Vaginal brachytherapy (VB) for early stage endometrial cancer (EC) has technologically evolved with the incorporation of 3D imaging into treatment planning. When our institution transitioned from phantom-based 2D template planning to 3D CT-based planning, we implemented selective optimization of standard dwell times for cases with bowel adhered to the vaginal apex. Our hypothesis was that optimizing 3D plans with unfavorable bowel dosimetry would result in no detriment to local control and potentially improve treatment tolerability. We reviewed records of patients with stage I-II EC treated with VB at our institution from 1/2017-6/2019 (3D group). We compared patient, tumor, and treatment characteristics to a historic cohort treated at our institution between 2011-2015 (2D group). Patients had minimum 6 months follow-up. Patients with stage III-IV disease, prior or concurrent pelvic radiation, or synchronous pelvic primary were excluded. Recurrence-free survival (RFS), local control, and treatment tolerability (any acute toxicity recorded prospectively on treatment) were analyzed. Statistical analyses were performed in Stata and included Chi-square and Wilcoxon Rank Sum tests to compare between-group characteristics and Cox proportional hazards analysis to assess RFS. A total of 403 patients were included (n = 99/3D and n = 304/2D). The 3D vs. 2D groups were similar in age, weight, gravidity, cylinder sizes, histology, and chemotherapy, though differed in race (p = 0.004) and % stage IA/IB/II (p = 0.022). VB dose was similar, with majority receiving either 18Gy/3 fractions (46% vs. 46%) or 21Gy/3 fractions (40% vs. 38%) to 5mm depth. In the 3D cohort, however, the proportion of vagina treated was less (3D median 50%, IQ range 43-56% vs. 2D median 60%, IQ range 50-67%; p<0.001). and standard dwell times were decreased to improve bowel dosimetry in 21%. Median follow-up times were 524 and 903 days in the 3D and 2D groups, respectively. There was no significant difference in RFS (HR 0.48, 95% CI 0.19-1.22, p = 0.122). Vaginal recurrence occurred in 1 and 2 patients, respectively, with recurrence distal to treated mucosa in all cases. There were no significant differences between groups in rates of acute GI (14.1% vs. 10.9%, p = 0.376), GU (8.1% vs. 11.2%, p = 0.380), or vaginal (20.2% vs. 19.4%, p = 0.827) toxicities. All toxicities were mild (grade 1-2). In the 3D patients, there were no significant differences (p>0.05) in bowel, rectal, or bladder Dmax or D2cc values between those with vs. without symptoms on treatment. 3D vs. 2D planning maintained excellent local control despite selectively reducing the volume of prescription isodose overlapping with bowel. Treatment tolerability was favorable in both groups with no benefit observed in the short term, however, longer follow-up is needed to compare late outcomes.