Abstract Purpose: Phyllodes tumor is a relatively rare disease of the breast, and the risk factors for local versus distant failure are unclear. The goal of this study was to identify factors influencing local, distant, and cause-specific survival outcomes in patients with phyllodes tumors, and to assess the impact of local radiotherapy (RT). Methods: We retrospectively reviewed the records of 387 patients recorded in the MD Anderson tumor registry as having "cystosarcoma phyllodes" or "phyllodes". After excluding patients with recurrent disease at presentation to the institution, concurrent invasive ductal carcinoma, absence of confirmation of phyllodes tumor, or inadequate follow up after surgery, 229 patients diagnosed from 1964-2011 were evaluable. The median tumor size was 4 cm (range, 0.7-28 cm); histology was benign in 29%, indeterminate in 7%, and malignant in 41% (23% unknown). Stromal overgrowth was present in 18% (n=40). Local therapy consisted of breast-conserving surgery (n=184) or mastectomy (n=43); 15% (n=34) also received local RT. Chemotherapy was administered in 9% (n=21). We used Kaplan-Meier analyses and Cox proportional hazards models to estimate the associations between patient/tumor characteristics and treatment on local control (LC), distant metastasis-free survival (DMFS), and cause specific survival (CSS). Results: At a median follow-up of 76 months (range, 1-485 months), the actuarial 5-yr LC for the entire cohort was 75%; 5-yr DMFS was 78%; and 5-yr CSS was 86%. Factors influencing LC included receipt of RT (5-yr LC 97% versus 72%, p=.005) and age over 50 (85% versus 72%, p=.044). Neither malignant histology nor stromal overgrowth increased the risk of local failure. Factors influencing DMFS included malignant histology (5-yr DMFS 65% versus 91%, p=.001) and stromal overgrowth (50% versus 84%, p<.001). Factors influencing CSS included malignant histology (5-yr CSS 77% versus 100%, p=.002) and stromal overgrowth (58% versus 92%, p<.001). Patients with stromal overgrowth or malignant histology were much more likely to receive RT (p<.001 for both factors). RT improved 5-yr LC for both breast conserved and mastectomy patients, but did not improve DMFS or CSS. In multivariable models, the following associations were noted (HR=hazard ratio; CI= 95% confidence interval): LC – RT (HR 0.13, CI .03-.55); DMFS – malignant histology (HR 2.99, CI 1.41-6.34); CSS – malignant histology (HR 4.18, CI 1.43-12.24). Conclusions: Adjuvant RT after surgery improves local control in phyllodes tumor; however, it does not improve DMFS or CSS. Stromal overgrowth and malignant histology are associated with worse DMFS and CSS, but do not impact local control. Therefore malignant histology and stromal overgrowth should not be deciding factors in the use of local RT, as escalated local therapy in patients with these tumor characteristics does not improve DMFS or CSS. Other strategies should be considered for patients with phyllodes tumor who are at high risk of distant failure, including systemic therapy. Citation Format: Bobbi A Porche, Pamela K Allen, Simona Shaitelman, B Ashleigh Guadagnolo, Wendy A Woodward, Constance Albarracin, Abenaa M Brewster, Kelly K Hunt, Welela Tereffe. Long term outcomes in patients with phyllodes tumor of the breast: The UT MD Anderson experience [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P1-15-03.