A Phase I/II Study of Erlotinib in Combination with the Anti

CORE Metadata, citation and similar papers at core.ac.uk

Provided by Elsevier - Publisher Connector ORIGINAL ARTICLE

A Phase I/II Study of Erlotinib in Combination with the Anti-Insulin-Like Growth Factor-1 Receptor Monoclonal Antibody IMC-A12 (Cixutumumab) in Patients with Advanced Non-small Cell Lung Cancer

Andrew Weickhardt, MBBS, DMedSc,* Robert Doebele, MD, PhD,* Ana Oton, MD,* Janice Lettieri, BSN,* DeLee Maxson, BS,* Michele Reynolds, BS,* Amy Brown, PhD,* Mary K. Jackson, AD,* Grace Dy, MD,† Araba Adjei, PhD,† Gerald Fetterly, PhD,† Xian Lu, MSc,‡ Wilbur Franklin, MD,* Marileila Varella-Garcia, PhD,* Fred R. Hirsch, MD, PhD,* Murry W. Wynes, PhD,* Hagop Youssouﬁan, MD,§ Alex Adjei, MD, PhD,† and D. Ross Camidge, MD, PhD*

growth factor-1 levels as seen with other insulin-like growth fac- Introduction: This phase I/II study evaluated the safety and anti- tor-1R inhibitors. tumor effect of the combination of erlotinib with cixutumumab, a Conclusions: The combinations of cixutumumab at 6 mg/kg every recombinant fully humanized anti-insulin-like growth factor-1 re- 7 days and 15 mg/kg every 21 days and full-dose erlotinib are not ceptor IgG1 monoclonal antibody, in advanced non-small cell lung tolerable in unselected patients with NSCLC, as measured by DLT. cancer (NSCLC). Cixutumumab at 5 mg/kg every 7 days was tolerable per DLT, but Methods: Patients with advanced NSCLC were treated in an initial dose delays were common. Efficacy in unselected patients with safety-lead and drop-down cohorts using erlotinib 150 mg/d with NSCLC seems to be low. cixutumumab 6 or 5 mg/kg on days 1, 8, 15, and 22 in 28-day cycles (cohorts 1 and 2). Emerging pharmacokinetic data led to an addi- Key Words: Non-small cell lung cancer, IGF1R monoclonal anti- tional cohort (3 ϩ 3 design) with cixutumumab at 15 mg/kg on day body, EGFR, Metastatic disease. 1 in 21-day cycles (cohort 3). (J Thorac Oncol. 2012;7: 419–426) Results: Eighteen patients entered the study (6 at 6 mg/kg, 8 at 5 mg/kg, and 4 at 15 mg/kg), with median age of 65 years. Four of six patients at 6 mg/kg experienced dose-limiting toxicities (DLTs), ung cancer causes more than 150,000 deaths annually in whereas at 5 mg/kg, one of eight patients experienced DLT but three Lthe United States.1 Epidermal growth factor receptor ty- of eight patients still required a dose delay during cycle 1. At 15 rosine kinase inhibitors (EGFR-TKIs) produce responses in mg/kg every 21 days, two of four patients experienced DLTs. In all patients with non-small cell lung cancer (NSCLC) with cohorts, DLTs were either G3 rash or fatigue. Five patients had EGFR-activating mutations; however, all responders eventu- stable disease as best response and 14 patients had progressive ally develop resistance, most commonly because of the emer- disease. The median progression-free survival was 39 days (range gence of a secondary T790M mutation or mesenchymal-epithe- 21–432ϩ days). Biomarkers analyses showed a trend toward better lial transition factor (cMET) amplification.2–4 Responses in progression-free survival seen with higher free baseline insulin-like EGFR wild-type (EGFR-WT) patients also occur but are rare.5 EGFR-TKIs in unselected patients produce improvements in the second- and third-line setting in progression-free survival (PFS) *University of Colorado Cancer Center, Aurora, Colorado; †Roswell Park and overall survival (OS) compared with placebo, although the Cancer Institute, Buffalo, New York; ‡Department of Biostatistics and gains are modest.6 Informatics, Colorado School of Public Health and University of Colo- rado, Aurora, Colorado; and §Imclone, Bridgewater, New Jersey. In vitro and in vivo studies have implicated the insulin- Disclosure: Dr. Andrew Weickhardt and Dr. D. Ross Camidge have received like growth factor receptor (IGF-1R), its ligands IGF I and II, honorarium from Pfizer in the previous 12 months. No other authors have and its related downstream signaling in the development and relevant affiliations or financial involvement with any organization or progression of cancer.7 Furthermore, the IGF-1R pathway is entity with a financial interest in or financial conflict with the subject matter or materials discussed in the article. implicated as an alternate possible cause of acquired resis- Address for correspondence: Andrew Weickhardt, MBBS, DMedSc, Uni- tance to EGFR-targeted therapy.8,9 Preclinical work demon- versity of Colorado Cancer Center, Aurora, CO 80045. E-mail: strates that the combination of an IGF-1R inhibitor with an [email protected] EGFR-TKI is synergistic in both NSCLC cancer cell lines Copyright © 2012 by the International Association for the Study of Lung Cancer that are EGFR-WT and those with mutations in EGFR exons ISSN: 1556-0864/12/0702-0419 19 and 21 (EGFR mutation [MT]).10–15

Journal of Thoracic Oncology • Volume 7, Number 2, February 2012 419 Weickhardt et al. Journal of Thoracic Oncology • Volume 7, Number 2, February 2012

FIGURE 1. Dose cohorts in the phase I trial.

Cixutumumab (previously IMC-A12; Imclone; Bridge- hematopoietic, hepatic, and renal function. Exclusion criteria water, NJ) is a fully humanized recombinant IgG1/␭ mono- included previous exposure to other EGFR or IGF-1R inhib- clonal antibody with high affinity to the extracellular domain itors; chemotherapy within five half-lives of the agent or 28 of IGF-1R with a mean effective concentration of 0.04 days if the half-life was unknown or was a monoclonal nmol/liter, acting as an antagonist of the IGF-I and IGF-II antibody; radiotherapy within 28 days; major surgery within ligand binding.16 In vitro, cixutumumab treatment induced 4 months of the study; ongoing previous treatment-related apoptosis in sensitive human tumor cell lines and elicited side effects/AEs that were Նgrade 2; known brain metastases both tumor growth inhibition and tumor regression across a unless adequately treated and stable off anticonvulsants and broad range of human tumor xenograft models.16 In combi- steroids; pregnancy; and poorly controlled diabetes mellitus nation with EGFR-TKI therapy, cixutumumab showed in- or other uncontrolled intercurrent illness. creased cytoxicity compared with either agent alone in cell The protocol was approved and monitored by all local lines and in xenografts.16 institutional review boards. All patients provided written Cixutumumab monotherapy, administered intravenously consent before enrollment. (IV) every 7 days, is tolerated at doses of up to 10 mg/kg.17 In a phase II trial of 10 mg/kg of IV cixutumumab every 7 Study Design and Drug Dosing and days in pretreated patients with chemotherapy refractory Administration advanced colon cancer, there were few grades 3 and 4 side Patients were enrolled at two institutions into one of the effects attributable to the drug, with asymptomatic hypergly- three cohorts (Figure 1). To assess safety and tolerability, the cemia the most frequent drug-related adverse event (AE).17 initial safety-lead and drop-down cohort used erlotinib 150 The mean terminal elimination half-life of cixutumumab is mg orally daily with cixutumumab 6 mg/kg (cohort 1) or 5 148 hours, with trough concentrations above those shown mg/kg (cohort 2) IV on days 1, 8, 15, and 22 in 28-day cycles. effective in xenograft models (data on file, Imclone). In a Tolerability required the completion of one cycle in 8 of 10 subsequent phase I study, cixutumumab was shown to be patients without dose-limiting toxicity (DLT). Emerging tolerable up to and including 20 mg/kg when given every 21 pharmacokinetic data related to 21-day dosing of cixutu- days, although these data were not available until midway mumab led to an amendment and enrollment of a third cohort through accrual for this clinical trial (data on file, Imclone). with erlotinib 150 mg orally daily with cixutumumab 15 The primary objectives of this study were to explore the mg/kg IV in 21-day cycles (cohort 3). In this cohort, for safety and efficacy of the combination of cixutumumab with toxicity and feasibility assessment, a standard 3 ϩ 3 phase I full-dose erlotinib, before planned expansion into a random- design was used. At least three patients were required to have ized phase II study comparing full-dose erlotinib with the completed one cycle for dose escalation to occur to a planned combination. cohort 4 (cixutumumab 20 mg/kg IV in 21-day cycles). If there was one DLT among the first three patients, an addi- PATIENTS AND METHODS tional three patients were enrolled at the same dose level. If two DLTs were observed in up to six patients, then dosing at Patient Selection that level was stopped. The highest tolerated dose regimen Patients with histologically or cytologically docu- was to be declared the maximum tolerated dose (MTD). mented locally advanced or metastatic NSCLC, age older DLT was defined as any treatment delay of Ն7 days than or equal to 18 years, and those who had progressed on or because of any drug-related toxicity within cycle 1, or any after receiving platinum-containing chemotherapy were eli- nonhematological toxicity Նgrade 3 within cycle 1, exclud- gible. A later amendment allowed the enrollment of patients ing rash, diarrhea, nausea, hyperglycemia, hypocalcemia, with EGFR-MTs who were untreated. Other criteria included hypokalemia, hypomagnesemia, hyponatremia, hypophos- Eastern Cooperative Oncology Group performance status phatemia, or hypercholesterolemia if controllable within 7 Յ2; life expectancy Ն3 months; presence of evaluable or days of holding drug, and no dose reductions occurred. For measurable disease as defined by the RECIST (version 1.0); both erlotinib and cixutumumab, dose intensity (%) for each fasting serum glucose less than 120 mg/dL; and adequate patient was calculated by the amount of drug delivered (mg)

420 Copyright © 2012 by the International Association for the Study of Lung Cancer Journal of Thoracic Oncology • Volume 7, Number 2, February 2012 Phase I/II Study of Erlotinib divided by the proposed dose delivered (mg) over the dura- Clinical Laboratory Improvement Amendments-certified Col- tion they were on study, accounting also for protocol-defined orado Molecular Correlates Laboratory.18 dose reductions if there was more than 10% change in weight Free IGF-1 in the pretreatment-collected serum was during treatment. measured using an enzyme-linked immunosorbent assay kit Cixutumumab was supplied in sterile 250 mg (50 ml) from Beckman-Coulter Diagnostic System Laboratories single-use vials by the Division of Cancer Treatment and (Webster, TX, Cat# DSL-10-9400). This kit requests ethyl- Diagnosis, National Cancer Institute, under a collaborative enediaminetetraacetic acid plasma samples; however, only agreement between Imclone Systems and the Division of serum samples were collected and used in this study. Follow- Cancer Treatment and Diagnosis. Erlotinib was commercially ing the kit’s instructions, no dilution of samples resulted in available (OSI Pharmaceuticals, Farmingdale, NY). poor recovery of the standard exogenously added to a patient sample and low free IGF-1 values for the patient samples, Dose Modification suggesting an inhibitory effect within the sample. Therefore, Dose modification of cixutumumab and erlotinib was samples were diluted 1:2 in the zero standard buffer, which allowed according to the protocol. In the event of grade 3 or resulted in accurate recovery of the spiked standard and 4 DLT, both cixutumumab and erlotinib were held until the higher levels of free IGF-1. Further dilutions resulted in no toxicity resolved to grades 0 and 1 (or baseline grade if higher change of the rank order of the specimens. The concentration than grade 1). Patients with greater than a 3-week delay in of free IGF-1 reported accounts for the dilution and repre- treatment were removed from study unless the delay was sents the concentration within the serum sample. deemed by the principal investigator to be unrelated to study In addition, archived patient paraffin-embedded tumor drug and the patient had not experienced clinical progression blocks were retrieved where possible, and serum was collected of disease. Individual dose modification of cixutumumab and for pharmacokinetic assays at cycle 1 on day 8 and cycle 2 on erlotinib was allowed and performed to the next lowest dose day 1. Supplemental materials (http://links.lww.com/JTO/A199) cohort at investigator discretion. If a patient discontinued include the methods used for silver in situ hybridization assay from cixutumumab because of cixutumumab-related toxici- and fluorescent in situ hybridization assay for insulin growth ties, erlotinib could continue if appropriate. factor receptor (IGFR) and immunohistochemical assessment of IGFR1 expression. Safety and Efficacy Assessments Statistical Analyses Patients were evaluated with history, physical exami- Descriptive statistics summarized the demographics, nation, vital signs, and blood tests weekly through cycle 1 and safety data, and efficacy outcomes. Visual representation of on day 1 of each subsequent cycle. Incidence and severity of PFS, response, and free IGF-1 levels and PFS was performed AEs were collected at each study visit and graded according on Prism 5.0 software (Graphpad Software, La Jolla, CA). to National Cancer Institute Common Toxicity Criteria for Kaplan-Meier curves were generated for time to progression Adverse Events version 3. Patients were evaluable for cohort within wild-type and mutant EGFR groups, respectively. Cox dose escalation/de-escalation decision making either if they proportional hazards regression was used to obtain hazard experienced DLTs in cycle 1 or if they had completed 24 of ratio estimates for the EGFR-mutant group versus the WT the planned 28 days (85%) dosing of erlotinib and three of the group. Statistical analyses were performed using SAS/BASE four planned days of weekly dosing of cixutumumab (75%) and SAS/STAT software, version 9.2 of the SAS System for in cycle 1 in cohorts 1 and 2 in the absence of DLTs. In cohort Windows (SAS Institute Inc., Cary, NC). 3, patients were evaluable for tolerability if they had completed 18 days (85%) dosing of erlotinib and had received the planned day 1 dose of cixutumumab. RESULTS Baseline evaluations and radiologic evaluations were Patient Characteristics and Trial Overview conducted within 4 weeks of study entry. Disease was as- Between December 2008 and October 2010, 18 patients sessed after every second cycle, with response evaluated were enrolled across cohorts 1 to 3. Because of DLT issues, according to the RECIST criteria (version 1.0). Patients were cohort 4 was not explored (see below). Because of concerns considered evaluable for objective response if they had mea- about the achievable dose intensity at the MTD and the surable disease at baseline, had received at least one dose of apparent low efficacy in unselected patients with NSCLC (see study related therapy, and had a second scan. subsequent sections), in conjunction with similar concerns PFS was defined from first dose of drug until time of arising with other IGF-1R inhibitors in combination with progression (as measured by time of computed tomogra- erlotinib, the expansion part of the trial beyond the dose phy [CT] assessment by RECIST demonstrating progres- escalation phase was cancelled. The characteristics of the sion) or death. enrolled patients are listed in Table 1. Biomarker Assessment Safety EGFR-MT analysis was performed where possible The most frequent AEs were fatigue (72%), acne- from archived tumor specimens with either bidirectional iform rash (56%), diarrhea (50%), anorexia (28%), and sequencing of polymerase chain reaction-amplified DNA nausea (28%; Table 2). The majority of these were grade 1 (Roswell Park, Buffalo, NY) or performed through the on-site or 2 severity. Four patients in cohort 1 experienced DLTs

TABLE 1. Patient Demographics and Clinical Characteristics Cohort 1 Cohort 2 Cohort 3 Cixutumumab, Cixutumumab, Cixutumumab, All Patients 6 mg/kg IV q7 5 mg/kg IV q7 15 mg/kg IV q21 (4 ؍ N) (8 ؍ N) (6 ؍ N) (18 ؍ N) Gender Male 8 4 2 2 Female 10 2 6 2 Age, median (minimum– 65 (48–77) 67 (48–77) 66 (58–71) 59 (50–67) maximum), yr Race White 14 6 5 3 Others 4 0 3 1 Ever smokers 14 3 7 4 ECOG performance status 05212 113472 Prior chemotherapy (no. of regimens) 01010 15410 24130 Ͼ38134 Prior radiotherapy 6 2 2 2 Prior surgery 4 1 2 1 Tumor types Nonsquamous 16 5 7 4 Squamous 2 1 1 0 EGFR status EGFR mutation (19–21) 3 1 2 0 EGFR wild type 13 5 5 3 Not assessed 2 0 1 1

IV, intravenous; ECOG, Eastern Cooperative Oncology Group; EGFR, epidermal growth factor receptor.

(three with grade 3 fatigue and one with grade 3 acneiform Dose Modifications rash, which failed to resolve within 7 days to Յgrade 1 Fourteen patients (78%) had a delay in treatment or severity on holding the drug), and this dose level in the dose modification while on study (Table 2). The most com- q7-dosing regimen was declared nontolerated. One patient mon reasons for delay/dose modification were fatigue (five in cohort 2 experienced a DLT (grade 3 acneiform rash, patients) and rash (five patients). In cohort 1, the mean dose which failed to resolve within 7 days to Յgrade 1 severity intensity of cixutumumab was 72% and of erlotinib was 73%. on holding the drug), and this dose level was declared the In cohort 1, five of six patients (83%) required a dose MTD for the every 7 day-dosing regimen. There were two reduction in the first cycle. In cohort 2, the mean dose patients in cohort 3 who experienced a DLT (one with intensity of cixutumumab was 88% and of erlotinib was 78%. grade 3 fatigue and one with grade 3 acneiform rash that In cohort 2, three of eight (38%) patients required a dose failed to resolve within 7 days to Յgrade 1 severity on reduction in the first cycle. In cohort 3, the mean dose holding the drug), and this dose level in the every 21 intensity of cixutumumab was 95% and of erlotinib was 79% day-dosing regimen was also declared nontolerated. Only and two of four (50%) patients required a dose reduction in three patients developed hyperglycemia (all grade 2 sever- the first cycle. Seventeen patients (94%) maintained the ity); none required treatment. two-drug combination until progression; one patient received Three patients, all from cohort 2, withdrew from the one cycle of erlotinib alone before progressing after eight study. One patient withdrew after 1 cycle because of a grade cycles. 2 infusion reaction to cixutumumab. One patient with an EGFR-MT withdrew consent after eight cycles while re- Antitumor Activity sponding to therapy to have resection of oligometastatic All 18 patients were evaluable for response (Table 3; disease, and one patient with an EGFR-MT withdrew consent Figure 2). The median PFS was 39 days (range 21–432ϩ after nine cycles while responding to therapy because of days). Of the 18 patients, 5 patients achieved stable disease patient preference. and 13 patients had progressive disease. There were no

TABLE 2. Adverse Events and Treatment Exposure Cohort 1 Cohort 2 Cohort 3 All Patients, Cixutumumab, Cixutumumab, Cixutumumab, All Grades 6 mg/kg q7 5 mg/kg q7 15 mg/kg q21 (4 ؍ N) (8 ؍ N) (6 ؍ N) 18 ؍ N AE N % G1/2 G3/4 G1/2 G3/4 G1/2 G3/4 Fatigue 13 (4) 72 2 3 (3) 5 2 1 (1) Acneiform rash 10 (3) 56 1 1 (1) 2 2 (1) 2 2 (1) Diarrhea 9 50 2 4 3 Anorexia 5 28 2 2 1 Nausea 5 28 4 1 Dry skin 4 22 1 1 2 Mucositis 4 22 3 1 Sinusitis 4 22 3 1 Taste alteration 3 17 3 Hyperglycemia 3 17 1 2 Rigors/chills 2 11 1 1 Nail changes 2 11 1 1 Dehydration 2 11 1 1 Constipation 2 11 1 1 Nausea 2 11 2 Heartburn 2 11 2 Renal failure 2 11 1 1 Muscle cramps 2 11 2 Dizziness 2 11 2

All adverse events (AEs) are reported if more than 10%, or if grade 3 or 4. There were no grade 5 toxicities. Dose-limiting toxicities (number of occurrences) are indicated by parenthesis.

TABLE 3. Treatment Summary Cohort 1 Cohort 2 Cohort 3 All Patients, Cixutumumab, Cixutumumab, Cixutumumab, mg/kg, q7 5 mg/kg q7 15 mg/kg q21 6 ,18 ؍ N (4 ؍ N) (8 ؍ N) (6 ؍ N (%) (N Stable disease 5 (28) 1 4a 0 Progression 13 (72) 5b 44 Median PFS (d) 39 32 46 40 Range (d) 21–432ϩ 21–112 21–432ϩ 21–40

a Includes two patients with EGFR-MT. b Includes I patient with EGFR-MT. PFS, progression-free survival; EGFR-MT, epidermal growth factor receptor mutation. confirmed responses to therapy. One patient remained on patients with EGFR-WT tumors was 37 days (range 21– treatment with ongoing stable disease after more than 14 432ϩ). One patient remained on study (432ϩ days), proven months of treatment. Seven patients had restaging of their to be EGFR-WT by direct sequencing. disease with CT scans before the protocol-defined time at the Thirteen of 18 patients had pretreatment free IGF-1 end of cycle 2, three patients because of clinical suspicion of plasma samples collected. The median PFS in the 13 of 18 progression, and four patients because intercurrent illness patients evaluable for biomarker analysis of free-IGF-1 prompted the earlier use of CT imaging. (fIGF-1) was 40 days, which was similar to that of the overall Biomarker Analyses trial population (39 days). Correlation of efficacy (measured Retrospective analysis of the EGFR-MT status and by PFS) and tolerability (using dose intensity of cixutu- correlation with efficacy was undertaken. Three patients were mumab as a surrogate marker) outcomes with baseline fIGF-1 found to have EGFR-MTs of 16 patients assessable. Median indicated a nonstatistically significant trend for benefit in duration on study in the EGFR-mutant group was 217 days terms of PFS in patients with the highest quartile of baseline (range 27–245 days). Median duration of study in the 13 fIGF-1, but no correlation with achievable dose intensity of

FIGURE 2. A, Progression-free survival and (B) waterfall plot of best response.

or have sensitizing mutations to EGFR-TKI therapy.10,12,13,15 Cixutumumab is an antibody directed against the IGF-1R that has been demonstrated to be safe up to 10 mg/kg IV every 7 days and has been used at this dose in a randomized phase II trial in conjunction with EGFR-monoclonal antibodies in metastatic colorectal cancer with few grade 3 and 4 AEs.17 When used as a single agent, the side effects attributable to cixutumumab have been fatigue and hyperglycemia.17,19,20 Similar AEs were reported with an alternate IGF-1R monoclonal antibody figitumumab (Pfizer; La Jolla, CA); however, addi- tional AEs such as increased rates of dehydration, infection, and cardiovascular toxicity were observed when figitumumab was combined with chemotherapy compared with chemotherapy alone.21 In our single-arm study, the combination had a rela- tively high level of EGFR-related side effects including acneiform rash and diarrhea and a significant association with grades 3 and 4 fatigue (Table 2). Interestingly, neither the randomized trial of cixutumumab in combination with the EGFR monoclonal antibody cetuximab (Imclone) nor the randomized trial of R1507 with erlotinib reported any significant higher increased FIGURE 3. A, Progression-free survival and (B) dose inten- incidence of rash or fatigue.17,22 sity of patients by baseline free IGF-1 quartiles. The middle Activity was limited with no objective responses, and 7 line of the box plots represents the median. The box repre- of 18 patients progressed before the completion of two cycles sents the upper and lower quartiles, and the line extends (Table 3). The lack of benefit from the combination of this from the lowest to the highest PFS values. †Includes two class of antibody with erlotinib in an unselected NSCLC EGFR-MT patients and one EGFR unknown patient. IGF, insulin-like growth factor; PFS, progression-free survival; EGFR- population was confirmed by a recently presented random- MT, EGFR mutation. ized trial of the combination of erlotinib 150 mg orally daily with an another anti-IGF-1R monoclonal antibody—R1507 (Hoffmann La Roche, Switzerland), which failed to demon- cixutumumab (Figure 3). There were no patients with EGFR- strate an increase in response rate, PFS, or OS with the MTs in the highest quartile of baseline fIGF-1. Other demo- addition of the IGF-1R monoclonal antibody to erlotinib.22 In graphic differences between the groups are summarized in the addition, in 2010, a press release announced that a large phase supplementary material (http://links.lww.com/JTO/A199). III study (ADVIGO 1018) in nonadenocarcinoma NSCLC, comparing full-dose erlotinib to erlotinib plus figitumumab, DISCUSSION was closed early after concluding that the combination was The IGF ligand, receptor and related downstream path- unlikely to fulfill its primary end point of improving OS ways have been implicated as important drivers of oncogenic relative to erlotinib alone. The discrepancy between preclin- growth.7 Preclinical work has demonstrated synergy with ical data demonstrating synergy and these clinical trials EGFR-TKIs in both NSCLC cell lines that are both wild type maybe because of toxicity preventing achieving optimum

424 Copyright © 2012 by the International Association for the Study of Lung Cancer Journal of Thoracic Oncology • Volume 7, Number 2, February 2012 Phase I/II Study of Erlotinib dosing for synergy or the failure of preclinical models to expand the trial in unselected patients with NSCLC and account for complex in vivo biological feedback loops. proceed to a randomized comparison to single-agent erlo- In our study, although not randomized, there seemed to tinib. The combination was tolerable with a dose of cixutu- be little benefit in an otherwise unselected EGFR-WT popu- mumab at 5 mg/kg IV weekly as judged on DLT criteria, but lation (median PFS 37 days). However, the patient with the many patients (3/8) in this cohort still required dose reduc- longest PFS on study (432ϩ days) was EGFR-WT (Figure 2). tions in the first cycle, and no patients experienced a partial The patient was confirmed EGFR-WT by direct sequencing, response. Although EGFR-MT patients had a longer mean and while a missed mutation is possible, the fact that the PFS than EGFR-WT patients, it is unclear, in the absence of patient did not manifest a dramatic response per RECIST a randomized study, whether EGFR-MT patients are deriving raises the possibility that a rare and as yet undefined group any more benefit than would be expected from exposure to an within the EGFR-WT population may be deriving particular EGFR-TKI alone. Of note, the patient with the longest PFS benefit from either or both drugs in the combination. This on the study was EGFR-WT. Given the difficulties seen with patient did not have baseline free IGF-1 levels assessed. The several other anti-IGF-1R monoclonal antibodies when com- PFS in the EGFR-MT group was longer (median PFS 217 bined with full-dose erlotinib in NSCLC, further develop- days) but not dramatically different from that expected from ment of the combination in otherwise unselected patients is EGFR-MT patients treated with EGFR-TKI monotherapy.5,23 not recommended. In addition, given the negative results of a In the absence of a randomized study, it is impossible to phase III trial of an anti-IGF-1R agent used in combination determine whether the addition of cixutumumab was adding with chemotherapy in NSCLC,21 the future of these agents in or subtracting from the benefit of an EGFR-TKI alone in the other combinations in unselected patients with NSCLC is EGFR-MT population. uncertain. However, as several studies including ours have With regard to groups that may be deriving maximal noted a trend toward patients with elevated baseline free benefit from IGF-1R inhibitors, a previous biomarker analy- IGF-1 levels having longer PFS than those with lower lev- 21,22,24 sis within a randomized phase II study of carboplatin and els, this biomarker may represent a means of identify- paclitaxel plus or minus figitumumab correlated increased ing patients who will develop the most benefit from such baseline free IGF-1 levels with the largest improvement in agents in the future. PFS from the addition of figitumumab.24 A similar pattern was also noted in the phase III study of figitumumab in ACKNOWLEDGMENTS combination with carboplatin and paclitaxel in a nonadeno- Supported by National Institutes of Health Cancer carcinoma population.21 The reason for this correlation is Therapy Evaluation Program (NCI 8148). unknown, but may be due to tumors that develop in patients REFERENCES with high free IGF-1 developing an addiction to IGF-1 1. Howlader N, Noone AM, Krapcho M, et al. SEER cancer statistics 24 receptor signaling. Given the caveat of the small numbers review, 1975–2008. National Cancer Institute, Bethesda, MD. involved, within our study a comparable trend was also 2. Doebele RC, Oton AB, Peled N, et al. New strategies to overcome observed, with there being a higher median PFS (83 days) limitations of reversible EGFR tyrosine kinase inhibitor therapy in among patients within the highest quartile of free baseline non-small cell lung cancer. Lung Cancer 2010;69:1–12. 3. Pao W, Miller VA, Politi KA, et al. Acquired resistance of lung IGF-1 compared with the other quartiles (Figure 3) suggest- adenocarcinomas to gefitinib or erlotinib is associated with a second ing that this may be a class effect.21,22,24 Whether free IGF-1 mutation in the EGFR kinase domain. PLoS Med 2005;2:e73. levels represent a biomarker for who will derive the most 4. Engelman JA, Zejnullahu K, Mitsudomi T, et al. MET amplification benefit from such drugs in terms of a direct correlation with leads to gefitinib resistance in lung cancer by activating ERBB3 signaling. Science 2007;316:1039–1043. activity against the tumor, or simply a marker for who it is 5. Mok TS, Wu YL, Thongprasert S, et al. Gefitinib or carboplatin-paclitaxel safest to give these agents to is currently unclear. Of note, in pulmonary adenocarcinoma. N Engl J Med 2009;361:947–957. within our small study, there was no apparent interaction 6. Shepherd FA, Rodrigues Pereira J, Ciuleanu T, et al. Erlotinib in between free IGF-1 and achievable dose intensity (as a previously treated non-small-cell lung cancer. N Engl J Med 2005;353: 123–132. surrogate for overall toxicity). Either way, free IGF-1 levels 7. Jacobs CI. A review of the role of insulin-like growth factor 2 in should be explored further for their potential for optimal malignancy and its potential as a modifier of radiation sensitivity. Clin patient preselection for IGF-1R inhibitors in the future. Oncol (R Coll Radiol) 2008;20:345–352. Data on the predictive value of other potential biomark- 8. Kato Y, Wynes MW, Mascaux C, et al. IGF-1R is a predictive biomarker for EGFR TKI (gefitinib) resistance in non-small cell lung cancer using IHC ers, including IGFR fluorescent in situ hybridization and and AQUA technology. 14th World Conference on Lung Cancer, Amster- silver in situ hybridization expression, and IGFR protein dam, July 3–7, 2011;1. expression determined by IHC were uninformative because 9. Guix M, Faber AC, Wang SE, et al. Acquired resistance to EGFR of the low number of patients with sufficient material for tyrosine kinase inhibitors in cancer cells is mediated by loss of IGF- binding proteins. J Clin Invest 2008;118:2609–2619. analysis, limiting conclusions that can be made to their 10. Morgillo F, Kim WY, Kim ES, et al. Implication of the insulin-like relative utility in predicting benefit from the combination of growth factor-IR pathway in the resistance of non-small cell lung cancer cixutumumab and erlotinib (see Supplemental Material, cells to treatment with gefitinib. Clin Cancer Res 2007;13:2795–2803. http://links.lww.com/JTO/A199). 11. Choi YJ, Rho JK, Jeon BS, et al. Combined inhibition of IGFR enhances the effects of gefitinib in H1650: a lung cancer cell line with EGFR Despite promising preclinical activity, the combination mutation and primary resistance to EGFR-TK inhibitors. Cancer Che- of erlotinib and cixutumumab in this trial failed to demon- mother Pharmacol 2010;66:381–388. strate either the safety profile or the activity necessary to 12. Camirand A, Zakikhani M, Young F, et al. Inhibition of insulin-like

growth factor-1 receptor signaling enhances growth-inhibitory and pro- FISH and mutation) to predict response to targeted therapy in NSCLC. apoptotic effects of gefitinib (Iressa) in human breast cancer cells. Breast The United States and Canadian Academy of Pathology Annual Meet- Cancer Res 2005;7:R570–R579. ing, Boston, March 7–13, 2009, Poster 233. 13. Chakravarti A, Loeffler JS, Dyson NJ. Insulin-like growth factor recep- 19. Abou-Alfa GK, Gansukh B, Chou JF, et al. Phase II study of cixutu- tor I mediates resistance to anti-epidermal growth factor receptor therapy mumab (IMC-A12, NSC742460; C) in hepatocellular carcinoma (HCC). in primary human glioblastoma cells through continued activation of ASCO Meeting Abstracts 2011;29:4043. phosphoinositide 3-kinase signaling. Cancer Res 2002;62:200–207. 20. Higano CS, Yu EY, Whiting SH, et al. A phase I, first in man study of 14. Desbois-Mouthon C, Baron A, Blivet-Van Eggelpoe¨l MJ, et al. Insulin- weekly IMC-A12, a fully human insulin like growth factor-I receptor like growth factor-1 receptor inhibition induces a resistance mechanism IgG1 monoclonal antibody, in patients with advanced solid tumors. via the epidermal growth factor receptor/HER3/AKT signaling pathway: ASCO Meeting Abstracts 2007;25:3505. rational basis for cotargeting insulin-like growth factor-1 receptor and 21. Jassem J, Langer CJ, Karp DD, et al. Randomized, open label, phase III epidermal growth factor receptor in hepatocellular carcinoma. Clin trial of figitumumab in combination with paclitaxel and carboplatin Cancer Res 2009;15:5445–5456. versus paclitaxel and carboplatin in patients with non-small cell lung 15. Jones HE, Goddard L, Gee JM, et al. Insulin-like growth factor-I cancer (NSCLC). ASCO Meeting Abstracts 2010;28:7500. receptor signalling and acquired resistance to gefitinib (ZD1839; Iressa) 22. Ramalingam SS, Spigel DR, Steins M, et al. Randomized, double-blind, in human breast and prostate cancer cells. Endocr Relat Cancer 2004; phase II study of erlotinib in combination with placebo or R1507, a 11:793–814. monoclonal antibody to insulin-like growth factor receptor-1 (IGF-1R), 16. Burtrum D, Zhu Z, Lu D, et al. A fully human monoclonal antibody to for advanced-stage non-small cell lung cancer (NSCLC). ASCO Meeting the insulin-like growth factor I receptor blocks ligand-dependent signal- Abstracts 2011;29:7527. ing and inhibits human tumor growth in vivo. Cancer Res 2003;63: 23. Inoue A, Kobayashi K, Usui K, et al. First-line gefitinib for patients with 8912–8921. advanced non-small-cell lung cancer harboring epidermal growth factor 17. Reidy DL, Vakiani E, Fakih MG, et al. Randomized, phase II study of receptor mutations without indication for chemotherapy. J Clin Oncol the insulin-like growth factor-1 receptor inhibitor IMC-A12, with or 2009;27:1394–1400. without cetuximab, in patients with cetuximab- or panitumumab-refrac- 24. Gualberto A, Hixon ML, Karp DD, et al. Pre-treatment levels of tory metastatic colorectal cancer. J Clin Oncol 2010;28:4240–4246. circulating free IGF-1 identify NSCLC patients who derive clinical 18. Lucas D, Varella-Garcia M, Camidge DR. Triple-platform testing (IHC, benefit from figitumumab. Br J Cancer 2011;104:68–74.