279 - Utilization of a Lumpectomy Bed Boost after Whole Breast Radiotherapy Does Not Improve Ipsilateral Breast Tumor Recurrence Risk in Patients with Ductal Carcinoma In-Situ Treated on Three NSABP Trials
Presenter(s)
A. J. Khan1, X. Gu2, G. Tang3, G. Yu2, R. S. Cecchini2, T. B. Julian4, E. P. Mamounas5, J. R. White6, and F. A. Vicini7; 1Memorial Sloan Kettering Cancer Center, New York, NY, 2NRG Oncology Statistical and Data Management Center; Department of Biostatistics and Health Data Science, University of Pittsburgh, Pittsburgh, PA, 3NRG Oncology Statistics and Data Management Center, and University of Pittsburgh School of Public Health, Department of Biostatistics and Health Data Science, Pittsburgh, PA, 4Allegheny Health Network Cancer Institute, Pittsburgh, PA, 5AdventHealth Cancer Institute, Orlando, FL, 6Department of Radiation Oncology, University of Kansas Medical Center, Kansas City, KS, 7Michigan Healthcare Professionals, Farmington Hills, MI
Purpose/Objective(s):
The seminal studies of adjuvant therapies in patients with ductal carcinoma in-situ (DCIS) did not routinely mandate or prohibit use of a lumpectomy bed boost after whole breast radiotherapy (RT). Delivery of this additional RT improves in-breast recurrence risk in invasive breast cancer patients. A recent trial demonstrated lower recurrence risk with boost in DCIS patients with adverse risk features. We studied the effect of boost utilization in women with DCIS treated on NRG trials.Materials/Methods:
We examined data from three NSABP (now NRG) trials (B-24, B-35, and B-43), which exclusively studied women with DCIS. The primary endpoint of this cross-protocol study was time to ipsilateral breast tumor recurrence (IBTR, invasive or in situ), censored by regional/distant recurrence, ipsilateral mastectomy, and either death or date of last mammogram, whichever came first. Stratified log-rank test was used to test whether boost reduced patients’ risk of IBTR, vs patients without boost. Stratification factors included study and potential study-specific confounders. Study-specific propensity score models were developed to predict boost status and were subsequently used in the inverse probability-weighted Cox model to assess effect of boost on risk of IBTR within each study. Then those results were summarized using random-effects meta-analysis.Results:
A total of 6,922 randomized patients with follow-up data were included: 1,804 from B-24, 3,104 from B-35, and 2,014 from B-43. Use of boost was less frequent in B-24 (38.2%) vs B-35 (82.5%) and B-43 (83.4%). IBTR was more prevalent in B-24 (14.9%) than B-35 (2.3%) or B-43 (6.5%). Prevalence of IBTR was also balanced between the no-boost and boost groups: 15.1% vs 14.6% in B-24, 1.4% vs 2.5% in B-35, and 6.8% vs 6.4% in B-43. There was no evidence of an effect of boost on the risk of IBTR in the stratified log-rank test (p=0.9) or risk of recurrence (p=0.89). Meta-analysis did not reveal any effect of boost on the risk of IBTR (p=0.31).Conclusion:
In this retrospective cohort study re-examining the correlation of boost utilization with IBTR risk, we did not find an association of boost with recurrence. We demonstrate a decreasing rate of IBTR over time and low overall rates of recurrence in boost and non-boost patients. This study represents the largest examination to date of the effect of boost on IBTR risk in DCIS. Important limitations include inconsistent use of boost within and across studies, heterogeneity of patients/treatment, non-standard definitions of boost volume/doses, and lack of pre-specified RT QA. Although these data do not definitely exclude benefit of boost in good-risk DCIS patients, the likelihood of a demonstrable benefit in patients at the low end of the risk spectrum appears to be low.