SAN ANTONIO – The 10-year incidence of leukemia after adjuvant chemotherapy for breast cancer appears to hover around 0.5%, twice as high as previously reported.
A review of more than 20,000 patient records included in the National Comprehensive Cancer Network (NCCN) database found 51 cases of leukemia that developed within 10 years of treatment. Women who had only chemotherapy were at the greatest risk – almost six times more likely to develop the disease than the surgery-only control group, Dr. Antonio Wolff said at that annual San Antonio Breast Cancer Symposium.
Dr. Antonio C. Wolff
"We know that the observed survival benefit with adjuvant therapy does take into account the potential mortality associated with leukemia," said Dr. Wolff of Johns Hopkins Sidney Kimmel Comprehensive Cancer Center, Baltimore. "But often we are in the situation where we make the decision to give chemotherapy ‘just because.’ We need to be very careful here, because some of those patients will potentially derive none of the benefits of chemo, and be at risk of its toxicities."
Dr. Wolff and colleagues examined the 1997-2008 data from eight facilities in the NCCN database – a total of 20,533 women with a first diagnosis of breast cancer. The median follow-up was 5 years, but some data were available for up to 15 years after treatment.
About half of the cohort had stage I disease; in most, the cancer was invasive ductal. Slightly more than half had hormone receptor–positive/HER2 negative disease.
The only baseline characteristic significantly different between the groups was age. Women who developed leukemia were significantly older at the time of their cancer treatment (60 vs. 54 years; P = .02); 78% of those with leukemia were older than 50 years.
Of the 51 cases, 44 were acute myeloid leukemia (AML), with a median onset time of 3.5 years after treatment. One-third of these occurred in women with a family history of breast or ovarian cancer. The median time to onset in the AML cases was 2 years.
None of the tumor characteristics were significantly associated with the development of leukemia. There were no significant differences among those who had breast-conserving surgery, mastectomy, or no surgery; or between those who had no radiation therapy, radiation without surgery, radiation after breast-conserving surgery, or radiation after mastectomy, though larger numbers of patients are needed for some of these smaller subset analyses.
The overall rate of leukemia per 1,000 patient/years was 0.46; in the surgery-only group, the rate was 0.16 per 1,000 patient-years. "Of interest is that the rates in each of the treatment groups were similar to the overall rate," – 0.43 in the radiation-only group, 0.52 in the chemotherapy-only group, and 0.54 in the combination therapy group.
At 5 years, the cumulative incidence of leukemia was 0.05% in the surgery-only group; 0.19% in the radiation-only group; 0.30% in the chemotherapy-only group; and 0.32% in the combination therapy group.
But the onset accelerated over the study period, Dr. Wolff noted, with the bulk of cases developing from years 6 to 10. .By the end of the 10-year period, the incidence was 0.2% in the surgery-only group; 0.44% in the radiation-only group; 0.52% in the chemotherapy group; and 0.51% in the combination group.
In a risk analysis using surgery as the control, the overall hazard ratio of leukemia was 2.7 in the radiation-only group; 5.68 in the chemotherapy only group; and 5.64 in the combination group.
Radiation appeared to confer no significant additional risk when it was combined with chemotherapy, Dr. Wolff added.
The findings are strikingly different than previously identified. The largest study of the association was published in 2003, when researchers from the National Surgical Adjuvant Breast and Bowel Project Operations Center reviewed six trials that examined rates of acute myeloid leukemia and myelodysplastic syndrome after doxorubicin and cyclophosphamide therapy (Clin. Breast Cancer 2003;4:273-9).
The subgroup that received anthracycline/cyclophosphamide in that review had a cumulative 8-year incidence of 0.24%. In Dr. Wolff’s chemotherapy subgroup, the cumulative 8-year incidence was 0.52%.
"It’s very important to keep in mind that the known leukemia latency period for anthracycline is 1-3 years, but that of alkylating drugs like cyclophosphamide is 4-6 years and there are case reports of it developing after 10 years," he said. "We need to be very careful when we talk about survival at 5 years vs. 10 years in exposed patients, because they could be at risk for a decade and, potentially, even longer."