Patterns of Care and Predictors of Adjuvant Radiation Therapy in Phyllodes Tumor of the Breast

Abstract

Wide excision without axillary staging remains the recommended treatment for phyllodes tumors of the breast, a rare malignancy that accounts for less than a percent of all breast cancers. There are no prospective data to support the role for adjuvant radiotherapy; however, National Comprehensive Cancer Network (NCCN) guidelines suggest its use in situations where the morbidity of local recurrence would be significant. The purpose of our study is to evaluate the patterns of care, utilization and predictors of adjuvant radiation therapy for phyllodes tumors. We identified 3,108 patients in the National Cancer Database who were diagnosed with phyllodes tumor of the breast from 2004 through 2015 and treated with surgery. Patients treated with neoadjuvant therapy were excluded. Temporal trends were evaluated using the Cochran-Armitage test. Multivariable logistic regression was utilized to examine predictors of adjuvant RT use. Predictors of overall survival were analyzed for the 2,823 patients for whom survival data are available using multivariable Cox proportional hazards regression. Overall, 53.4% of patients received breast-conserving surgery and 35.9% received mastectomy. The type of surgery was unknown in 10.7% of the cohort. Lymph node (LN) sampling or dissection was performed in 25.9% of patients. Surgical margin status was negative in 90.6% of patients, positive in 6.7%, and unknown in 2.8%. Overall, 23.2% of patients received adjuvant RT. Its use has increased from 15.1% in 2004 to 32.3% in 2015 (P <0.0001). The median dose was 60.0 Gy (IQR 50.4 to 60.4). On multivariable analysis, adjuvant RT use was more common for patients with tumors ≥5cm (adjusted odds ratio [AOR] 1.34, P =0.004); with poorly-differentiated (AOR 3.14, P <0.001) or anaplastic tumors (AOR 3.17, P <0.001); treated with nodal sampling or dissection (AOR 1.75, P <0.001); and treated in an academic/research center (AOR 1.37, P =0.006). Its use was less common for patients age >70 years (AOR 0.54, P =0.003); and with Charlson-Deyo comorbidity score ≥1 (AOR 0.73, P =0.046). Positive surgical margins were not a predictor of adjuvant RT use (AOR 1.01, P =0.957). Adjuvant RT was not independently associated with improvement in overall survival (adjusted hazard ratio 1.21, 95% confidence interval 0.97-1.53, P =0.097). From 2004-2015, the use of adjuvant radiation in phyllodes tumor has doubled and appears to be utilized in younger patients with fewer comorbidities, less favorable pathologic features, and who receive treatment at academic centers. Margin status was not a predictor for adjuvant radiation. Despite NCCN guidelines recommending against it, one quarter of patients underwent LN sampling. Adjuvant radiotherapy does not appear to have a benefit in terms of overall survival.