IntroductionNephroureterectomy is commonly performed for high-grade (HG) upper tract (UT) urothelial carcinoma (UC). However, some patients may benefit from a de-escalation of surgical management, particularly for noninvasive disease and carcinoma in situ (CIS). Bacillus Calmette–Guerin (BCG) is currently the only guideline-recommended endoluminal treatment option. Gemcitabine/Docetaxel (Gem/Doce) has shown promising efficacy as a treatment for noninvasive HG UTUC, though a comparison to BCG is lacking. We report the outcomes of patients treated with endoluminal Gem/Doce vs. BCG for UT-CIS. MethodsA single-institutional retrospective review of patients treated with Gem/Doce vs. BCG for UT-CIS was performed. Treatment was instilled via nephrostomy or retrograde ureteral catheter. In both treatment groups, induction consisted of 6 weekly instillations. Maintenance was initiated if disease-free and consisted of 6 monthly instillations in the Gem/Doce group and a reduced dose (one-tenth) 3-week course at 3 months in the BCG group. Recurrence was defined as biopsy-proven disease or HG cytology. ResultsThe final cohort included 53 patients with 65 upper tract units; 31 received BCG and 34 received Gem/Doce. Median follow-up was 88 and 29 months in the BCG and Gem/Doce groups, respectively. Presenting pathology included biopsy-proven CIS and HG cytology in 9.7% and 90% of the BCG group, and 8.8% and 91% of the Gem/Doce group, respectively. The 2-year estimates for recurrence-free and nephroureterectomy-free survival were 61% and 89% for the BCG group and 54% and 100% for the Gem/Doce group, respectively. Upon multivariable analysis, instillation via percutaneous nephrostomy tube was associated with an increased risk of recurrence (HR 3.89, 95% CI 1.59–9.53). The development of any symptom was not statistically different between treatment groups (P = 0.12). There were 2 treatment-related deaths that occurred, 1 within each treatment group. ConclusionEndoluminal Gem/Doce and BCG have similar oncological outcomes and major adverse event rates in the treatment of UT-CIS. Further prospective evaluation is warranted.