Progression-Free Survival Called Unsteady Endpoint for Cancer Trials

NEW YORK, March 21 -- Progression-free survival may be a blurry yardstick for the efficacy of cancer drugs because the definition of when progression begins is subjective, cautioned investigators here.

When overall survival is a trial's endpoint, the date of death of each patient is a fixed event that can be accurately measured, but when progression-free survival is used, the outcome depends on when and how often patients are monitored for signs of disease, wrote Katherine S. Panegeas, Dr.P.H., and colleagues, of the Memorial Sloan-Kettering Cancer Center.

"Progression-free survival is often seen as a desirable outcome measure because it is available earlier than overall survival and can thereby shorten drug development time, resulting in more rapid availability of efficacious therapies," the investigators wrote in the March 21 issue of the Journal of the National Cancer Institute.

"Furthermore, unlike overall survival, progression-free survival is not influenced by second-line treatment choices," they said. "But in contrast to overall survival, disease progression is subject to measurement error; it is also influenced by the timing of scheduled reassessments."

The authors said that if progression-free survival is used as study endpoint for cancer drug trials, assessment intervals should be standardized by the type of cancer studied and cancer stage.

A determination that cancer has progressed is usually on the basis of radiographic findings from studies conducted at fixed intervals, such as every eight, 12, or 24 weeks. But the actual point of disease progression is rarely so obliging as to occur just at the time of a scheduled follow-up, the investigators noted.

The authors reviewed all phase II and III clinical trials in breast cancer published in 2005 to determine how often progression-free survival was reported as a trial endpoint, and to determine whether tumor assessment across trials was consistent.

They identified 43 studies reporting original trials results, 10 of which listed progression-free survival or time to progression as the primary endpoint, 19 of which listed these outcomes as secondary endpoints, and 14 of which did not specify the use of progression-free survival.

The trials generally conducted assessments every two or three treatment cycles, but the treatment cycles themselves varied from every six weeks in some studies to every three months in others, and in four of the studies no assessment interval was specified.

"Although these differences in cycle length may seem modest, there is a close relationship between cycle length and progression date that can lead to bias," the authors wrote.

They cited the example of a recent phase II study evaluating erlotinib (Tarceva) and bevacizumab (Avastin) in 37 women with metastatic breast cancer. The trialists in the study defined the time to progression as the date of the CT scan in which the progression was first detected, at a point that statisticians refer to as the upper limit of the true progression interval. On a graph using the Kaplan-Meier method, the clinical events appeared to occur at nine-week intervals -- the specified interval for CT scans in the study.

Looked at this way, using the upper limit of the progression interval, the estimated median progression-free survival was 10 weeks (95% confidence interval, eight to 17 weeks).

"However, another way to analyze progression would be to use the other end of the progression interval so that progression would be defined as occurring on the date of the scan before the one at which progression was identified, namely the 'lower limit' of the progression interval," the authors wrote. "Using the lower limit, the estimated median progression-free survival from Kaplan-Meier analysis is 0 weeks (95% CI, 0 to 9 weeks). This estimate of median progression-free survival is markedly different from the 10-week estimate of progression-free survival obtained by the use of the upper limit and is likely to be an underestimate of the true median progression-free survival."

The trials designers could have as easily chosen the five-week midpoint between scan as the progression date, which would yield an estimated median progression-free survival of five weeks (95% CI, 4 to 13 weeks).

"There is no way to evaluate which of the three median progression-free survival estimates given by these three approaches most accurately represents the true underlying median. In effect, we are ignoring the interval censoring by arbitrarily basing the progression date on the interval reassessment date," the investigators pointed out. "Reliance on the final surveillance date to determine progression will necessarily result in inflated estimates of median progression-free survival," they said.

They called on trialists to consider reporting the lower limit, midpoint, or upper limit when using progression-free survival as a study endpoint. They also recommended standardizing study assessment intervals, "e.g., every eight weeks for chemotherapy trials in metastatic pancreatic or metastatic breast cancer, every 10 weeks for metastatic colon cancer trials, and every 12 weeks for metastatic breast cancer hormonal therapy trials."

They acknowledged that "to accommodate the realities of the clinical setting, there needs to be a reasonable margin for scheduling around these specified time points. When designing a phase II study for a disease with short median progression-free survival, it may be better to use a binary endpoint, such as the three-month progression-free survival rate, and to use a traditional phase II design, as is common for tumor response endpoints. All patients would then be assessed by this fixed time, allowing for greater uniformity of results across trials."