© Oxford University Press 2006.
NEWS |
Negative Data From Lung Cancer Trial May Change Practice Guidelines, Study Designs
Revised practice guidelines for early non–small-cell lung cancer (NSCLC) are due this fall from the National Comprehensive Cancer Network (NCCN), triggered by new uncertainties surrounding one of the big questions in early-stage NSCLC: Does adjuvant therapy benefit patients at stage IB?
Oncologists thought they had the answer to that question 2 years ago when preliminary results from a randomized trial showed a definite survival benefit in IB—so strong, in fact, that the trial was stopped early, and adjuvant therapy became standard of care.
But the positive findings were short-lived. As data continued to accumulate, the benefit shrank, and by last spring, the results of the trial, CALGB-9633, no longer showed a statistically significant overall survival benefit for adjuvant therapy in stage IB.
"While we do see trends that continue to favor the chemotherapy arm, there is no question that the difference is less," said principal investigator Gary Strauss, M.D., associate professor of medicine at Brown University in Providence, R.I., who presented the updated survival data at this year's meeting of the American Society of Clinical Oncology. "The results do not mandate adjuvant chemotherapy as standard of care in stage IB non–small-cell lung cancer."
Clinicians and researchers have been struggling with how best to cope with the new uncertainty surrounding stage IB, but one point is clear: The unexpected turnaround of the data is having a substantial impact on both practice and the next generation of trials.
Changing Clinical Practice
In the wake of the trial results, the NCCN lung cancer committee met in August, and the new recommendations for clinicians will be official by the end of the year.
Following the early closing of CALGB-9633 in 2004, the NCCN had strongly endorsed adjuvant therapy for stage IB (as well as II and IIIA). That will change, said Duke University's Thomas D'Amico, M.D., a member of the NCCN's lung cancer committee who spoke at a recent meeting on the guidelines. However, the NCCN guidelines will not rule out adjuvant therapy for stage IB, he said.
|
"There is a gray area here for clinical decision making," D'Amico said. Variables indicating increased risk, such as the size of the tumor or lymph system invasion without lymph node involvement, can be taken into account, he said. Other signs of increased risk, such as poorly differentiated tumors and those with large-cell histology, could play a role in a physician's decision.
The CALGB-9633 researchers agree that adjuvant therapy should continue to be considered for some patients. "We still believe that the results of the study support consideration of adjuvant paclitaxel and carboplatin in stage IB NSCLC, particularly among those with tumors 4 cm and larger in diameter," said Strauss in his ASCO presentation. An unplanned subset analysis of the findings showed that patients with tumors 4 cm and larger had a statistically significant survival benefit.
But not all experts agree. Joan Schiller, M.D., of the University of Texas Southwestern Medical Center in Dallas and the ASCO discussant of the new CALGB-9633 data, said she disagreed with that conclusion and argued that adjuvant therapy in stage IB should now be reserved for clinical trials.
"Until we know why the study was negative, be it because of a small treatment effect, wrong stage, or wrong drugs ... I don't think we can routinely recommend carbotaxol off of a clinical study," she said.
Which platinum drug to use is another issue that has resurfaced in the wake of the new CALGB data. That study was the only one of four recent adjuvant studies to use carboplatin instead of cisplatin, so its updated results remove the only evidence in favor of carboplatin. What's more, in a meta-analysis also presented at ASCO this year, the trend favored cisplatin for overall survival when looking at cisplatin versus carboplatin in stage III and IV disease. In patients with nonsquamous tumors and those who had third-generation drugs, the survival difference was statistically significant.
"It was a minimal difference but enough of one to make a difference in a disease where you can aim for a cure," said Heather Wakelee, M.D., assistant professor at the Stanford University School of Medicine in California. "In early-stage disease, you want to do all you can ... a 1% or 2% difference might make a difference when cure is the goal."
The NCCN guidelines will recommend cisplatin for patients who can tolerate it, D'Amico said.
Designing the Next Trials
The new CALGB-9633 data have also changed future research plans. "There's been a lot of discussion and reassessment over the last few months," said the National Cancer Institute's Claudio Dansky Ullmann, M.D., who coordinates the cooperative groups' lung cancer trials.
"Although other positive adjuvant trials have shown an overall benefit in early-stage NSCLC, it is clear that not all subgroups of patients will benefit and that current therapies are not good enough to be applied across the board," he said. "Research questions in future trials should be more focused on addressing these issues."
And in fact there have been key changes in the plans for two proposed NCI intergroup trials. One of the trials, ECOG-1505, will test cisplatin-based chemotherapy with or without bevacizumab. Originally designed to include patients with stage IB–IIIA, it has been modified to include only some stage IB patients—those with tumors 4 cm or larger—on the basis of a subset analysis in CALGB-9633.
Carboplatin also will not be an option in the redesigned ECOG-1505. Investigators may choose among three cisplatin doublets—cisplatin–vinorelbine, cisplatin–gemcitabine, or cisplatin–docetaxel—before randomization to bevacizumab or no bevacizumab. Plans call for accruing 1,500 patients over 4 years with 3.5 years of follow-up.
The decision to include the subset of IB patients was hotly debated, said Wakelee, who is the principal investigator. "Basically, it's making a new substage," she said, and many people are wary of that. "Some say it should not be included at all; some say all IB should be included."
Including all stage IB patients did not make sense because there was no longer any reliable evidence that stage IB patients benefited from chemotherapy. But there were reasons for keeping the question open. These included the positive (though non–statistically significant) trends in survival seen both in CALGB-9633 and in another meta-analysis presented at ASCO this year. Also, positive results in Japan from large trials with uracil–tegafur in stage IA and IB have kept alive hope that adjuvant therapy could play a role in the earliest stages.
"We didn't want to cut out IB completely, so we compromised by including only IBs that were 4 cm or greater," Wakelee said.
The other adjuvant intergroup trial, which is still in the planning stage, will test the use of a gene expression profile to determine whether early-stage NSCLC patients will benefit from adjuvant therapy. Originally proposed just for IA patients, it is now being modified to include IB patients as a result of the new CALGB data, said David Harpole, M.D., professor of surgery at Duke University Medical Center in Durham, N.C., and the principal investigator on the proposed trial. It may be limited to IB patients with smaller tumors, he said.
Using a microarray-based profile developed at Duke, called the metagene model, the trial will divide participants into two groups after surgery: those at low risk for recurrence and those at higher risk. The low-risk group will be observed, and the high-risk group will be randomized to either observation or cisplatin-based adjuvant therapy. They hope to recruit 1,000 patients over about 4 years with 2 years of follow-up, Harpole said.
A third adjuvant study is OSI Pharmaceutical's Radiant Trial, which will randomly assign patients to either erlotinib or a placebo after surgery with or without chemotherapy. This trial, which opened on Sept. 21, will include 945 patients at stages IB–IIIA. Only patients who are positive for erlotinib's target, the epidermal growth factor receptor, will be eligible.
Impact of Early Closing
Closing trials early sometimes generates controversy, with critics citing the loss of data, especially long-term toxicity data. But there has been "surprisingly little" debate over the wisdom of early closure in this case, said Paul Bunn, M.D., director of the University of Colorado Comprehensive Cancer Center in Denver, and a Southwest Oncology Group leader.
|
One reason may be that the trial was small and, as it turned out, underpowered. Its original recruitment goal had been reduced from 500 to 384 patients because of slow accrual, and the early closing reduced the sample further, from 384 to 344 patients. That 10% reduction probably did not reduce the ability of the trial to detect a survival difference, Strauss said, "because, based on the magnitude of the effect we've seen, the study was too small." If the trial had not been halted, it still would have had only 40% power to demonstrate a significant survival difference, he said; to have 80% power, the aim, of this and other trials, 1,000 patients would have been needed.
However, the loss of statistical significance in CALGB-9633 has raised awareness of the rules for early closure, say CALGB researchers.
"We are being more conservative in our group," said Harpole. "We're making sure that when we make stopping rules, statistically they are conservative." And, he noted, the proposed metagene trial will have 90% power to detect statistically significant overall survival differences in the two arms. In the past, many trials, including CALGB-9633 and the other large lung trials that established adjuvant therapy as the standard of care for stages II–IIIA, were designed with 80% power.
Stephen George, Ph.D., chief of statistics for both CALGB and the Duke Comprehensive Cancer Center, said there is more awareness of the standards for early closure. "I think what's happening is a greater awareness that these calculations [of preset closing points] are based on assumptions," he said. One such assumption, for example, is that hazard ratios will remain the same over time—an assumption proved wrong in CALGB-9633.
"There's not a sea change going on," he said, "but it's fair to say there is greater awareness."
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K. Miyoshi, S. Moriyama, T. Kunitomo, and S. Nawa Prognostic impact of intratumoral vessel invasion in completely resected pathologic stage I non-small cell lung cancer. J. Thorac. Cardiovasc. Surg., February 1, 2009; 137(2): 429 - 434. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Ramsey and J. Scoggins Commentary: Practicing on the Tip of an Information Iceberg? Evidence of Underpublication of Registered Clinical Trials in Oncology Oncologist, September 1, 2008; 13(9): 925 - 929. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||