Phase II study of dasatinib in the treatment of head and neck squamous cell carcinoma (HNSCC)

Abstract

Purpose

Treatment options for patients with advanced HNSCC are scarce. This phase II study was conducted to evaluate the safety, tolerability, PK, and efficacy of dasatinib in this setting.

Experimental Design

Patients with recurrent and/or metastatic HNSCC after platinum-based therapy were treated with dasatinib either orally or via percutaneous feeding gastrostomy (PFG). Primary endpoints were 12-week progression-free survival (PFS) and objective response rate (ORR) with a 2 stage design and early stopping if 12-week PFS was 20% or less and no patients had an objective response (OR). Forty-nine serum cytokines and angiogenic factors (CAFs) were analyzed from treated patients.

Results

Of fifteen patients enrolled, twelve were evaluable for response, and all were evaluable for toxicity. No OR was observed and two patients (16.7%) had stable disease (SD) at eight weeks. Median treatment duration was 59 days, median time to progression (TTP) 3.9 weeks, and survival 26 weeks. One patient required dose reduction, 3 required interruptions, and 4 were hospitalized for toxicity. Dasatinib inhibited c-Src both when ingested and via PFG. Greater mean drug exposure, decreased half-life, and greater maximum concentration was observed in patients receiving dasatinib via PFG. Eleven baseline CAFs were associated with treatment outcome and one, MIF, was differentially modulated in correlation with SD versus progression.

Conclusions

Single-agent dasatinib failed to demonstrate significant activity in advanced HNSCC, despite c-Src inhibition. The toxicity profile was consistent with that reported in other solid tumors, and the drug can be given via PFG tube.

Introduction

Recurrent or metastatic disease is common in patients with advanced HNSCC, and although some are candidates for palliative chemotherapy, most will die within 1 year of recurrence (1). Molecular targeting has been confirmed as a relevant strategy in cancer therapeutics for other solid tumors and has been an active area of pre-clinical and clinical investigation in the past decade in the setting of HNSCC. To date, only the addition of anti-EGFR antibody to standard platinum-based chemotherapy has improved overall survival (OS) (2). Single-agent trials with targeted agents show promising but modest results (5–6).

c-Src is a non-receptor cytoplasmic tyrosine kinase that regulates signals from cell surface molecules including growth factor receptors and G-protein-coupled receptors. They play a key role in modulating multiple cellular functions, including invasion, adhesion, motility, migration proliferation, and survival, by activating the signal transducer and activator of transcription (STAT) family of transcription factors. Preclinical evaluations have provided a strong rationale for targeting c-Src in HNSCC (7). c-Src inhibition using dasatinib (BMS-354825), a potent inhibitor of src-family kinases EphA2, platelet-derived growth factor (PDGFR), Abl, BTK, and c-kit., in HNSCC cell lines in vitro has led to cell cycle arrest and apoptosis (8). In human HNSCC, c-Src (9, 10) is over expressed and correlated with lymph node metastases (10).

Based on this encouraging preclinical information, we conducted a phase II trial of single-agent dasatinib in patients with platinum-refractory recurrent metastatic HNSCC. The primary objective of this trial was to determine the 12-week PFS and ORR to single agent dasatinib. Secondary objectives were to measure metabolic response rate by PET scan at 8 and 12 wks, define OS, define duration of response, determine if there is correlation between clinical benefit from dasatinib and PK, PD, or changes in serum levels of cytokines, growth factors, and growth factor receptors relevant to the c-Src signaling pathway, and to describe the PK profile, relative bioavailability, safety of dasatinib suspension in patients receiving the drug through PFG tube.

Materials and Methods

Study design and Treatment

This was a 2-stage, open-label phase II study to characterize the efficacy and safety of dasatinib treatment in patients with recurrent or metastatic HNSCC previously treated with no more than one prior chemotherapeutic regimen. The primary dual endpoint of this study was 12-week PFS and ORR. A 2-stage Bayesian design was used which required stop the trial for futility if fewer than 5 of 15 patients had not progressed by 12 weeks and if less than 2 of these 15 had experienced an OR. All patients received dasatinib at a dose of 100mg PO BID and one cycle was 28 days. The dasatinib was generously provided by Bristol Myers Squibb and the National Cancer Institute, NIH. If patients were unable to swallow, dasatinib could be administered via PFG tube after mixed in lemonade. Treatment was intended to continue until progression of disease, illness that prevented further treatment, unacceptable agent-related adverse event, drug related toxicity not alleviated by two consecutive dose reductions, or failure to recover to less than grade 2 toxicity within 14 days of interruption of treatment, withdrawal of consent, or by judgment of investigator.

Dose reductions were performed for grade 3 hematologic toxicity or grade 2 hemorrhage, bleeding or coagulopathy without thrombocytopenia to 150 mg daily (−1 dose level) and 50mg BID (−2 dose level). For grade 3 or 4 bleeding or coagulopathy, treatment was discontinued, while for significant QTc prolongation (>550 milliseconds) where dasatinib was thought to have contributed, a two dose level reduction was implemented. Fluid retention, including pleural effusions was controlled early with institution of diuresis, steroids were used at the discretion of the treating physician. Thoracentesis was performed for persistent and symptomatic effusions.

Eligibility

Patients with histologically proven HNSCC recurrent after surgery and/or radiation therapy or chemoradiation therapy or with metastatic disease were eligible for the trial if they had received no more than one prior systemic therapy regimens. Other eligibility requirements included measurable disease by RECIST criteria, age ≥18 years, Eastern Cooperative Oncology Group performance status 0–1, and adequate organ and marrow function (ANC≥ 1.5 ×10⁹ /L, platelets ≥ 100 ×10⁹ /L, Hgb ≥ 9 g/dL, total bilirubin ≤ 1.5 × institutional upper limit of normal (ULN) albumin ≥2.5 g/dL, AST (SGOT)/ALT (SGPT) ≤1.5 × ULN, creatinine ≤3mg/dL). Pregnant, breastfeeding, and patients of child bearing potential not practicing contraception or sexual abstinence were excluded. Brain metastases were permitted if patient did not require anticonvulsants or corticosteroids, were either >4 weeks beyond cranial irradiation or felt not to require it at the time of enrollment. Exclusion criteria included chemotherapy or palliative radiotherapy for recurrent and/or metastatic disease within 3 weeks, failure to recover from adverse events (to great 1 or less), concurrent treatment with potent inhibitors or inducers of CYP3A4, myocardial infarction or ventricular arrhythmia within 6 months, QTc prolongation >480 milliseconds or major conduction abnormality. All patients gave written, informed consent prior to trial enrollment.

Study assessments

Baseline evaluations within 28 days prior to initiation of dasatinib included full medical history, physical examination, performance status assessment, electrocardiogram (EKG), laboratory tests, baseline PET/computerized tomography (CT), and any additional radiology studies as indicated. Radiographic response, determined by RECIST criteria, was planned at 4 weeks and then every 8 weeks thereafter, while metabolic response, determined by PET, was planned for 12 weeks. Patients were assessed at the beginning of each cycle of therapy for adverse events, by physical examination, laboratory tests and EKG. All adverse events were recorded and graded using the National Cancer Institute's Common Toxicity Criteria version 3.0 and followed for at least 4 weeks after removal from study or for those removed from study for adverse events until resolution or stabilization of event. Patients had off-study visit 3 to 5 weeks after discontinuation and were followed via mail or telephone until death. Hospice patients were followed via telephone.

Pharmacokinetic Sampling and Assay

Pharmacokinetic samples were collected In PO patients on cycle 1, day 1 at pre-dose (baseline), 1, 2, 4, 6, and 10 hrs following the time of ingestion. In PFG patients, cycle 1, day 1 sampling included pre-dose (baseline), 2, and 6 hrs following the time of administration. Whole blood was collected into sodium heparin BD Vacutainer tubes (Becton Dickenson, Franklin Lakes NJ). Each sample was centrifuged to isolate plasma at 1500 rpm for 10 minutes at 4 °C. Two aliquots of plasma were transferred to labeled polypropylene cryovials (Corning, Corning, NY) and stored at −70 °C until analysis.

Pharmacokinetic Analysis Methods

Dasatinib concentrations in plasma were determined using a validated LC/MS/MS assay with modifications (11). An Agilent 1100 series LC pump (Agilent, Santa Clara, CA) with a 4 °C autosampler was run in tandem with a Micromass Quattro Micro mass spectrometer (Waters, Milford, MA) using MassLynx/QuanLynx software for data acquisition and processing. Prior to drug extraction, the plasma was acidified by adding 250 μL 0.4% formic acid to 250 μL plasma. Solid-phase extraction was performed on an Oasis HLB 30 mg, 1 mL cartridge (Waters, Milford, MA) conditioned with 2 mL methanol followed by 2 mL water, washed with 2 mL water then 1 mL 5% methanol in water, and eluted with 1 mL 50:50, methanol:acetonitrile (v:v). The eluent was dried at 40 °C under gentle nitrogen gas stream and reconstituted with 250 μL 70:30, 0.2% formic acid:methanol (v:v). Chromatographic separation was carried out following 20 μL injection on a Phenomenex Luna C8, 50 × 2.1 mm, 3 μm analytical column heated to 60 °C over 8 minutes at 0.35 mL/min using 85:15, 0.2% formic acid:methanol, (v:v) with a linear gradient to 100% methanol over the first 3 minutes. The monitored transition (MRM) of 488.2 > 401.1 (m/z) had a retention time of 3.1 ± 0.1 minutes in positive electrospray ionization mode (ESI+). The calibration curve ranged from 1 – 1000 ng/mL in single donor human plasma. Pharmacokinetic parameter estimates were generated from individual patient concentration-time data using a non-compartmental model (WinNonLin v5.2, Pharsight Corp., Mountain View, CA).

Pharmacodynamics and CAF Analysis

Serum sample collection and preparation

Blood was drawn to evaluate circulating levels of serum cytokines and activated c-Src expression (pSrcY416) in platelets pre-treatment (cycle 1, day 1) and at cycle 2, day 1 (pre-dose) of treatment in patients receiving dasatinib PO and per PFG. Baseline serum was isolated from blood samples from patients at cycle 2 day 1 (pre-dose) in 14 patients. After the sample was allowed to coagulate for 30–60 minutes, it was centrifuged for 15 minutes at 1000g to separate serum which was then placed in a cryovial and stored at −70°C until analysis.

Whole blood was collected into EDTA-vacutainer tubes and immediately supplemented with citric acid dextrose (8.5mM sodium citrate, 11mM dextrose, 7nM citric acid) and prostacyclin (50ng/mL). Platelets were isolated using serial centrifugation. Platelets were lysed and activated c-Src measured using an enzyme-linked immunosorbent assay according to the manufacturer's protocol (ELISA, Calbiochem, San Diego CA). Samples were compared with a manufacturer-prepared standard of platelets stimulated with ATP.

Multiplex bead assay and ELISA

Forty nine CAFs were measured by multiplex bead assay (12, 13). These included hypoxia-induced factors, chemokines, interleukins, angiogenic factors, apoptosis mediators, and hematopoietic growth factors (Table 1). Multiplex bead assays were performed with BioSource Multiplex Assays for Luminex (Invitrogen, Calsbad, CA) in a 96-well format according to the BioSource protocol. Multiplex bead assay is a high-throughput technology that uses antibody-coupled microspheres identified by a unique ratio of two fluorescent dyes to measure up to 100 biomolecules simultaneously. Performance characteristics are comparable to ELISA (14, 15).

TABLE 1.

Cytokines and angiogenic factors in serum biomarker analysis

β-NGF	CTACK	Eotaxin	SCF	SCGF-b	ICAM-1	VCAM-1
HGF	M-CSF	G-CSF	GM-CSF	RANTES	GRO-a	PDGF-bb
TNF-a	TNF-b	TRAIL	IP-10	IFN-a2	IFN-g	IL-1a
IL-2	IL-1RA	IL2RA	IL-3	IL-4	IL-6	IL-7
IL-8	IL-9	IL-10	IL-12 (p40)	IL-12	IL-13	IL-15
IL-16	IL-17	IL-18	LIF	MIG	MIF	MCP-1
MCP-3	MIP-1a	MIP1-b	SDF-1a	Osteopontin	VEGFR2	VEGF

Abbreviations: nerve growth factor (NGF), cutaneous T cell-attracting chemokine (CTACK), stem cell growth factor (SCGF), inter-cellular adhesion molecule (ICAM), vascular cell adhesion molecule (VCAM-1), hepatocyte growth factor (HGF), monocyte colony stimulating factor (M-CSF), granulocyte colony stimulating factor (G-CSF), regulated upon activation, normal T-cell expressed and secreted (RANTES), growth-related oncogene (Gro), platelet derived growth factor (PDGF), tumor necrosis factor (TNF), tumor-necrosis-factor related apoptosis inducing ligand (TRAIL), interferon-inducible protein (IP), interferon (IFN), interleukin (IL), LIF (lymphocyte inhibitory factor), mitogen-indicuble gene (MIG), macrophage migration inhibitory factor (MIF), monocyte chemotactic protein (MCP), macrophage inflammatory protein (MIP), stem cell factor (SCF), stromal cell derived factor (SDF), vascular endothelial growth factor receptor (VEGFR).

Calculations

CAF concentrations were calculated based on a standard curve derived by performing six serial dilutions of a protein standard in assay diluent. Serum samples were tested in duplicate and the mean value used for analysis. If the mean of the duplicate values for all factors in an individual sample varied by >25%, the sample was re-tested. Individual factors were excluded from analysis if ≥50% of the serum values measured for that factor samples were either out of range or extrapolated. When out of range values were included in the analysis, the highest value (measured or extrapolated) across all samples was substituted for values above the upper limit of detection. For values below the lower limit of detection, half the lowest value was substituted. Both analytes were measured in duplicate and mean value used for analysis.

Endpoints and Statistical analysis

The dual primary endpoint was objective response rate, defined by RECIST and PFS evaluated on an intent-to treat basis, defined as the duration of time from start of treatment to time of progression. Patient tumor characteristics were tabulated. Descriptive statistics, including mean, median and range are provided for the continuous variable of patient age, whereas frequency and percentage for categorical variables. Overall survival (OS) was calculated from the date of study entry to the date of death or last follow-up. The Kaplan-Meier method (16) was used to estimate PFS and OS and the 95% confidence intervals (CI) were calculated. Differences between responders and non-responders with respect to PD endpoints were assessed using 2-sample independent t-tests, while differences between pre- and post-baseline measurements were assessed used a paired t-test. For all assessments p-values of less than 0.05 were deemed statistical significant.

Results

Patient, disease, and prior therapy characteristics

Table 2 summarizes the patient characteristics of the fifteen patients enrolled in this phase II study, since the predefined criteria for OR and PFS was not fulfilled. All patients had received radiation therapy previously and most had received chemotherapy or had undergone surgical resection previously, 93.3% and 73.3% respectively. Two patients were current smokers, ten were former smokers, and three were never smokers. Evaluation of recurrence at the time of study entry revealed that 80% of patients had local recurrence. No patients went on for further therapy after discontinuing the trial.

TABLE 2.

Baseline Patient Demographics and Characteristics

	No.	%
Gender
Male (ages 52–73)	11	73.3
Female (ages 54–78)	4	26.7
Smoking Status
Current	2	13.3
Former	10	66.7
Never	3	20.0
Alcohol Use
Current	10	66.7
Former	2	13.3
Never	3	20
Number of Prior Therapies
Radiation therapy	15	100.0
Chemotherapy	14	93.3
Surgery	11	73.3
Site of Metastasis
Lymph Node	2	13.3
Bone	3	20.0
Lungs	9	60.0
Other	1	6.7
Site of Recurrence
Local	12	80.0

Objective Response, Progression-Free Survival, and Overall Survival

No complete or partial responses were reported among the twelve patients evaluable for response. Two patients had SD at 8 weeks (16.7%). One patient had SD for greater than three cycles of treatment. The 12-week PFS based on the Kaplan-Meier Curve for all patients was 13.3% (95% CI (0.022, 0.346); Fig 1A). Median PFS was 0.91 months (95% CI: 0.91–1.97 months). The median OS was 6.05 months (95% CI:3.05–8.31 months) Fig 1B). At the time of data analysis, a total of 14 patients were deceased. Median follow-up time was 4.63 months (95% CI 3.05–7.29 months).

Figure 1.

Kaplan-Meier curves for (A) PFS and (B) Overall survival

Toxicity

Table 3 summarizes the most common toxicities considered to be drug-related. The most common toxicities were nausea and pleural effusion (both 40%), although the severity was grade 1–2 in most cases. Grade 3 pleural effusion, dehydration, and anemia occurred in one patient each. Overall, dasatinib was not associated with severe hematological toxicities or laboratory changes. Seven patients discontinued therapy because of disease progression, four patients discontinued for toxicity, three patients had treatment interruptions for toxicity, and four were hospitalized for toxicity.

TABLE 3.

Most common Drug Related Adverse events

	Grade 1–2	Grade 3–4
	No. (%)	No. (%)
Pleural Effusion	5 (33.3)	1 (6.7)
Dehydration	0	1 (6.7)
Anemia	1 (6.7)	1 (6.7)
Nausea	6 (40.0)	0
Vomiting	4 (26.7)	0
Diarrhea	3 (20.0)	0
Dyspnea	3 (20.0)	0
Limb Edema	3 (20.0)	0

PFG Pharmacokinetic Analysis

Plasma concentrations of dasatinib were available on a small number of patients (n=3) who received the drug via PFG. Due to sparse sampling in the patients administered dasatinib by mouth, we were not able to estimate any pharmacokinetic parameters. When compared to oral ingestion of drug in a previous phase I study in patients with advanced solid tumors (17), there is greater systemic drug exposure via PFG based on AUC comparison (AUC_0–10 = 448.9 vs. AUC_0–10 = 218.0 ng•hr/mL). Maximum dasatinib plasma concentration is also greater in PFG patients when compared to the phase I study (168.2 vs. 56.0 ng/mL) while elimination half-life was shorter (2.1 vs. 4.3 hours). These differences are most likely due to peak effects within 1.5 hrs following administration caused by enhanced drug absorption via PFG or difference in method of administration, but it does warrant further investigation to determine their nature. For those patients we were able to compare dasatinib plasma concentrations in this study, there was no significant difference in drug concentrations at 2 and 6 hour post-dose when comparing oral vs. PFG patients (Figure 2).

Figure 2.

Mean plasma concentration-time profiles of dasatinib following PFG and PO administration

Pharmacodynamic Analysis

Four patients had measurement of activated c-Src (p-Src) in platelets prior to starting therapy and on day 21 at 0, 2, and 6h (Fig. 3). One patient only had p-Src levels drawn pretreatment and pre-dose on day 21. The level of p-Src on day 21 at 0h (i.e. about 12 h since the prior day's dose) was variable when compared to the pre-treatment level – in 3 of 5 samples the value was significantly lower (p<0.05) and in 2 cases there was no statistical difference [one lower (p=0.15), one higher (p=0.06)]. In all four patients with the full panel drawn, the level of p-Src was reduced below the level of detection 2h after dasatinib administration consistent with the rapid absorption of dasatinib observed in the PK studies. In 3 of 4 patients, the level of p-Src remained undetectable 6h after dasatinib administration. Overall these studies demonstrated that dasatinib led to a universal, rapid and complete inhibition of c-Src in patients' platelets with a variable recovery 6–12 hours later. (Figure 3)

Figure 3.

Graph of platelet p-Src levels (arbitrary units) baseline versus day 21.

Cytokine Results

Fourteen pts had baseline levels, and nine had paired samples. CAF levels (log₂) were log₂ transformed and compared between patients with (n=12) and without (n=2) PD by t-test. Eleven CAFs were significantly higher in patients who had PD (p≤0.05): CTACK, IFNa2, interleukin (IL)-2, IL-12, IL-3, LIF, MCP-3, macrophage migration inhibitory factor (MIF), stem cell factor (SCF)-beta, TNF β and b-NGF. Cytokine modulation was compared between patients with SD and those that rapidly progressed. Increase in MIF was seen in patients with rapid disease progression (2,945.7 pg/ml) compared to patients with disease stabilization where MIF was decreased (−19,399 pg/ml) (p=0.07, t-test). (Table 4)

Table 4.

Baseline cytokines associated with Disease Progression

Covariate	Progression	n	Mean +/- std*	p-value+
CTACK	Yes	12	1130.9 +/- 242.08	0.0060
	No	2	536.2 +/- 105.56
IFN-a2	Yes	12	270.58 +/- 21.886	0.0015
	No	2	205.09 +/- 5.5932
IL-2RA	Yes	12	359.53 +/- 69.985	0.0300
	No	2	228.83 +/- 64.262
IL-12p40	Yes	12	1277.2 +/- 215.48	0.0292
	No	2	865.09 +/- 243.24
IL-3	Yes	12	1858.2 +/- 214.01	0.0129
	No	2	1389.1 +/-169.15
LIF	Yes	12	100.54 +/- 40.921	0.0312
	No	2	25.75 +/-30.109
MCP3	Yes	12	169.17 +/- 35.76	0.0417
	No	2	198.27 +/- 12.056
MIF	Yes	12	4432 +/- 4109.7	0.0391
	No	2	20199 +/- 27716
MCSF	Yes	12	491.12 +/- 96.481	0.1006
	No	2	362.89 +/- 67.288
SCF	Yes	12	432.48 +/- 53.776	0.0122
	No	2	304.54 +/- 83.538
TNFb	Yes	12	63.824 +/- 16.063	0.0851
	No	2	41.345 +/- 10.699

^*values are pg/mL log2 transformed

⁺p-value from t- test

Discussion

This single-center, open-label, phase II study met its objectives to evaluate the safety, tolerability, PK, and biological efficacy of dasatinib administered as a single agent in patients with recurrent HNSCC. Dasatinib as a single agent failed to demonstrate significant activity in advanced, metastatic HNSCC. The observed toxicity profile was consistent with that reported in other solid tumors, and the drug can be given successfully via PFG tube.

Many challenges exist in the development of novel therapeutics against HNSCC. Phase II studies of molecularly targeted compounds in patients with recurrent or metastatic HNSCC are beset with the challenge of identifying patients whose tumors demonstrate an OR, which is unlikely to occur in more than 10–15% of the patients based on existing literature (3–6). Another approach in this setting would be the use of progression-based endpoints. We chose to incorporate a dual endpoint in our trial in order to identify patients that may derive benefit from dasatinib despite the lack of objective response. We were, however, unable to demonstrate that dasatinib can reach the benchmark set by agents such as cetuximab, gefitinib or erlotinib (targeting for 6-month PFS rates of ≥15%) (3–6).

We also planned to investigate baseline expression and possible modulation of serum cytokines that may serve as biomarkers in order to define patients that may derive more benefit from this agent in the current or future trials. We analyzed cytokines in the serum and although no significant activity was observed we were able to characterize certain profiles of cytokines correlating with outcome. All baseline cytokines and growth factors with the exception of SCF-beta, IL-17 and PDGF were higher in non-responders, however several were significantly higher in patients who demonstrated rapid progression including MIF.

Although pro-inflammatory cytokines were generally increased post-treatment, MIF, a tumor growth regulator was higher after treatment in patients with rapid progression and lower in patients whose disease stabilized on treatment. MIF is a lymphokine involved in cell-mediated immunity, immunoregulation and inflammation, and plays a role in the regulation of macrophage function in host defense through suppression of anti-inflammatory effects of glucocorticoids. Although the exact role of MIF in tumorigenesis remains ambiguous, several reports are now linking MIF to processes such as cell proliferation, differentiation and tumor progression and metastasis and overexpression has been reported in several solid tumors (18–21). MIF enhances neoplastic cell invasion by inducing the expression of matrix metalloproteinase 9 and IL- 8. It is also known to stimulate angiogenic factor expression (VEGF and IL-8) (22). Despite evidence of effective inhibition of p-Src, dasatinib did not appear to have effects on tumor growth. It is possible that several signaling pathways are activated in advanced HNSCC (23) and therefore targeting p-Src is not sufficient. In designing clinical trials in this setting, it is important to consider the complexities of cell signaling in advanced malignancies including HNSCC, which may result in upregulation of several feedback mechanisms when a single target is hit. Taking that into consideration, an alternative strategy in this setting is to simultaneously target two or more pathways, ideally with agents that do not possess overlapping toxicity. Along these lines further investigation of dasatinib in combination with EGFR inhibitors is ongoing.

Statement of Translational Relevance.

This study establishes the feasibility and efficacy of administering the c-Src inhibitor dasatinib via percutaneous feeding gastrostomy (PFG), presenting pharmacokinetic (PK) data to show comparable treatment concentration levels. Pharmacodynamic (PD) data shows that activated c-Src is inhibited by dasatinib administration through both routes. Although dasatinib as a single agent failed to demonstrate significant activity in advanced and/or metastatic head and neck squamous cell carcinoma (HNSCC), we investigated baseline expression and modulation of serum cytokines in order to identify those that may serve as clinically useful biomarkers for current or future trials. Most baseline cytokines and growth factors were higher in non-responders, and macrophage migration inhibitory factor (MIF) was significantly higher in patients who demonstrated rapid progression. The observed toxicity profile was consistent with that reported in other solid tumors.