Variation in the use of boost irradiation in breast conserving therapy in the Netherlands

Background: A boost dose is applied after breast conserving surgery (BCS) and whole breast irradiation to further reduce the risk of local recurrences in breast cancer. In the NABON Breast Cancer Audit (NBCA) variation in the use of the boost was seen. Identification of factors explaining this variation can be useful to reduce possible undesirable variance in clinical practice. Aims of the current study were to determine the variation of the use of boost irradiation over time and between radiation oncology departments. We also investigated which patient and tumour factors were associated with the use of a boost. Material and Methods: From the NBCA cohort, all patients with primary DCIS or invasive breast cancer without distant metastatic disease diagnosed between January 1, 2011 and December 31, 2014 were selected. Funnel plots were used to evaluate the variation in the use of a boost between departments and over time. Logistic regression was performed to determine factors influencing the variation. Analyses were performed in STATA (version 13.1 2013, Texas). Results: During the study period, 33,902 female patients were treated with radiation for DCIS or invasive breast cancer after BCS. In total 51% of the patients received a boost (45% DCIS, 54% invasive, Table 1). Variation between the 23 departments of radiation oncology was seen for both DCIS and invasive tumours. For DCIS patients, the use of the boost slightly increases over time while for invasive tumours a decrease was seen. Multivariable logistic regression (Table 2) demonstrates that lower age (<50 years), larger tumours (>2 cm diameter), grade 3 and irradical surgery increased boost application for both DCIS and invasive breast cancers. Positive lymph nodes significantly influences boost prescription for invasive breast cancers. Variation could not completely be explained by patient and tumour characteristics. Table 1. DCIS (n = 4,568) Invasive (n = 29,334) % Boost 45% 54% Variation between institutions (n = 23) 5−92% 21−80% % boost in 2011 (range) 45% (5–100%) 61% (14–100%) % boost in 2014 (range) 46% (5−86%) 44% (25−59%) Table 2. Multivariable logistic regression boost vs no boost DCIS (n = 4,568) Invasive (n = 29,334) OR 95%CI OR 95%CI Age <50 vs >50 years 1.8 1.6−2.0 8.0 7.4−8.7 Larger (>2 cm) vs smaller (