A likely reason for that finding is that mammography can detect tumors that are both slower growing and less biologically lethal than those found symptomatically, say the researchers, who published their findings in the Aug. 17 issue of the Journal of the National Cancer Institute.
The study is important because the survival benefit seen in this analysis is much greater than one would expect for screen-detected breast cancer, says the study's lead author Donald Berry, Ph.D., chair of the Department of Biostatistics and Applied Mathematics. Berry is well-known for his work in designing breast cancer clinical trials sponsored by the National Cancer Institute and for his research in evaluating the effects of screening mammography.
"We know that screening picks up many tumors before they can be detected in other ways and women may benefit from early treatment, but the advantage we found is much larger than what would be expected from the so-called stage shift that is associated with screening mammography," Berry says.
Based on the results, Berry suggests that method of detection should be considered when a treatment plan for newly diagnosed breast cancer is being devised, and that this information also should be collected by researchers conducting clinical trials of experimental therapeutic strategies. "That information may be just as important as other variables, such as the number of positive lymph nodes. If you don't account for method of detection, the results may not be as accurate as they would be otherwise," Berry says.
"The important message here for clinicians and patients is that breast cancer detected through mammography has a substantially better survival prognosis," he says. "Of two women who have the same age, size of tumors, and similar stage of cancer and spread to lymph nodes, the one whose cancer was detected with mammography has a reason to be happier than the woman whose cancer was detected symptomatically," Berry says.
While that sounds like good news for some patients, Berry says the conclusion should not be over interpreted.
"The paradox is that this result does not mean screening is beneficial," he says. "Without screening, some of the women would not have been diagnosed with breast cancer at all, and in that group, some of them could have avoided surgery and treatment without detriment. The rub is that we don't know which ones they are."
This issue has long plagued screening mammography, especially when the detected tumors are very small and have not spread. But this study appears to add a new element to the debate, Berry says. "Our conclusions apply generally, and are as important in node-negative breast cancer as they are in node-positive disease," he says.
In this study, researchers examined data from three large randomized breast cancer screening trials - the Health Insurance Plan (HIP) of New York, which assigned about 62,000 women to screening or to a control group; and two Canadian National Breast Cancer Screening Studies (NBSS), which included a total of 44,790 women in the screening groups and 44,961 women in the control groups.
They then looked only at women in these studies who were eventually diagnosed with breast cancer, and adjusted for stage and other tumor characteristics as a way to eliminate what is known as "lead-time bias." Lead time is the time between when the tumor was detected by mammography and when the tumor would have been detected in the absence of screening. Lead-time bias occurs because lead time is added to the survival time of women detected by mammography but not to women whose tumors are detected clinically, Berry says. "Lead time is an artifact of screening and not necessarily a benefit of screening," he says.
If lead-time bias was responsible for improved survival, and if it was eliminated from the screening mammography group, then those patients should have the same survival, statistically, as women whose cancer was detected outside of mammography, the researchers say.
But that is not what they found.
Instead they discovered that all things being equal, the method of detection was a statistically significant independent predictor of breast cancer survival. After adjusting for stage of disease, patients whose breast tumors were discovered after a previous negative mammography screen had a 53 percent greater risk of death from the cancer than women with screen-detected cancer. Patients in the control group (where no mammography was used) had a 36 percent increased risk of death compared to screened patients.
"What is new here is that we found an effect that is beyond stage shift," Berry says. "All breast oncologists know that tumors detected by mammography have a better survival than tumors detected otherwise because they are smaller and more likely to be node-negative. Our study shows that these patients are even better off than their clinical characteristics suggest."
The difference likely is due to what is called a "length bias" which occurs because screening detects disproportionately more slowly growing tumors, the researchers say. This bias statistically incorporates biological characteristics of the tumors which have not yet been discovered, or cannot yet be easily diagnosed, Berry says."If we were able to get down to the molecular level to assess just the right factors at play in cancer development and progression, that might tell us how lethal a tumor is, but we can't do that right now."
In the meantime, Berry suggests that information on method of detection should be collected for every clinical trial in order to improve the accuracy of findings, or at least as a way that might account for unexpected results. "I am always surprised that clinical trials of new agents result in better outcomes than expected, and it may just be that screen-detected breast tumors are becoming more common over time," he says.
The study was funded by the National Cancer Institute. Co-authors include first author Yu Shen, Ph.D., Ying Yang, Ph.D., Mark F. Munsell, M.S., all from M. D. Anderson; Lurdes Y. T. Inoue, Ph.D., from the University of Washington, Seattle; and Anthony B. Miller, M.D., from the University of Toronto.