In Early Breast Cancer, Outcomes Affected by Tumor Biology and Choice of Local Treatment

September 5, 2014

Breast conserving therapy (BCT) improves overall survival compared with other local treatments in patients with stage I breast cancer, according to an examination of the National Cancer Database (NCDB). The benefit associated with BCT appears to be limited to women with hormone receptor (HR)-positive disease, said Catherine Parker, MD, at the 2014 ASCO Breast Cancer Symposium.

Mastectomy and BCT (lumpectomy and whole breast radiation) were established as appropriate treatments for early stage breast cancer in the 1980s, with equivalent survival outcomes. At the time these clinical studies were conducted, however, little was understood about breast cancer biology. In the past 10 to 15 years, genomic studies have provided new insights into the biology of breast cancer. “From these studies, breast cancer subtypes have emerged, which are important predictors of distant metastasis, locoregional recurrences, and radiobiology,” said Parker, an oncology fellow at the University of Texas MD Anderson Cancer Center, Houston, where the study was performed. She’s now at the University of Alabama, Birmingham.

Specifically, HR-positive cancers appear to be more sensitive to radiation compared with HR-negative tumors, she said.

In 2008, a re-analysis of data from the Danish Breast Cancer Cooperative Group (J Clin Oncol 2008;26:1419-1426) found a significant reduction in the probability of locoregional recurrence and an improvement in overall survival after post-mastectomy radiation in women with HR receptor-positive breast cancer. A moderate benefit to radiation on locoregional recurrence in patients with triple-negative disease did not translate into a survival benefit, which suggests the potential for differential effects locally and systemically to radiation.

Given the dramatic progress made in the understanding of breast cancer, Parker’s group hypothesized that the choice of local treatment may have animpact on survival outcomes if reviewed in the context of tumor biology.

The NCDB is a nationwide registry of the American College of Surgeons, the American Cancer Society, and the Commission on Cancer. It captures about 70% of newly diagnosed cases of cancer nationwide. From the registry, 16,646 women in 2004-2005 with stage I breast cancer who underwent mastectomy excluding post-mastectomy radiation, BCT, or breast conserving surgery (BCS) without radiation were identified. These years were chosen to allow adequate follow-up to assess survival outcome.

Tumors were classified as HR-positive or HR-negative. Patients were matched using propensity score for a broad range of covariates, including age at diagnosis, ethnicity, Charlson/Deyo comorbidity score (a variation of the Charlson comorbidity index), tumor grade, HR status, receipt of hormone therapy and/or chemotherapy, receipt of endocrine therapy, and the facility where cancer treatment was delivered.

The mean age at diagnosis of the 16,646 women who met the inclusion criteria was 59 years in those who underwent BCT, 62 years in those who underwent mastectomy, and 66 years in those treated with BCS. The median follow-up was 83.5 months.

Two-thirds of the cohort underwent BCT compared with 22% who underwent mastectomy and 11% who were treated with BCS. Nearly 90% of the cohort was white.

One hundred percent of the cohort had negative surgical margins, 85% of the women had HR-positive disease, 28% received chemotherapy, and 63% received endocrine therapy.

In the matched cohort of 1,706 patients in each arm, patients undergoing BCT had improved overall survival (OS) compared with the other treatment groups (P<.0001). The estimated 5-year OS was 93% for BCT, 88% for mastectomy, and 86% for BCS. A similar pattern of improved OS with BCT was observed among HR-positive patients within the matched cohort (P<.0001), who had an estimated 5-year OS of 94% with BCT, 89% with mastectomy, and 87% with BCS. The differences across treatment groups among the matched cohort of HR-negative patients were not as pronounced, with estimated 5-year OS of 90% with BCT, 87% with mastectomy, and 81% with BCS (P=.032 for BCT vs. the other groups).

After multivariate adjustment, BCT retained its OS advantage compared with mastectomy (hazard ratio [HR]: 0.76; P<.01) and BCS (HR: 0.63; P<.01). Overall survival was more favorable for women with HR-positive tumors than HR-negative tumors (HR: 0.80; P=.03).

In the matched cohort, after multivariate adjustment, the OS benefit of BCT over mastectomy (HR: 0.73; P<.01) and BCS (HR: 0.63; P<.01) remained in the HR-positive subset. In the HR-negative subset of patients, multivariate analysis revealed no significant difference in OS between BCT and mastectomy although BCT remained associated with better OS compared with BCS.

“Differences in survival across HR subtypes may reflect differences in radiation sensitivity,” said Parker.

<<<