Retrospective Statistical Analysis Corroborates Early Results Of Randomized Breast Cancer Trials
ATLANTA, G.A. -- Using data on almost 1,300 patients with advanced breast cancer, researchers from Duke University and other institutions have found no benefit from high dose chemotherapy followed by bone marrow transplantation.
They prepared their findings for presentation at the annual meeting of the American Society of Clinical Oncology (ASCO).
The researchers' conclusions are supported by early results of two metastatic breast cancer trials also being presented at ASCO. Unlike these ongoing randomized trials, however, the Duke conclusions come from analysis of metastatic breast cancer patients treated in the 1980s and early 1990s. Patients with metastatic cancer are those in whom cancer cells have moved from the original malignancy to other parts of the body.
According to the American Cancer Society, the five-year relative survival rate for patients with early stage breast cancer is 97%, but the rate for patients with distant metastases is just 22%, similar to the survival of patients in Berry's study.
Half of the analyzed patients had been included in one of four clinical trials and were all treated with standard doses of chemotherapy. The other half were treated with high-dose chemotherapy and bone marrow transplant outside of clinical trials and were included in a patient registry, a compilation of treatment information and patient characteristics.
"We found no benefit for metastatic breast cancer patients treated with high-dose chemotherapy and bone marrow transplant, even adjusting for patient characteristics such as estrogen receptor status, age, and metastatic sites," said lead author Donald Berry, professor in the Institute of Statistics and Decision Sciences at Duke. "Moreover, we could find no subset of metastatic patients for whom transplant was beneficial."
Berry emphasized that the analysis shows it is possible to determine two treatments' relative abilities without a randomized clinical trial. He does not believe that randomized trials can be replaced, but careful statistical analysis of patient registries, databases of patient characteristics and treatment information, can provide information to direct further inquiry, he said.
"There are circumstances, in which 'non-randomized' comparisons can be informative, and they can be even more informative when the sample sizes are large," Berry said. "These circumstances are present when assessing possible benefits of bone marrow transplant in metastatic breast cancer."
Berry's statistical analysis compared metastatic breast cancer patients who participated in four clinical trials conducted by the Cancer and Leukemia Group B (CALGB) with that of metastatic breast cancer patients included in the Autologous Blood and Marrow Transplant Registry (ABMTR). The 663 CALGB patients included had been treated with various conventional chemotherapy regimens from 1980 to 1992, while the 638 ABMTR patients were treated with various regimens of high-dose chemotherapy with bone marrow or stem cell transplants from 1989 to 1995.
The CALGB trials were initially conducted to compare various chemotherapy regimens, but no survival differences were seen for patients with metastatic breast cancer, within each trial or across the trials. Berry looked for treatment-related differences within ABMTR, but found none. Thus the two separate patient groups were compared to see if the high dose chemotherapy used in ABMTR provided any survival benefit for metastatic breast cancer patients.
As the standard for evaluating new treatments, randomized clinical trials usually pit a new cancer treatment against the standard treatment to see if one is more effective or has fewer or more-tolerable side effects. Patients who meet requirements and agree to enter the trials are randomly assigned into the group getting the standard therapy or the group getting the experimental therapy.
Hundreds of patients are usually needed for researchers to detect any differences in survival, response to therapy or other measures. For some rare cancers, there frequently aren't enough patients to detect small differences.
"Randomized trials are more credible, but there are a number of obstacles to overcome," said Berry, who is a statistician in the CALGB. "It is difficult to enroll sufficient numbers of patients, the trials are difficult to run, they are expensive, and they take many years to yield results."
While the right circumstances exist to evaluate high dose chemotherapy and bone marrow transplant in non-randomized metastatic breast cancer patients, that is not the case for patients with earlier stage disease because of how these patients are selected for transplant, Berry said. Therefore only patients with proven metastatic disease and who were treated accordingly were included in the group's analysis.