How Do U.S. Medical Oncologists Learn and Apply New Clinical Trials Information from Press Releases in Nonmedical Media? A Case Study Based
http://www.100md.com
《肿瘤学家》
LEARNING OBJECTIVES
After completing this course, the reader will be able to:
Discuss new vehicles for the initial release of practice-relevant phase III clinical trial data.
Describe the limited educational impact of such information transfer vehicles on the medical oncology community.
Explain the potential risks to medical oncologists and their patients because of underinformed or non–evidence-based therapeutic recommendations involving new indications for commercially available agents.
Discuss the development of new strategies for effective knowledge transfer of preliminary, practice-relevant, clinical trial data.
ABSTRACT
Background. Practicing oncologists are expected to easily assimilate large amounts of rapidly evolving clinical data. We hypothesized that U.S. oncologists rapidly familiarize themselves with new, practice-relevant, phase III clinical trial data. We tested this hypothesis in relation to the release of phase III data from the Eastern Cooperative Oncology Group 4599 trial on the role of bevacizumab in advanced non-small cell lung cancer (NSCLC).
Methods. We queried approximately 310 medical oncologists concerning their awareness of the bevacizumab data within 1 and 3 weeks after the data release or immediately after the 2005 Annual Meeting of the American Society of Clinical Oncology (ASCO).
Results. Prior to the ASCO meeting, 57% and 56% of the oncologists in the two research meetings, respectively, indicated "awareness" of the data release. Less than 25% selected an accurate descriptor of the released information from a short list of plausible options. After the ASCO meeting, the figures were 88% and 34%. Over 50% said they plan to use bevacizumab in NSCLC treatment as soon as reimbursement is secure. Eighty-two percent said they plan to use it in second- or third-line treatment; 56% said they plan to use it during second-line chemotherapy despite progression during first-line use. A large majority intend to use bevacizumab in dosages, tumor types, drug combinations, and/or patients not specifically supported by phase III data.
Conclusion. Release of clinically relevant phase III data through electronic and print media is a poor vehicle for informing U.S. medical oncologists. For a commercially available agent, this can have important implications for potential use in untested and potentially unsafe clinical settings. Effective educational strategies for dealing with the new paradigm of "instant" release of clinical data need to be developed.
INTRODUCTION
Practicing oncologists are expected to assimilate an increasingly large amount of rapidly evolving clinical data. Patients, families, and loved ones hope that every cancer treatment will maximize survival. At the same time, they expect their physicians to control side effects to the greatest extent possible. A large segment of oncology clinical practice is off-label, allowing physicians to prescribe treatment in the absence of level 1 evidence. Individual clinicians need to keep current on a veritable flood of new clinical information and integrate the data into their daily practice. Because U.S. oncologists work in a very dynamic environment, we hypothesized that they would rapidly familiarize themselves with practice-relevant, phase III clinical trial data on new therapy options whenever such information became publicly available. The introduction of bevacizumab as a core component of colon cancer therapy is a case in point.
Bevacizumab is a humanized monoclonal antibody that binds vascular endothelial growth factor. It is the first antiangiogenic molecule to be approved by the U.S. Food and Drug Administration (FDA) for cancer treatment. Its approval was based on findings from a Genentech-sponsored, pivotal phase III trial involving more than 800 patients with advanced colorectal cancer [1]. In that trial, the addition of i.v. bevacizumab given every other week to the standard chemotherapy regimen of irinotecan -fluorouracil (5-FU), and leucovorin resulted in significantly longer median and overall survival times than with chemotherapy alone. These results, and the subsequent FDA approval of bevacizumab, had a rapid and profound impact on clinical practice. Within months, the use of bevacizumab as first-line treatment in combination with a 5-FU–based chemotherapy regimen became standard. By 2005, essentially every practicing oncologist in the U.S. who cared for patients with colorectal cancer had used the drug [M. Green, personal communication].
The theoretical benefits of bevacizumab therapy are not limited to any single solid tumor. In 1999, the Eastern Cooperative Oncology Group (ECOG) began development of a phase III clinical trial (ECOG 4599) [2] assessing the impact on survival of adding bevacizumab to chemotherapy in patients with previously untreated advanced non-small cell lung cancer (NSCLC). Patients were randomized to the ECOG standard chemotherapy regimen of carboplatin (Paraplatin®; Bristol-Myers Squibb, Princeton, NJ and paclitaxel (Taxol®; Bristol-Myers Squibb) alone or with bevacizumab. The eligibility criteria and dosing schedules used in ECOG 4599 were influenced by a previous NSCLC phase II trial of bevacizumab and chemotherapy in which four patients with squamous cell carcinoma of the lung had a fatal pulmonary hemorrhage [3]. Thus, squamous cell patients were excluded from the phase III trial. In addition, patients with brain metastases or prior/ongoing anticoagulation therapy were also excluded. The bevacizumab dose of 15 mg/kg i.v. every 3 weeks on the same day as chemotherapy was based on favorable response and survival outcomes in the phase II trial.
In March 2005, the ECOG Data Monitoring Committee (DMC) for study ECOG 4599 reviewed the confidential data from a planned interim analysis of the trial data. The primary end point of ECOG 4599 was overall survival. The interim analysis showed a difference in survival favoring bevacizumab plus chemotherapy. The level of significance (p = .007) crossed a prespecified boundary for data release. The ECOG DMC recommended that the results be released immediately, prior to achieving the prespecified number of events required for the final trial analysis. On March 14, 2005, two findings from the trial (significant overall survival benefit and median survival times of 10.2 vs. 12.5 months) were made public via the National Cancer Institute website [4] and a press announcement from Genentech [5], the manufacturer of bevacizumab (see addenda). The findings were widely disseminated by news media within hours of the data release [6]. Within a few days, some large health care insurers had indicated that they would reimburse off-label use of bevacizumab for the care of patients with lung cancer.
Given the very rapid integration of bevacizumab into colorectal cancer therapy, we hypothesized that U.S. oncologists rapidly familiarize themselves with new, practice-relevant, phase III clinical trial data. We were interested in testing this hypothesis in relation to the release of phase III data on the role of bevacizumab in patients with advanced NSCLC. The test set initially included approximately 210 academic and community-based oncologists who were queried concerning their awareness of these data. The assessments were made in two live research meetings held within 1 and 3 weeks after the data release. Approximately 2 months later, immediately after the 2005 Annual Meeting of the American Society of Clinical Oncology (ASCO), we did a similar assessment in a third cadre of practicing oncologists (Fig. 1). Our findings from all three groups are reported below.
METHODS
Thousands of U.S. medical oncologists have participated in 1-day clinical case–based oncology research meetings organized by the Network for Medical Communication and Research (NMCR). Attendees self-select for participation, registering via the NMCR website. They are reimbursed for meeting expenses, including travel costs and specific out-of-pocket expenses, from revenues paid to NMCR by several pharmaceutical sponsors. Attendees do not receive honoraria. Meeting topics range from one disease (e.g., prostate cancer) to an entire field (e.g., hematology). Research data are acquired anonymously using a wireless electronic scoring system. Research questions are based on real-life clinical scenarios. The question-and-answer (Q&A) options are projected and read by the meeting moderator. Typically, there are four to eight response choices for each question. Default options such as "other" or "I don’t know" are sometimes included. Once the attendees have made their selection via the wireless keypad, the results are immediately tallied and projected in bar graph format. This is followed by expert commentary on the core clinical issue addressed in the question and then an active Q&A session involving all meeting participants.
The two pre-ASCO research meetings were held on March 19 and April 2, 2005. Specific goals of the bevacizumab research project were fourfold: (a) to assess physician awareness of a "release of material information" for an approved drug in new therapeutic indication; (b) to determine oncologists’ depth of awareness of the new data; (c) to explore physicians’ plans for applying the new data in their own practices; and (d) to profile the potential for the use of bevacizumab in practice settings not specifically addressed by the clinical trial findings. The third meeting was held on June 2, 2005, approximately 3 weeks after the ASCO Annual Meeting. Once again approximately 100 oncologists participated. The research goals were similar, but the surveyed physicians had had the potential for substantial additional exposure to the actual clinical data and extensive "expert" commentary.
The three groups of oncologists had largely similar demographics. In each group nearly three fourths identified themselves as being in community rather than academic practice. Approximately 25% were within 5 years of completion of their oncology training, 25% were within 5–10 years, 25% were within 10–20 years, and 25% had over 20 years of practice experience. The first two cohorts included approximately 5% solo practitioners. In the third cohort, this figure was 25%.
The research questions (or available answers) evolved during the assessment process. In the two pre-ASCO sessions, eight questions were used. To allow the question process to move forward with the participation of all attendees, stepwise information about the trial results and data release were provided by the meeting moderator after each question had been answered by the attendees. At the first meeting, several questions required "forced choices," because a default option was not included. Default choices were added in the subsequent meetings. The questions are shown in Table 1. The June assessment focused on questions 1, 3, and 8.
RESULTS
In group 1, studied less than 1 week after data were announced, 57% of the oncologists reported "awareness" of the data release. In the second group, awareness was essentially identical, at 56%. After the ASCO Annual Meeting, awareness rose to 88% (Fig. 2).
In a follow-up question about what data were actually revealed in the announcements (see addenda), two thirds or more of all respondents (including at the post-ASCO session) either indicated that they did not know what had been reported or recalled the data incorrectly (Fig. 3). At the first meeting, where the available answers "required" a forced choice (or abstention by not responding), 21% selected a completely irrelevant distractor "the study was stopped early because of toxicity." When an "I do not know" option was added in the April session, 43% selected that option. The default option was selected by only 12% during the post-ASCO session, and the accuracy of data recall rose to 34% among the third session attendees.
Data Integration and Use in Clinical Practice On- and Off-Label
Bevacizumab is commercially available and "indicated" for the care of patients with colorectal cancer. U.S. oncologists can consider using it off-label for lung cancer patients. Therefore, we explored when and how the meeting attendees were planning to integrate bevacizumab into their daily clinical practice. When asked when they planned to begin using bevacizumab as part of the first-line therapy for NSCLC patients, 9% of respondents in the March 19 study group said they were either doing it already or would begin immediately. Two weeks later, the figure was 4%. In all three meetings, over 50% of the respondents cited financial considerations as a critical driver of use, indicating that they would begin integrating bevacizumab as soon as reimbursement was secure (Table 2). Furthermore, 69% said that as soon as bevacizumab was available and reimbursed, they would add it to chemotherapy in patients already getting first-line treatment with chemotherapy alone. Fifty-six percent planned to continue using bevacizumab during second-line chemotherapy even after progression during first-line use. Eighty-two percent said they would use it (combined with other systemic therapy) as second- or third-line treatment in patients who had not received it first line (Fig. 4).
A large majority of participating oncologists reported their intent to use bevacizumab in dosages, tumor types, drug combinations, and/or patients not specifically supported by the available phase III data. At the March and April meetings, 82% and 85% of attendees, respectively, reported a comfort level for using bevacizumab in combination with chemotherapy regimens other than the one used in the pivotal ECOG trial. In June, after the ASCO presentation, only 43% of the oncologists reported considering the use of a combination other than that used in the pivotal study (Table 3).
Patients with squamous cell carcinoma were specifically excluded from the phase III trial of paclitaxel–carboplatin–bevacizumab in NSCLC. Because of concern for intracranial hemorrhage, patients with known brain metastases were also excluded. Over the course of the three assessment sessions, an increasing number of respondents reported they would use bevacizumab therapy only in patients with "ECOG study eligible" characteristics. Nonetheless, even in the June assessment, 38% of respondents indicated plans to use bevacizumab in settings in which the bleeding risk profile had not been definitively tested (Table 4).
DISCUSSION
The preliminary results of ECOG 4599, a pivotal phase III clinical trial comparing paclitaxel and carboplatin with or without bevacizumab as first-line therapy for patients with advanced, nonsquamous, non-small cell lung cancer, were released to the public media on March 14, 2005 [4, 5]. Relevant eligibility criteria for study entry and some preliminary toxicity information were also provided. The study demonstrated a statistically significant difference in overall survival favoring bevacizumab plus chemotherapy [2]. Extensive media coverage followed [6]. Based on the preliminary reports, some health insurers promptly announced plans to reimburse physicians for the use of bevacizumab in selected lung cancer patients [7]. Within 2 months of the first information release, a formal presentation of the study data took place at a plenary session of the 2005 ASCO Annual Meeting [3]. Once again, media coverage of the presentation and the findings was extensive.
We hypothesized that nearly all U.S. oncology physicians would rapidly become familiar with such new, statistically significant, clinical trial data involving a high-profile, commercially available, anticancer agent. We presumed that awareness would be high regardless of the specific medium of information delivery (e.g., large-scale medical meeting, publication in peer-reviewed medical journal, or the news media and financial press). Further, we expected that the release of such data would lead to rapid integration of the new therapy into evidence-based treatment focused on appropriately selected patients. Our survey of three separate groups of experienced medical oncology physicians concerning their familiarity with the bevacizumab clinical trial data in NSCLC and their plans to use bevacizumab for their lung cancer patients failed to support our pretest hypotheses.
At the first two meetings, completed within 3 weeks of the release of the preliminary data, just over half of the oncologists indicated that they were "aware" of the data release. Less than 25% were able to select an accurate descriptor of the released information from a short list of plausible options. Approximately 10 weeks later, just after the 2005 ASCO Annual Meeting, we used the live meeting format to survey a third group of approximately 100 U.S. medical oncologists. Even after the ASCO Annual Meeting presentation and attendant publicity, a large majority of the physicians surveyed were not fully aware of the relevant clinical findings. A small cadre still reported that they were unaware that phase III data on the subject had been reported.
Despite being unable to select the accurate descriptor of the actual findings of the ECOG trial, most respondents were able to delineate their plans for use of bevacizumab in their lung cancer patients. A majority reported that they were ready to use the drug as soon as reimbursement for its use in lung cancer treatment was secure. Even after the formal ASCO presentation, 43% indicated they would be comfortable combining bevacizumab with a chemotherapy regimen other than paclitaxel–carboplatin (the regimen used in ECOG 4599). Thirty-eight percent reported plans to treat at least some patients with types (squamous cell) or patterns (brain metastases) of NSCLC specifically excluded from the ECOG 4599 study population. Plans for extending the use of bevacizumab to second- and third-line therapy were common.
These findings highlight just some of the challenges physicians face in the lifelong learning process. Gaining simple awareness, let alone mastery over the constant flow of new clinical information and its implications for clinical practice, is a formidable task. Our initial testing suggests that public media releases are a poor vehicle for informing doctors about new clinical data. An awareness level of barely 50% and an accurate data recall rate of <25% are far below what we had anticipated. Repetitive exposure to clinical results does appear to increase awareness. However, even after the ASCO Annual Meeting, only a minority of respondents had sufficient functional recall of the data to select the correct articulation of the key findings. Of greatest concern was the apparent comfort with the use of bevacizumab in patients and settings completely unexplored in the pivotal clinical trial. These findings highlight the need for more effective strategies for timely, focused, relevant, accessible, continuing medical education in dynamic fields like medical oncology and very likely in every other area of medicine in the U.S.
In the U.S., sharing off-label product information with medical practitioners and other stakeholders is carefully regulated by the FDA. Compliance with these regulations is an extremely high priority for the pharmaceutical industry and for medical education providers. These guidelines, while understandable and necessary, may impact the safe application of new, "late-breaking" medical information. This may be particularly relevant in cases in which the new data relate to a potentially expanded indication for an already approved drug. Clearly, in the "Internet Age," in which more and more is "available" if you know where to find it, the data are there. But making the time to do the digging, or spending the money to have it "delivered" to your electronic front door, is only the beginning of the complex learning process. Our data suggest that "just" hearing or reading second-hand reports of new clinical trial data, especially as presented in nonclinical publications, does not support the critical learning needs of busy clinicians. The lack of information may even increase risk to patients during early adoption of off-label use. Effective educational strategies for dealing with the new paradigm of "instant" release of clinical data through media-based or electronic outlets need to be developed.
ADDENDA
Bevacizumab Combined with Chemotherapy Prolongs Survival for Some Patients with Advanced Lung Cancer [4]
Preliminary results from a large, randomized clinical trial for patients with previously untreated advanced nonsquamous, non-small cell lung cancer show that those patients who received bevacizumab (Avastin®) in combination with standard chemotherapy lived longer than patients who received the same chemotherapy without bevacizumab.
The clinical trial was sponsored by the National Cancer Institute (NCI), part of the National Institutes of Health, and conducted by a network of researchers led by the Eastern Cooperative Oncology Group. Genentech, Inc., South San Francisco, CA, who manufactures bevacizumab, provided bevacizumab for the trial under the Cooperative Research and Development Agreement (CRADA) with NCI for the clinical development of bevacizumab.
The Data Monitoring Committee overseeing the trial (known as E4599) recommended that the results of a recent interim analysis be made public because the study had met its primary end point of improving overall survival. Researchers found that patients in the study who received bevacizumab in combination with standard chemotherapy (a treatment regimen of paclitaxel and carboplatin) had a median overall survival time of 12.5 months compared with patients treated with the standard chemotherapy alone, who had a median survival time of 10.2 months. This difference is statistically significant.
Interim Analysis of Phase III Trial Shows Avastin® Plus Chemotherapy Extends Survival of Patients with First-Line NonSquamous, NonSmall Cell Lung Cancer [5]
First Positive Phase III Results with an Antiangiogenesis Therapy in Lung Cancer
South SanFrancisco, CA and Basel, Switzerland–March 14, 2005--Genentech, Inc. (NYSE: DNA) and Roche (SWX Zurich) today announced that an interim analysis of a phase III study of Avastin® (bevacizumab) plus paclitaxel and carboplatin chemotherapies in first-line nonsquamous, non-small cell lung cancer (NSCLC) met its primary efficacy end point of improving overall survival, or a reduction in the risk of death, compared with chemotherapy alone. The trial was sponsored by the National Cancer Institute (NCI), part of the National Institutes of Health, under a Cooperative Research and Development Agreement between NCI and Genentech, Inc., and conducted by a network of researchers led by the Eastern Cooperative Oncology Group (ECOG). According to ECOG, data from this study will be submitted to the annual meeting of the American Society of Clinical Oncology (ASCO), May 13–17, 2005.
"These results represent the first study combining a targeted biologic therapy with chemotherapy to show an overall survival improvement in the first-line non-small cell lung cancer setting, and the first time that any treatment has improved upon the standard, two-drug chemotherapy regimen in this disease," said Hal Barron, M.D., Genen-tech senior vice president, development and chief medical officer. "We would like to thank our collaborators at NCI and ECOG for their work on this study, as well as the many patients and their families who made the decision to participate in this study. We plan to share these data with the FDA to discuss the possibility of filing a supplemental Biologics License Application for Avastin® plus chemotherapy in first-line non-small cell lung cancer."
This is the first phase III study to evaluate the therapeutic antibody Avastin® in combination with chemotherapy in NSCLC. This was a randomized, controlled, multicenter trial that enrolled 878 patients with previously untreated advanced NSCLC. The patients enrolled in this trial were randomized to receive treatment with paclitaxel and carboplatin chemotherapies with or without Avastin®.
Adverse Events
In previous clinical experience with Avastin® in combination with paclitaxel and carboplatin in NSCLC, life-threatening or fatal pulmonary bleeding was identified as a severe adverse event apparently unique to this disease. Certain characteristics, including any significant pulmonary bleeding prior to receiving treatment with Avastin® or the presence of a specific type of NSCLC called squamous cell carcinoma appeared to predispose patients to experiencing this adverse event. Patients with these characteristics were excluded from this phase III study, and the rate of life-threatening or fatal pulmonary bleeding was substantially reduced from prior clinical studies. However, some patients did experience fatal pulmonary bleeding in this trial and this event was more common in the patient group that received Avastin® in combination with chemotherapy than in the patient group that received chemotherapy only. Other adverse events observed in this study were similar to those identified in previous phase II and phase III studies of Avastin®. More detailed information about adverse events in this study will be presented at the ASCO meeting in May.
Genentech Drug Shows Effect on a 2nd Cancer [6]
By Andrew Pollack
Genentech Inc.’s cancer drug Avastin® prolonged the lives of patients with advanced lung cancer in a clinical trial, a result that sent the company’s shares soaring and could pave the way for wider use of what is widely considered the company’s most promising product.
In the trial, sponsored by the National Cancer Institute, patients who received Avastin® along with standard chemotherapy drugs had a median survival of 12.5 months, compared with 10.2 months for those receiving only the chemotherapy.
In the trial, 878 patients received either the chemotherapy drugs paclitaxel and carboplatin, or those drugs plus Avastin®.
The biggest side effect was fatal or life-threatening bleeding from the lungs. The National Cancer Institute said this occurred "infrequently." Dr. Scott Saxman of the institute said that meant it happened in fewer than 3% of the patients.
But the deaths and bleeding cases occurred despite efforts to minimize them by excluding certain types of patients who, in an earlier trial, had seemed most prone to the side effect. The exclusions, of people with a type of cancer called squamous cell carcinoma and of patients whose cancer had spread to the brain, would amount to roughly one third of patients with advanced non-small cell lung cancer, specialists estimated.
Doctors and Genentech officials said they would not discuss the results in detail. They said they wanted to present them in May at the annual meeting of the American Society of Clinical Oncology, and that the society would not allow presentation of results that had previously been presented.
Copyright ©2005 by The New York Times Co. Reprinted with permission.
DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
Dr. Green is an employee of the Network for Medical Communication and Research and receives honoraria from and is a consultant for Genentech. All other authors are also employees of the Network for Medical Communication and Research. Carmen Allegra is a consultant for Genentech.
REFERENCES
Hurwitz H, Fehrenbacher L, Novotny W et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004;350:2335–2342.
Sandler AB, Gray R, Brahmer J et al. Randomized phase II/III trial of paclitaxel plus carboplatin with or without bevacizumab in patients with advanced, non-squamous non-small cell lung cancer. An Eastern Cooperative Oncology Group trial – E4599. Proc Am Soc Clin Oncol 2005;23:1090s.
Johnson DH, Fehrenbacher L, Novotny WF et al. Randomized phase II trial comparing bevacizumab plus carboplatin and paclitaxel with carboplatin and paclitaxel alone in previously untreated locally advanced or metastatic non-small-cell lung cancer. J Clin Oncol 2004;22:2184–2191.
National Cancer Institute. Bevacizumab Combined With Chemotherapy Prolongs Survival for Some Patients with Advanced Lung Cancer, March 14, 2005.
Genentech. Interim Analysis of Phase III Trial Shows Avastin Plus Chemotherapy Extends Survival of Patients with First-Line Non-Squamous, Non-Small Cell Lung Cancer March 14, 2005.
Pollack A. Genentech Drug Shows Effect on a 2nd Cancer. The New York Times, March 15, 2005.
Blue Cross to Pay for Life-Prolonging Cancer Drug. Newsday.com, March 16,2005.(Daniel Dornbuscha, Carmen)
After completing this course, the reader will be able to:
Discuss new vehicles for the initial release of practice-relevant phase III clinical trial data.
Describe the limited educational impact of such information transfer vehicles on the medical oncology community.
Explain the potential risks to medical oncologists and their patients because of underinformed or non–evidence-based therapeutic recommendations involving new indications for commercially available agents.
Discuss the development of new strategies for effective knowledge transfer of preliminary, practice-relevant, clinical trial data.
ABSTRACT
Background. Practicing oncologists are expected to easily assimilate large amounts of rapidly evolving clinical data. We hypothesized that U.S. oncologists rapidly familiarize themselves with new, practice-relevant, phase III clinical trial data. We tested this hypothesis in relation to the release of phase III data from the Eastern Cooperative Oncology Group 4599 trial on the role of bevacizumab in advanced non-small cell lung cancer (NSCLC).
Methods. We queried approximately 310 medical oncologists concerning their awareness of the bevacizumab data within 1 and 3 weeks after the data release or immediately after the 2005 Annual Meeting of the American Society of Clinical Oncology (ASCO).
Results. Prior to the ASCO meeting, 57% and 56% of the oncologists in the two research meetings, respectively, indicated "awareness" of the data release. Less than 25% selected an accurate descriptor of the released information from a short list of plausible options. After the ASCO meeting, the figures were 88% and 34%. Over 50% said they plan to use bevacizumab in NSCLC treatment as soon as reimbursement is secure. Eighty-two percent said they plan to use it in second- or third-line treatment; 56% said they plan to use it during second-line chemotherapy despite progression during first-line use. A large majority intend to use bevacizumab in dosages, tumor types, drug combinations, and/or patients not specifically supported by phase III data.
Conclusion. Release of clinically relevant phase III data through electronic and print media is a poor vehicle for informing U.S. medical oncologists. For a commercially available agent, this can have important implications for potential use in untested and potentially unsafe clinical settings. Effective educational strategies for dealing with the new paradigm of "instant" release of clinical data need to be developed.
INTRODUCTION
Practicing oncologists are expected to assimilate an increasingly large amount of rapidly evolving clinical data. Patients, families, and loved ones hope that every cancer treatment will maximize survival. At the same time, they expect their physicians to control side effects to the greatest extent possible. A large segment of oncology clinical practice is off-label, allowing physicians to prescribe treatment in the absence of level 1 evidence. Individual clinicians need to keep current on a veritable flood of new clinical information and integrate the data into their daily practice. Because U.S. oncologists work in a very dynamic environment, we hypothesized that they would rapidly familiarize themselves with practice-relevant, phase III clinical trial data on new therapy options whenever such information became publicly available. The introduction of bevacizumab as a core component of colon cancer therapy is a case in point.
Bevacizumab is a humanized monoclonal antibody that binds vascular endothelial growth factor. It is the first antiangiogenic molecule to be approved by the U.S. Food and Drug Administration (FDA) for cancer treatment. Its approval was based on findings from a Genentech-sponsored, pivotal phase III trial involving more than 800 patients with advanced colorectal cancer [1]. In that trial, the addition of i.v. bevacizumab given every other week to the standard chemotherapy regimen of irinotecan -fluorouracil (5-FU), and leucovorin resulted in significantly longer median and overall survival times than with chemotherapy alone. These results, and the subsequent FDA approval of bevacizumab, had a rapid and profound impact on clinical practice. Within months, the use of bevacizumab as first-line treatment in combination with a 5-FU–based chemotherapy regimen became standard. By 2005, essentially every practicing oncologist in the U.S. who cared for patients with colorectal cancer had used the drug [M. Green, personal communication].
The theoretical benefits of bevacizumab therapy are not limited to any single solid tumor. In 1999, the Eastern Cooperative Oncology Group (ECOG) began development of a phase III clinical trial (ECOG 4599) [2] assessing the impact on survival of adding bevacizumab to chemotherapy in patients with previously untreated advanced non-small cell lung cancer (NSCLC). Patients were randomized to the ECOG standard chemotherapy regimen of carboplatin (Paraplatin®; Bristol-Myers Squibb, Princeton, NJ and paclitaxel (Taxol®; Bristol-Myers Squibb) alone or with bevacizumab. The eligibility criteria and dosing schedules used in ECOG 4599 were influenced by a previous NSCLC phase II trial of bevacizumab and chemotherapy in which four patients with squamous cell carcinoma of the lung had a fatal pulmonary hemorrhage [3]. Thus, squamous cell patients were excluded from the phase III trial. In addition, patients with brain metastases or prior/ongoing anticoagulation therapy were also excluded. The bevacizumab dose of 15 mg/kg i.v. every 3 weeks on the same day as chemotherapy was based on favorable response and survival outcomes in the phase II trial.
In March 2005, the ECOG Data Monitoring Committee (DMC) for study ECOG 4599 reviewed the confidential data from a planned interim analysis of the trial data. The primary end point of ECOG 4599 was overall survival. The interim analysis showed a difference in survival favoring bevacizumab plus chemotherapy. The level of significance (p = .007) crossed a prespecified boundary for data release. The ECOG DMC recommended that the results be released immediately, prior to achieving the prespecified number of events required for the final trial analysis. On March 14, 2005, two findings from the trial (significant overall survival benefit and median survival times of 10.2 vs. 12.5 months) were made public via the National Cancer Institute website [4] and a press announcement from Genentech [5], the manufacturer of bevacizumab (see addenda). The findings were widely disseminated by news media within hours of the data release [6]. Within a few days, some large health care insurers had indicated that they would reimburse off-label use of bevacizumab for the care of patients with lung cancer.
Given the very rapid integration of bevacizumab into colorectal cancer therapy, we hypothesized that U.S. oncologists rapidly familiarize themselves with new, practice-relevant, phase III clinical trial data. We were interested in testing this hypothesis in relation to the release of phase III data on the role of bevacizumab in patients with advanced NSCLC. The test set initially included approximately 210 academic and community-based oncologists who were queried concerning their awareness of these data. The assessments were made in two live research meetings held within 1 and 3 weeks after the data release. Approximately 2 months later, immediately after the 2005 Annual Meeting of the American Society of Clinical Oncology (ASCO), we did a similar assessment in a third cadre of practicing oncologists (Fig. 1). Our findings from all three groups are reported below.
METHODS
Thousands of U.S. medical oncologists have participated in 1-day clinical case–based oncology research meetings organized by the Network for Medical Communication and Research (NMCR). Attendees self-select for participation, registering via the NMCR website. They are reimbursed for meeting expenses, including travel costs and specific out-of-pocket expenses, from revenues paid to NMCR by several pharmaceutical sponsors. Attendees do not receive honoraria. Meeting topics range from one disease (e.g., prostate cancer) to an entire field (e.g., hematology). Research data are acquired anonymously using a wireless electronic scoring system. Research questions are based on real-life clinical scenarios. The question-and-answer (Q&A) options are projected and read by the meeting moderator. Typically, there are four to eight response choices for each question. Default options such as "other" or "I don’t know" are sometimes included. Once the attendees have made their selection via the wireless keypad, the results are immediately tallied and projected in bar graph format. This is followed by expert commentary on the core clinical issue addressed in the question and then an active Q&A session involving all meeting participants.
The two pre-ASCO research meetings were held on March 19 and April 2, 2005. Specific goals of the bevacizumab research project were fourfold: (a) to assess physician awareness of a "release of material information" for an approved drug in new therapeutic indication; (b) to determine oncologists’ depth of awareness of the new data; (c) to explore physicians’ plans for applying the new data in their own practices; and (d) to profile the potential for the use of bevacizumab in practice settings not specifically addressed by the clinical trial findings. The third meeting was held on June 2, 2005, approximately 3 weeks after the ASCO Annual Meeting. Once again approximately 100 oncologists participated. The research goals were similar, but the surveyed physicians had had the potential for substantial additional exposure to the actual clinical data and extensive "expert" commentary.
The three groups of oncologists had largely similar demographics. In each group nearly three fourths identified themselves as being in community rather than academic practice. Approximately 25% were within 5 years of completion of their oncology training, 25% were within 5–10 years, 25% were within 10–20 years, and 25% had over 20 years of practice experience. The first two cohorts included approximately 5% solo practitioners. In the third cohort, this figure was 25%.
The research questions (or available answers) evolved during the assessment process. In the two pre-ASCO sessions, eight questions were used. To allow the question process to move forward with the participation of all attendees, stepwise information about the trial results and data release were provided by the meeting moderator after each question had been answered by the attendees. At the first meeting, several questions required "forced choices," because a default option was not included. Default choices were added in the subsequent meetings. The questions are shown in Table 1. The June assessment focused on questions 1, 3, and 8.
RESULTS
In group 1, studied less than 1 week after data were announced, 57% of the oncologists reported "awareness" of the data release. In the second group, awareness was essentially identical, at 56%. After the ASCO Annual Meeting, awareness rose to 88% (Fig. 2).
In a follow-up question about what data were actually revealed in the announcements (see addenda), two thirds or more of all respondents (including at the post-ASCO session) either indicated that they did not know what had been reported or recalled the data incorrectly (Fig. 3). At the first meeting, where the available answers "required" a forced choice (or abstention by not responding), 21% selected a completely irrelevant distractor "the study was stopped early because of toxicity." When an "I do not know" option was added in the April session, 43% selected that option. The default option was selected by only 12% during the post-ASCO session, and the accuracy of data recall rose to 34% among the third session attendees.
Data Integration and Use in Clinical Practice On- and Off-Label
Bevacizumab is commercially available and "indicated" for the care of patients with colorectal cancer. U.S. oncologists can consider using it off-label for lung cancer patients. Therefore, we explored when and how the meeting attendees were planning to integrate bevacizumab into their daily clinical practice. When asked when they planned to begin using bevacizumab as part of the first-line therapy for NSCLC patients, 9% of respondents in the March 19 study group said they were either doing it already or would begin immediately. Two weeks later, the figure was 4%. In all three meetings, over 50% of the respondents cited financial considerations as a critical driver of use, indicating that they would begin integrating bevacizumab as soon as reimbursement was secure (Table 2). Furthermore, 69% said that as soon as bevacizumab was available and reimbursed, they would add it to chemotherapy in patients already getting first-line treatment with chemotherapy alone. Fifty-six percent planned to continue using bevacizumab during second-line chemotherapy even after progression during first-line use. Eighty-two percent said they would use it (combined with other systemic therapy) as second- or third-line treatment in patients who had not received it first line (Fig. 4).
A large majority of participating oncologists reported their intent to use bevacizumab in dosages, tumor types, drug combinations, and/or patients not specifically supported by the available phase III data. At the March and April meetings, 82% and 85% of attendees, respectively, reported a comfort level for using bevacizumab in combination with chemotherapy regimens other than the one used in the pivotal ECOG trial. In June, after the ASCO presentation, only 43% of the oncologists reported considering the use of a combination other than that used in the pivotal study (Table 3).
Patients with squamous cell carcinoma were specifically excluded from the phase III trial of paclitaxel–carboplatin–bevacizumab in NSCLC. Because of concern for intracranial hemorrhage, patients with known brain metastases were also excluded. Over the course of the three assessment sessions, an increasing number of respondents reported they would use bevacizumab therapy only in patients with "ECOG study eligible" characteristics. Nonetheless, even in the June assessment, 38% of respondents indicated plans to use bevacizumab in settings in which the bleeding risk profile had not been definitively tested (Table 4).
DISCUSSION
The preliminary results of ECOG 4599, a pivotal phase III clinical trial comparing paclitaxel and carboplatin with or without bevacizumab as first-line therapy for patients with advanced, nonsquamous, non-small cell lung cancer, were released to the public media on March 14, 2005 [4, 5]. Relevant eligibility criteria for study entry and some preliminary toxicity information were also provided. The study demonstrated a statistically significant difference in overall survival favoring bevacizumab plus chemotherapy [2]. Extensive media coverage followed [6]. Based on the preliminary reports, some health insurers promptly announced plans to reimburse physicians for the use of bevacizumab in selected lung cancer patients [7]. Within 2 months of the first information release, a formal presentation of the study data took place at a plenary session of the 2005 ASCO Annual Meeting [3]. Once again, media coverage of the presentation and the findings was extensive.
We hypothesized that nearly all U.S. oncology physicians would rapidly become familiar with such new, statistically significant, clinical trial data involving a high-profile, commercially available, anticancer agent. We presumed that awareness would be high regardless of the specific medium of information delivery (e.g., large-scale medical meeting, publication in peer-reviewed medical journal, or the news media and financial press). Further, we expected that the release of such data would lead to rapid integration of the new therapy into evidence-based treatment focused on appropriately selected patients. Our survey of three separate groups of experienced medical oncology physicians concerning their familiarity with the bevacizumab clinical trial data in NSCLC and their plans to use bevacizumab for their lung cancer patients failed to support our pretest hypotheses.
At the first two meetings, completed within 3 weeks of the release of the preliminary data, just over half of the oncologists indicated that they were "aware" of the data release. Less than 25% were able to select an accurate descriptor of the released information from a short list of plausible options. Approximately 10 weeks later, just after the 2005 ASCO Annual Meeting, we used the live meeting format to survey a third group of approximately 100 U.S. medical oncologists. Even after the ASCO Annual Meeting presentation and attendant publicity, a large majority of the physicians surveyed were not fully aware of the relevant clinical findings. A small cadre still reported that they were unaware that phase III data on the subject had been reported.
Despite being unable to select the accurate descriptor of the actual findings of the ECOG trial, most respondents were able to delineate their plans for use of bevacizumab in their lung cancer patients. A majority reported that they were ready to use the drug as soon as reimbursement for its use in lung cancer treatment was secure. Even after the formal ASCO presentation, 43% indicated they would be comfortable combining bevacizumab with a chemotherapy regimen other than paclitaxel–carboplatin (the regimen used in ECOG 4599). Thirty-eight percent reported plans to treat at least some patients with types (squamous cell) or patterns (brain metastases) of NSCLC specifically excluded from the ECOG 4599 study population. Plans for extending the use of bevacizumab to second- and third-line therapy were common.
These findings highlight just some of the challenges physicians face in the lifelong learning process. Gaining simple awareness, let alone mastery over the constant flow of new clinical information and its implications for clinical practice, is a formidable task. Our initial testing suggests that public media releases are a poor vehicle for informing doctors about new clinical data. An awareness level of barely 50% and an accurate data recall rate of <25% are far below what we had anticipated. Repetitive exposure to clinical results does appear to increase awareness. However, even after the ASCO Annual Meeting, only a minority of respondents had sufficient functional recall of the data to select the correct articulation of the key findings. Of greatest concern was the apparent comfort with the use of bevacizumab in patients and settings completely unexplored in the pivotal clinical trial. These findings highlight the need for more effective strategies for timely, focused, relevant, accessible, continuing medical education in dynamic fields like medical oncology and very likely in every other area of medicine in the U.S.
In the U.S., sharing off-label product information with medical practitioners and other stakeholders is carefully regulated by the FDA. Compliance with these regulations is an extremely high priority for the pharmaceutical industry and for medical education providers. These guidelines, while understandable and necessary, may impact the safe application of new, "late-breaking" medical information. This may be particularly relevant in cases in which the new data relate to a potentially expanded indication for an already approved drug. Clearly, in the "Internet Age," in which more and more is "available" if you know where to find it, the data are there. But making the time to do the digging, or spending the money to have it "delivered" to your electronic front door, is only the beginning of the complex learning process. Our data suggest that "just" hearing or reading second-hand reports of new clinical trial data, especially as presented in nonclinical publications, does not support the critical learning needs of busy clinicians. The lack of information may even increase risk to patients during early adoption of off-label use. Effective educational strategies for dealing with the new paradigm of "instant" release of clinical data through media-based or electronic outlets need to be developed.
ADDENDA
Bevacizumab Combined with Chemotherapy Prolongs Survival for Some Patients with Advanced Lung Cancer [4]
Preliminary results from a large, randomized clinical trial for patients with previously untreated advanced nonsquamous, non-small cell lung cancer show that those patients who received bevacizumab (Avastin®) in combination with standard chemotherapy lived longer than patients who received the same chemotherapy without bevacizumab.
The clinical trial was sponsored by the National Cancer Institute (NCI), part of the National Institutes of Health, and conducted by a network of researchers led by the Eastern Cooperative Oncology Group. Genentech, Inc., South San Francisco, CA, who manufactures bevacizumab, provided bevacizumab for the trial under the Cooperative Research and Development Agreement (CRADA) with NCI for the clinical development of bevacizumab.
The Data Monitoring Committee overseeing the trial (known as E4599) recommended that the results of a recent interim analysis be made public because the study had met its primary end point of improving overall survival. Researchers found that patients in the study who received bevacizumab in combination with standard chemotherapy (a treatment regimen of paclitaxel and carboplatin) had a median overall survival time of 12.5 months compared with patients treated with the standard chemotherapy alone, who had a median survival time of 10.2 months. This difference is statistically significant.
Interim Analysis of Phase III Trial Shows Avastin® Plus Chemotherapy Extends Survival of Patients with First-Line NonSquamous, NonSmall Cell Lung Cancer [5]
First Positive Phase III Results with an Antiangiogenesis Therapy in Lung Cancer
South SanFrancisco, CA and Basel, Switzerland–March 14, 2005--Genentech, Inc. (NYSE: DNA) and Roche (SWX Zurich) today announced that an interim analysis of a phase III study of Avastin® (bevacizumab) plus paclitaxel and carboplatin chemotherapies in first-line nonsquamous, non-small cell lung cancer (NSCLC) met its primary efficacy end point of improving overall survival, or a reduction in the risk of death, compared with chemotherapy alone. The trial was sponsored by the National Cancer Institute (NCI), part of the National Institutes of Health, under a Cooperative Research and Development Agreement between NCI and Genentech, Inc., and conducted by a network of researchers led by the Eastern Cooperative Oncology Group (ECOG). According to ECOG, data from this study will be submitted to the annual meeting of the American Society of Clinical Oncology (ASCO), May 13–17, 2005.
"These results represent the first study combining a targeted biologic therapy with chemotherapy to show an overall survival improvement in the first-line non-small cell lung cancer setting, and the first time that any treatment has improved upon the standard, two-drug chemotherapy regimen in this disease," said Hal Barron, M.D., Genen-tech senior vice president, development and chief medical officer. "We would like to thank our collaborators at NCI and ECOG for their work on this study, as well as the many patients and their families who made the decision to participate in this study. We plan to share these data with the FDA to discuss the possibility of filing a supplemental Biologics License Application for Avastin® plus chemotherapy in first-line non-small cell lung cancer."
This is the first phase III study to evaluate the therapeutic antibody Avastin® in combination with chemotherapy in NSCLC. This was a randomized, controlled, multicenter trial that enrolled 878 patients with previously untreated advanced NSCLC. The patients enrolled in this trial were randomized to receive treatment with paclitaxel and carboplatin chemotherapies with or without Avastin®.
Adverse Events
In previous clinical experience with Avastin® in combination with paclitaxel and carboplatin in NSCLC, life-threatening or fatal pulmonary bleeding was identified as a severe adverse event apparently unique to this disease. Certain characteristics, including any significant pulmonary bleeding prior to receiving treatment with Avastin® or the presence of a specific type of NSCLC called squamous cell carcinoma appeared to predispose patients to experiencing this adverse event. Patients with these characteristics were excluded from this phase III study, and the rate of life-threatening or fatal pulmonary bleeding was substantially reduced from prior clinical studies. However, some patients did experience fatal pulmonary bleeding in this trial and this event was more common in the patient group that received Avastin® in combination with chemotherapy than in the patient group that received chemotherapy only. Other adverse events observed in this study were similar to those identified in previous phase II and phase III studies of Avastin®. More detailed information about adverse events in this study will be presented at the ASCO meeting in May.
Genentech Drug Shows Effect on a 2nd Cancer [6]
By Andrew Pollack
Genentech Inc.’s cancer drug Avastin® prolonged the lives of patients with advanced lung cancer in a clinical trial, a result that sent the company’s shares soaring and could pave the way for wider use of what is widely considered the company’s most promising product.
In the trial, sponsored by the National Cancer Institute, patients who received Avastin® along with standard chemotherapy drugs had a median survival of 12.5 months, compared with 10.2 months for those receiving only the chemotherapy.
In the trial, 878 patients received either the chemotherapy drugs paclitaxel and carboplatin, or those drugs plus Avastin®.
The biggest side effect was fatal or life-threatening bleeding from the lungs. The National Cancer Institute said this occurred "infrequently." Dr. Scott Saxman of the institute said that meant it happened in fewer than 3% of the patients.
But the deaths and bleeding cases occurred despite efforts to minimize them by excluding certain types of patients who, in an earlier trial, had seemed most prone to the side effect. The exclusions, of people with a type of cancer called squamous cell carcinoma and of patients whose cancer had spread to the brain, would amount to roughly one third of patients with advanced non-small cell lung cancer, specialists estimated.
Doctors and Genentech officials said they would not discuss the results in detail. They said they wanted to present them in May at the annual meeting of the American Society of Clinical Oncology, and that the society would not allow presentation of results that had previously been presented.
Copyright ©2005 by The New York Times Co. Reprinted with permission.
DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
Dr. Green is an employee of the Network for Medical Communication and Research and receives honoraria from and is a consultant for Genentech. All other authors are also employees of the Network for Medical Communication and Research. Carmen Allegra is a consultant for Genentech.
REFERENCES
Hurwitz H, Fehrenbacher L, Novotny W et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004;350:2335–2342.
Sandler AB, Gray R, Brahmer J et al. Randomized phase II/III trial of paclitaxel plus carboplatin with or without bevacizumab in patients with advanced, non-squamous non-small cell lung cancer. An Eastern Cooperative Oncology Group trial – E4599. Proc Am Soc Clin Oncol 2005;23:1090s.
Johnson DH, Fehrenbacher L, Novotny WF et al. Randomized phase II trial comparing bevacizumab plus carboplatin and paclitaxel with carboplatin and paclitaxel alone in previously untreated locally advanced or metastatic non-small-cell lung cancer. J Clin Oncol 2004;22:2184–2191.
National Cancer Institute. Bevacizumab Combined With Chemotherapy Prolongs Survival for Some Patients with Advanced Lung Cancer, March 14, 2005.
Genentech. Interim Analysis of Phase III Trial Shows Avastin Plus Chemotherapy Extends Survival of Patients with First-Line Non-Squamous, Non-Small Cell Lung Cancer March 14, 2005.
Pollack A. Genentech Drug Shows Effect on a 2nd Cancer. The New York Times, March 15, 2005.
Blue Cross to Pay for Life-Prolonging Cancer Drug. Newsday.com, March 16,2005.(Daniel Dornbuscha, Carmen)