seen at least once during the long-term follow-up period, which ranged from 0 months to 30 months. The median survival was 6 months (95% confidence interval, 4–7 months) (see Fig. 1).

Figure 1.

Kaplan–Meier survival curve of the patients on study.

DISCUSSION

The incidence and severity of the neutropenia observed in this study were similar to results from a study that used CAV to treat patients with recurrent SCLC.9 Grade 4 neutropenia was experienced by 71.7% of patients who received with CAV and in 51.4% of cycles with no treatment-related deaths. In addition, 4.8% of patients and 1.4% of cycles of CAV were associated with infection.13

However, the overall incidence of grade 4 neutropenia (with or without prophylactic G-CSF) that we observed with the VX-710 plus doxorubicin/vincristine combination was greater than expected based on results from a Phase I study. In that study, among 11 patients who received VX-710 120 mg/m2 per hour with doxorubicin 45 mg/m2, only 17% experienced grade 4 neutropenia.8

The addition of vincristine to the same VX-710/doxorubicin dose combination appeared to increase myelosuppression markedly in the current study. Vincristine pharmacokinetics data could not be evaluated effectively because of the inability to detect vincristine after 30 minutes in any of the patients tested. Therefore, it was not possible to determine whether a pharmacokinetic interaction could contribute, at least in part, to the increase in myelosuppression observed. Increased myelosuppression has been observed in other Phase II studies with VX-710, even after dose adjustments for pharmacokinetic interactions. Phase II studies with VX-710 in combination with paclitaxel for advanced breast cancer or ovarian cancer refractory to paclitaxel reported a 40% and 47% incidence of grade 4 neutropenia, respectively.19

An alternative explanation for the current observations is that CD34-positive myeloid precursor cells express P-gp, and it is possible that increased vincristine and doxorubicin retention in the presence of VX-710 may be a mechanism-based pharmacodynamic effect that contributes to myelosuppression.20–22 Finally, prior treatment history may have been a factor in the increased myelosuppression and clinical consequences for some patients on this study. Seventy-five percent of patients had received previous mediastinal radiation and, thus, began the study with potentially compromised bone marrow. This rate is higher than previous the rate reported in Phase II studies of CAV and topotecan, in which radiation frequency ranged from 36% to 62%3, 18; and it presumably is higher than the rate reported from a Phase I study of VX-710 and doxorubicin, which included multiple cancer types.12

Although this trial was not designed as a dose-finding study for the combination of doxorubicin, vincristine, and VX-710, an initial cohort of 3 patients was treated with half-dose vincristine. These patients had DLTs monitored for only 1 cycle, and no toxicities were noted. In retrospect, a Phase I component that incorporated more extensive DLT monitoring may have provided useful information about the hematologic toxicities that we observed in this study prior to implementing an amendment that required neutropenia prophylaxis. Without this prophylaxis, the combination could not be considered a safe regimen. After implementing this amendment, 4 of 21 patients (19%) did not complete >2 cycles because of treatment-related toxicities. This number is higher than the numbers reported from a study that compared topotecan with CAV11 but is comparable to the numbers reported from a recent study of currently used regimens in first-line therapy.23

The confirmed response rate in the current study of 19% was similar to the reported response rates to topotecan (24%) and CAV (18%) in a randomized study with similar inclusion criteria that enrolled patients with SCLC.3 Furthermore, the majority of these responses were sustained throughout the duration of treatment, with 1 patient showing a prolonged response of 3 years. The median survival in this population of 6 months was nearly identical to the median survival reported for both topotecan and CAV.1, 9

However, the proposed hypothesis for this study was that VX-710 would improve upon the objective response rate to the vincristine/doxorubicin combination. Several factors may have affected the effective inhibition of these MDR targets by VX-710. The proportion of diagnostic biopsy specimens that expressed P-gp was low (15%), and no specimens expressed high levels of P-gp (3+). MRP-1 expression was more frequent (60%), but only 10% of specimens expressed MRP-1 at high levels. Of the patients who responded, only 30% had any expression of either protein, and none had expression at a high level. Because pre-VX-710 biopsies were not available from patients in this study, the change in expression levels of P-gp and MRP-1 after first-line therapy could not be determined, which would have been more informative about the association of those levels with response to VX-710. In the absence of significantly up-regulated expression of MDR proteins, the addition of VX-710 would not necessarily be expected to improve the response rate to doxorubicin and vincristine.

Functional tumor imaging with 99mTc-sestamibi was another strategy we examined in this study to identify tumor lesions with a drug-accumulation defect, although only some centers were able to perform this study. The absence of significant 99mTc-sestamibi uptake in tumor tissue from patients in this study contrasts with Phase I data, in which a 20% to 60% increase in tumor retention of 99mTc-sestamibi was observed in 8 patients with imageable tumors given VX-710 at a similar dose and infusion schedule.10 The reason for the difference in results from the 2 studies with VX-710 are not clear. In contrast to previous studies that used this imaging modality in other solid tumors,16–18 a study of 23 patients with lung cancer (9 patients with SCLC and 15 patients with non-SCLC) showed no significant correlation between the percent retention of 99mTc-sestamibi and chemotherapeutic response,24 suggesting that the location of the tumor, local VX-710 tissue concentrations, or vascularity of the tumor all may affect the utility of 99mTc-sestamibi imaging. In addition, it is unknown how the levels of P-gp and MRP-1 in this study compared with the levels reported from the Phase I study, which included multiple tumor types. The lack of 99mTc-sestamibi uptake may reflect low MDR protein expression and, thus, may be the cause of the lack of efficacy observed with VX-710; however, because there were some therapeutic responses, alternatively, it may be an inappropriate biomarker in this setting.

One concern about the use of MDR inhibition in general is the effect of inhibiting only 1 or even 2 MDR modulators when tumors may express multiple MDR transporters with overlapping activities (for review, see Donnenberg and Donnenberg, 200525). Although VX-710 inhibits both P-gp and MRP, the expression of P-gp in SCLC is less common compared with the expression of MRP, as confirmed in this study. If alternative transporters are expressed in recurrent SCLC, then the MDR phenotype still may be unaffected despite the effective inhibition of P-gp and MRP-1 by VX-710.

In conclusion, although data suggest that acquired expression of P-gp or MRP1 may be associated with a poor response of SCLC to second-line therapy, the addition of VX-710 to doxorubicin and vincristine did not improve the response rate in this small study and was associated with significant hematologic toxicity. Although 7 of 36 total patients had PRs, we believed it was unlikely that 2 additional responses would be seen in 3 more evaluable patients to achieve the required 9 of 35 responses required to continue the study. Thus, based on the accumulated data, the study was stopped early. Although no further development is planned currently for this particular molecule, future evaluation of MDR inhibitors in SCLC should incorporate prestudy biopsies for analysis of MDR transporter expression to select patients who have a greater probability of achieving clinical benefit.

Ancillary