|Home | About | Journals | Submit | Contact Us | Français|
There is no effective therapy for patients with advanced medullary thyroid carcinoma (MTC). Vandetanib, a once-daily oral inhibitor of RET kinase, vascular endothelial growth factor receptor, and epidermal growth factor receptor signaling, has previously shown antitumor activity in a phase II study of patients with advanced hereditary MTC.
Patients with advanced MTC were randomly assigned in a 2:1 ratio to receive vandetanib 300 mg/d or placebo. On objective disease progression, patients could elect to receive open-label vandetanib. The primary end point was progression-free survival (PFS), determined by independent central Response Evaluation Criteria in Solid Tumors (RECIST) assessments.
Between December 2006 and November 2007, 331 patients (mean age, 52 years; 90% sporadic; 95% metastatic) were randomly assigned to receive vandetanib (231) or placebo (100). At data cutoff (July 2009; median follow-up, 24 months), 37% of patients had progressed and 15% had died. The study met its primary objective of PFS prolongation with vandetanib versus placebo (hazard ratio [HR], 0.46; 95% CI, 0.31 to 0.69; P < .001). Statistically significant advantages for vandetanib were also seen for objective response rate (P < .001), disease control rate (P = .001), and biochemical response (P < .001). Overall survival data were immature at data cutoff (HR, 0.89; 95% CI, 0.48 to 1.65). A final survival analysis will take place when 50% of the patients have died. Common adverse events (any grade) occurred more frequently with vandetanib compared with placebo, including diarrhea (56% v 26%), rash (45% v 11%), nausea (33% v 16%), hypertension (32% v 5%), and headache (26% v 9%).
Vandetanib demonstrated therapeutic efficacy in a phase III trial of patients with advanced MTC (ClinicalTrials.gov NCT00410761).
Medullary thyroid carcinoma (MTC), a malignancy of the parafollicular C cells of the thyroid gland, accounts for approximately 5% of all thyroid cancers and presents either sporadically (75% of patients) or in a hereditary pattern.1,2 The 10-year overall survival rate in unselected patients with MTC is approximately 75%, but it decreases to 40% or less in patients with locally advanced or metastatic disease.1,3,4 Neither radiotherapy nor chemotherapy has demonstrated durable objective responses in patients with advanced MTC.5,6
Germline mutations in the RET (rearranged during transfection) proto-oncogene occur in virtually all patients with hereditary MTC.7–9 Approximately 50% of patients with sporadic MTC have somatic RET mutations, and 85% of them have the M918T mutation.10,11 Evidence from preclinical studies of molecular targeted therapeutics with activity against RET demonstrate that RET kinase is a potential therapeutic target in MTC.12–14 Other signaling pathways likely to contribute to the growth and invasiveness of MTC include vascular endothelial growth factor receptor (VEGFR)–dependent tumor angiogenesis and epidermal growth factor receptor (EGFR)–dependent tumor cell proliferation.15
Vandetanib is a once-daily oral agent that selectively targets RET, VEGFR, and EGFR signaling.12,16 We report the results of an international, randomized, placebo-controlled, double-blind, phase III study (ZETA) to evaluate vandetanib 300 mg/d in patients with locally advanced or metastatic MTC.
Eligible patients were adults who had measurable, unresectable locally advanced or metastatic, hereditary or sporadic MTC. Submission of a tumor sample was required except for patients with hereditary MTC who had a documented germline RET mutation. Other key inclusion criteria were WHO performance status of 0 to 2 and serum calcitonin level ≥ 500 pg/mL. Exclusion criteria included significant cardiac, hematopoietic, hepatic, or renal dysfunction and administration of chemotherapy and/or radiation therapy within 4 weeks before random assignment. All patients provided written informed consent. The protocol was approved by all relevant institutional ethical committees or review bodies, and the study was conducted in accordance with the Declaration of Helsinki, Good Clinical Practice, and the AstraZeneca policy on bioethics.
Patients recruited to this multicenter phase III study were randomly assigned in a 2:1 ratio to receive oral vandetanib at a starting dose of 300 mg/d or placebo until disease progression. On objective disease progression based on investigator assessment, patients discontinued study treatment, were unblinded, and could elect to enter postprogression, open-label treatment with vandetanib until a withdrawal criterion was met. All patients were to be followed for survival.
The primary objective was to determine whether vandetanib, compared with placebo, prolonged progression-free survival (PFS) on the basis of independent central review. Secondary assessments included objective response rate, disease control rate at 24 weeks, duration of response, overall survival, biochemical response (decreases in serum levels of calcitonin and carcinoembryonic antigen [CEA]), and time to worsening of pain (for time to worsening of pain, see Methods and Results in the Appendix, online only).
The principal investigator in collaboration with the study sponsor, AstraZeneca, designed the clinical trial. The sponsor provided funding and organizational support, collected and managed the data, and performed the statistical analysis. Each author reviewed and approved the manuscript and the principal investigator had final responsibility for the decision to submit for publication. The senior academic authors developed the manuscript, and all coauthors contributed to the manuscript.
PFS was determined from objective tumor measurements performed at screening and then every 12 weeks until progression or withdrawal of consent. Tumor assessments were categorized by the investigator by using Response Evaluation Criteria in Solid Tumors v1.0 (RECIST).17 Responses were confirmed by central review of separate assessments performed at least 4 weeks apart. RECIST assessments derived from an independent central review of patient scans were the basis for the primary analysis. If, according to central review, progression had not occurred by the time a patient entered open-label treatment, open-label RECIST assessments were also used in the derivation of PFS, objective response rate, disease control rate, and duration of response. PFS was defined from the date of random assignment to the date of objective progression or death (by any cause in the absence of progression within 3 months of the last evaluable RECIST assessment). Patients who had not progressed or who had died at the time of analysis were censored at the time of their last evaluable RECIST assessment.
Overall survival was calculated from the date of random assignment to the date of death with patients followed every 12 weeks until withdrawal of consent or death. Patients who had not died at the time of analysis were censored at the time they were last known to be alive. A final analysis of overall survival is planned when 50% of the study patients have died.
The presence of an RET mutation was determined by a combination of two methods: (1) an amplification-refractory mutation system (ARMS) assay that specifically detects the most common RET mutation (M918T) found in sporadic MTC, and (2) direct DNA sequencing following polymerase chain reaction amplification of RET (exons 10, 11, and 13 to 16). A mutation-positive sample had either M918T by ARMS assay or an RET mutation in any of exons 10, 11, and 13 to 16. Conversely, a mutation-negative sample had no M918T mutation by ARMS and a wild-type RET sequence in each of exons 10, 11, and 13 to 16. The mutation status was declared unknown in cases in which an assay failed to yield a sequence at any of the tested exons (by sequencing or ARMS assay), and none of the successful assays demonstrated a mutation.
Blood samples for calcitonin and CEA analysis were collected at baseline (day 1), every 4 weeks until week 12, and then every 12 weeks thereafter. Serum levels of calcitonin and CEA were determined as previously described.18 A patient's best biochemical response for either calcitonin or CEA was defined as follows: complete response, normalization of serum levels following treatment confirmed a minimum of 4 weeks later; partial response, ≥ 50% decrease from baseline levels maintained over a minimum of 4 weeks; stable disease, between +50% and −50% change from baseline levels maintained for at least 4 weeks; and progressive disease, ≥ 50% increase from baseline maintained for at least 4 weeks.
Safety was assessed throughout the study by monitoring and recording adverse events, 12-lead ECG parameters, vital signs, clinical chemistry, hematology, and urinalysis. Adverse events were assessed by using the National Cancer Institute's Common Terminology Criteria for Adverse Events (CTCAE, v3). Scheduled 12-lead ECGs were performed during screening, at 1, 2, 4, 8, and 12 weeks and every 3 months thereafter. The QTc interval was evaluated centrally, and prolongation was defined as a single measurement of ≥ 550 ms or an increase of ≥ 100 ms from baseline, two consecutive measurements (within 48 hours of each other) that were ≥ 500 ms but less than 550 ms, or an increase of ≥ 60 ms but less than 100 ms from baseline to a value ≥ 480 ms. Specific dose reduction plans were in place for management of skin toxicity, GI toxicity, and QTc prolongation. There was also a general dose reduction scheme for any CTCAE grade 3 or 4 adverse event (patients started at vandetanib 300 mg/d or placebo, and dose was reduced to 200 mg/d for a grade 3 or 4 adverse event; if further grade 3 or 4 toxicity occurred, reduction to 100 mg/d was allowed).
The study was designed to have more than 80% power to detect a hazard ratio (HR) less than 0.50 at a 5% significance level; a minimum of 90 progression events were required, assuming a median PFS of 12 months in the placebo group and an overall sample size of 232 patients. Analyses of PFS and overall survival were conducted by using the log-rank test (unadjusted model with treatment factor only) in the intention-to-treat population. A sensitivity analysis of PFS was performed by using Cox's proportional hazards regression model, which allowed for the effect of treatment and included terms for RET mutation status, MTC status (hereditary or sporadic), prerandomization historic calcitonin and CEA changes, number of prior therapies, and response to prior therapy.19 The following predefined sensitivity analyses were also performed for PFS: per protocol that excluded significant protocol deviators; Whitehead method to assess the impact of a differential frequency of assessments in the two treatment arms; randomized phase alone (ie, excluding the open-label phase); and PFS derived from investigator assessments.20 The objective response rate and disease control rate were analyzed by using logistic regression (these variables included open-label assessments). All P values were two-sided. Subgroup analyses of PFS by prespecified baseline characteristics and ad hoc subgroup analyses of PFS and objective response rate by RET mutation status and M918T mutation status were performed.
Between December 7, 2006, and November 21, 2007, 331 patients recruited from 23 countries were randomly assigned to vandetanib (231) or placebo (100; Fig 1). Although not an exact 2:1 randomization, this imbalance occurred by chance. Patient characteristics and baseline demographics were similar in both arms (Table 1). The majority of patients presented with sporadic disease, and most had metastatic disease at study entry. At data cutoff (July 31, 2009), the median duration of follow-up was 24 months, and 139 patients were continuing blinded treatment: 111 (48%) randomly assigned to vandetanib and 28 (28%) randomly assigned to placebo. Among 123 patients who developed tumor progression and were eligible to receive open-label treatment, 93 (vandetanib, 41 of 67 [61%]; placebo, 52 of 56 [93%]) elected to enter postprogression open-label treatment with vandetanib. Overall, 48 deaths (32, vandetanib arm; 16, placebo arm) occurred at the time of data cutoff, including one patient randomly assigned to the placebo arm who died of progressive MTC before receiving study treatment and who was not included in the safety analysis population. All patients were included in the efficacy analysis.
At the time of analysis, 124 patients (37%) had progressed and 48 (15%) had died. Significant prolongation of PFS was observed for patients receiving vandetanib compared with placebo (HR, 0.46; 95% CI, 0.31 to 0.69; P < .001; Fig 2A; Table 2). The median PFS was 19.3 months in the placebo group and, although the median had not yet been reached for the vandetanib group, fitting a Weibull model indicated a predicted median of 30.5 months.21 The PFS at 6 months was 83% (vandetanib) and 63% (placebo). The Kaplan-Meier plot indicates that the relative hazards were larger at earlier time points. In addition to the primary analysis, Cox regression analysis as well as other sensitivity analyses detected an improvement in PFS with vandetanib versus placebo (Table 2). A total of 51 patients (23 vandetanib, 28 placebo) received open-label vandetanib before progression by central read was documented. Both visual inspection of the forest plot and the finding of a lack of statistical significance for the planned global interaction test (P = .177) suggest that the PFS benefits observed were generally consistent across all prespecified subgroups (Fig 2B).
Vandetanib also showed significant advantages compared with placebo in the secondary efficacy end points of objective response rate, disease control rate, and calcitonin and CEA biochemical response rates (Table 2). Objective responses were durable on the basis of the median duration of response not being reached at 24 months of follow-up (fitting a Weibull model gives a predicted median duration of response of 22 months). It is important to note that 12 of 13 responses observed in patients initially randomly assigned to placebo occurred while the patients were subsequently receiving vandetanib in the open-label phase. Overall survival data were immature at data cutoff (HR, 0.89; 95% CI, 0.48 to 1.65; Fig 3). A final survival analysis will take place when 50% of the patients have died.
Of the 33 patients with hereditary MTC, 32 had a documented RET germline mutation before study entry, and of the 28 receiving vandetanib, 13 (46.4%) had an objective response. Paraffin blocks or slides were available for analysis from 297 of 298 patients with sporadic MTC. An RET mutation was present in 155 patients (52.0%), no RET mutation was present in eight patients (2.7%), and the RET mutation status was unknown in 135 patients (45.3%). There was a high number of patients with unknown RET mutation status because their paraffin blocks or slides had an insufficient quantity or quality of DNA for complete analysis. The small number of RET-negative patients means that subgroup analyses of PFS (Fig 3B) and objective response rate (Table 3) by RET mutation status are inconclusive. In patients with sporadic MTC, however, a subgroup analysis of PFS by M918T mutation suggested that M918T mutation–positive patients had a higher response rate to vandetanib compared with M918T mutation–negative patients (Fig 2C; Table 3).
The median duration of treatment in the randomized phase was 90.1 weeks (vandetanib) and 39.9 weeks (placebo). Common adverse events (any grade and grade 3 or higher) are summarized in Table 4. Thirty-one patients discontinued treatment during the randomized phase because of an adverse event: 28 (12%) receiving vandetanib and three (3%) receiving placebo. Adverse events such as diarrhea, rash, nausea, and hypertension occurred in more than 30% of patients receiving vandetanib; adverse events leading to discontinuation of vandetanib reported in more than 1% of patients were asthenia (1.7%) and rash (1.3%). More patients required dose reduction of vandetanib compared with placebo for adverse events or QTc prolongation (35% v 3%). Nineteen patients (8%) developed protocol-defined QTc prolongation, but there were no reports of torsades de pointes. From entry, more patients on vandetanib compared with placebo were noted to have rising thyroid-stimulating hormone serum levels, and they required an increase in thyroid replacement (49.3% v 17.2%). Five patients on the vandetanib arm experienced adverse events leading to death during the randomized phase; these were single instances of aspiration pneumonia, respiratory arrest, respiratory failure, staphylococcal sepsis, and arrhythmia and acute cardiac failure in one patient. The two deaths due to an adverse event in the placebo arm were isolated cases of gastroenteritis and GI hemorrhage.
Patients with locally advanced or metastatic MTC are incurable, and chemotherapy and radiation therapy have been largely ineffective. Therefore, the ability to substantially prolong the time to disease progression would benefit such patients. Mutations in the RET proto-oncogene are central to the development of MTC in virtually all patients with hereditary MTC and in approximately half the patients with sporadic MTC.7–9,11 Following preclinical studies demonstrating that vandetanib inhibited signaling through RET kinase,12,22 a phase II clinical trial of oral vandetanib (300 mg) was initiated in patients with locally advanced or metastatic hereditary MTC. There were confirmed partial remissions in 20% of patients and stable disease of more than 24 weeks in 73% of patients.18 Published reports of early-phase clinical trials of other tyrosine kinase inhibitors in patients with advanced MTC have shown partial remission rates ranging from 0% to 25% in small single-arm trials.23–29
In this study, there was a significant prolongation of PFS in patients who received vandetanib compared with placebo, with an HR of 0.46 and an estimated 11-month prolongation of median PFS. In addition to the primary analysis, an improvement was also detected in all other predefined sensitivity analyses of PFS, including an HR of 0.27 if open-label scans were excluded. The secondary efficacy end points of objective response rate, disease control rate, and biochemical response also showed statistically significant benefit in the treatment group compared with the control group. Of the 13 objective responses in patients randomly assigned to placebo, 12 occurred while patients were receiving open-label vandetanib. Overall survival data are immature, and the final assessment will occur when 50% of patients have died; however, the analysis is likely to be confounded by the ability of patients randomly assigned to placebo to receive subsequent treatment with open-label vandetanib.
The benefit that was demonstrated in PFS for patients receiving vandetanib compared with placebo was observed in patients with the hereditary or the sporadic form of MTC. Because of the small number of patients with sporadic MTC who were RET negative and the large number of patients who were RET unknown, the subgroup analyses of PFS and objective response rate by RET mutation status are inconclusive. If data from the ARMS assay are taken into account, patients with sporadic MTC received benefit from vandetanib whether their tumors were M918T positive or negative; however, the response rate was greater in those who had an M918T mutation.
Treatment with vandetanib was generally well tolerated. The majority of adverse events were manageable according to standard clinical practice alone or in combination with vandetanib dose reductions, which allowed patients to continue receiving vandetanib treatment for extended periods of time. The rate of vandetanib treatment discontinuation because of toxicity was low (12%), despite a median duration of treatment of approximately 1 year and 9 months (90.1 weeks).
In this clinical study, vandetanib has shown efficacy in patients with locally advanced or metastatic MTC, a challenging group of patients for whom there has been no effective therapy.
The following clinical trialists participated in this study: Bruno Niederle, Universitätsklinik für Chirurgie, Vienna, Austria; Diana Learoyd, University of Sydney, Sydney, Australia; Ahmad Awada, Institut Jules Bordet, Brussels; Patrick Schöffski, Universitair Ziekenhuis Leuven, Leuven, Belgium; Léa Maciel, Hospital das Clínicas de Ribeirão Preto, Ribeirão Preto; Ana Maia, Hospital de Clinicas de Porto Alegre, Porto Alegre, Brazil; Monika Krzyzanowska, Ontario Cancer Institute/Princess Margaret Hospital, Toronto, Ontario; Marie-France Langlois, Centre Hospitalier Universitaire de Santé de l'Estrie Site Fleurimont, Sherbrooke, Quebec; Dean Ruether, Tom Baker Cancer Centre, Calgary, Alberta; Asif Shaikh, The Moncton Hospital Oncology Clinic, Moncton, New Brunswick; Eric Winquist, London Health Sciences Centre, London, Ontario, Canada; Petr Vlcek, Fakultni Nemocnine V Motol, Prague, Czech Republic; Lars Bastholt, Odense Universitetshospital, Odense, Denmark; Binh Bui, Institut Bergonié, Bordeaux; Christelle de la Fouchardière, Centre Léon Bérard, Lyon, France; Klaus Mann, Universitätsklinik Essen, Essen; Christoph Reiners, Klinikum der Julius-Maximilians-Universität Würzburg, Würzburg; Hans-Joachim Schmoll, Martin-Luther-Universität Halle-Wittenberg, Halle, Germany; László Mangel, Pécsi Tudományegyetem Onkoterápiás Intézet, Pécs, Hungary; Raju Chacko, Christian Medical College and Hospital, Vellore; Manju Sengar, Tata Memorial Hospital, Mumbai, India; Emilio Bombardieri, Istituto Nazionale Per Cura Tumori, Milan; Sebastiano Filetti, Azienda Universitaria Policlinico Umberto I, Rome; Furio Pacini, Policlinico Santa Maria alle Scotte, Siena; Gabriella Pellegriti, Azienda Ospedaliera Garibaldi, Catania; Domenico Salvatore, Università Degli Studi Di Napoli Federico II, Naples, Italy; Luis Mauricio Hurtado, Hospital General de México, México City, México; Thera P. Links, Universitair Medisch Centrum Groningen, Groningen; Bernard Zonnenberg, Universitair Medisch Centrum Utrecht, Utrecht, the Netherlands; Izabella Kozlowicz-Gudzinska, The Maria Sklodowska-Curie Memorial Cancer Centre and Institute of Oncology, Warsaw; Jerzy Sowinski, University of Medical Sciences in Poznan, Poznan, Poland; Maria João Bugalho, Instituto Português de Oncologia, Francisco Gentil EPE, Serviço de Endocrinologia, Lisbon, Portugal; Dae-Seog Heo, Seoul National University Hospital, Seoul, Korea; Constantin Dumitrache, Institute of Endocrinology “C.I. Parhon” Bucharest, Bucharest, Romania; Viktor Medvedev, Obninsk Radiology Research Center, Obninsk, Russia; Svetozar Damjanovic, Institute of Endocrinology, Diabetes and Metabolic Diseases, Clinical Center of Serbia, Belgrade; Radan Dzodic, Institute of Oncology and Radiology of Serbia, Belgrade, Serbia; Antonio Casado, Hospital Universitario San Carlos, Madrid; Lara Iglesias Docampo, Hospital Universitario 12 de Octubre, Madrid; Jesús García Foncillas, University of Navarra, Pamplona, Spain; Barbro Eriksson, Uppsala University Hospital, Uppsala; Arne Hallqvist, Karolinska University Hospital, Stockholm, Sweden; Thomas Krause, Inselspital Bern Nuklearmedizin, Bern; Rolf Stoeckli, University Hospital Basel, Basel, Switzerland; Kenneth Ain, University of Kentucky, Lexington, KY; Keith Bible, Mayo Clinic Rochester, Rochester, MN; Donald Bodenner, University of Arkansas for Medical Sciences, Little Rock, AR; Orlo Clark, Mount Zion Medical Center, San Francisco, CA; Ezra Cohen, University of Chicago Hospital, Chicago, IL; Jessica Gosnell, Mount Zion Medical Center, San Francisco, CA; Ann Gramza, Oregon Health Sciences University, Portland, OR; Steven Grunberg, Fletcher Allen Health Care, Burlington, VT; Robert Haddad, Dana-Farber Cancer Institute, Boston, MA; Joshua Hornig, Medical University of South Carolina, Charleston, SC; Zyad Kafri, Wayne State University-Karmanos Cancer Institute, Detroit, MI; Madeleine Kane, University of Colorado Denver, Aurora, CO; Jeffrey Moley, Washington University School of Medicine, St. Louis, MO; Robert Smallridge, Mayo Clinic Jacksonville, FL; Julie Sosa, Yale University School of Medicine, New Haven, CT; David Steward, University of Cincinnati, Cincinnati, OH.
Time to worsening of pain was a composite end point determined from patient-reported opioid analgesic use and responses to the Brief Pain Inventory (BPI) questionnaire (Daut RL, et al: Pain 17:197-210, 1983) administered at baseline and every week thereafter. Worsening of pain was defined as an adverse change of two or more points from baseline on the BPI worst pain scale (0 to 10 points) or an increase from baseline in opioid analgesic use of ≥ 10 mg/d morphine sulfate equivalent. Time to worsening of pain was the interval from the date of random assignment to the date of worsening of pain (with no evidence of improvement within 14 days). If worsening of pain was not observed at the time of analysis, time to worsening of pain was censored at the last evaluable assessment of pain or opioid usage.
Patients receiving vandetanib experienced a statistically significant delay in time to worsening of pain compared with those receiving placebo (hazard ratio, 0.61; 95% CI, 0.43 to 0.87; P = .006: Appendix Figure A1). Overall compliance, calculated as patients with baseline data and at least one evaluable composite assessment, was 89% in the vandetanib arm and 86% in the placebo arm.
See accompanying editorial on page 119 and article on page 200
Supported by AstraZeneca and by John Matthew, Mudskipper Bioscience (medical writing support).
Presented at the 46th Annual Meeting of the American Society of Clinical Oncology, Chicago, IL, June 4-8, 2010, and as a poster at the 35th Congress of the European Society for Medical Oncology, Milan, Italy, October 8-12, 2010.
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
Clinical trial information can be found for the following: NCT00322452.
Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Employment or Leadership Position: James R. Vasselli, AstraZeneca (C); Jessica Read, AstraZeneca (C); Peter Langmuir, AstraZeneca (C) Consultant or Advisory Role: Samuel A. Wells Jr, AstraZeneca (C); Bruce G. Robinson, AstraZeneca (C); Robert F. Gagel, AstraZeneca (C); Barbara Jarzab, AstraZeneca (C); Anderson J. Ryan, AstraZeneca (U); Martin J. Schlumberger, AstraZeneca (C) Stock Ownership: James R. Vasselli, AstraZeneca; Anderson J. Ryan, AstraZeneca Honoraria: Robert F. Gagel, AstreaZeneca; James A. Fagin, AstraZeneca; Barbara Jarzab, Exelixis, Eisai Research Funding: Robert F. Gagel, AstraZeneca; Massimo Santoro, AstraZeneca; Barbara Jarzab, AstraZeneca; Martin J. Schlumberger, AstraZeneca Expert Testimony: Robert F. Gagel, AstraZeneca (U); Anderson J. Ryan, AstraZeneca (C) Other Remuneration: Samuel A. Wells Jr, AstraZeneca
Conception and design: Samuel A. Wells Jr, Bruce G. Robinson, Robert F. Gagel, James A. Fagin, Massimo Santoro, Peter Langmuir, Anderson J. Ryan, Martin J. Schlumberger
Collection and assembly of data: Samuel A. Wells Jr, Bruce G. Robinson, Robert F. Gagel, Henning Dralle, Eric Baudin, Rossella Elisei, Barbara Jarzab, James R. Vasselli, Peter Langmuir, Anderson J. Ryan, Martin J. Schlumberger
Data analysis and interpretation: Samuel A. Wells Jr, Bruce G. Robinson, Robert F. Gagel, James A. Fagin, James R. Vasselli, Jessica Read, Peter Langmuir, Anderson J. Ryan, Martin J. Schlumberger
Manuscript writing: All authors
Final approval of manuscript: All authors