Presentation Authors: Matteo Soligo*, Michele Colicchia, Alessandro Morlacco, Padova, Italy, Luca Boeri, Vidit Sharma, Igor Frank, Stephen Boorjian, R. Jeffrey Karnes, Rochester, MN
Introduction: The 5-year bladder cancer recurrence rate after radical cystectomy (RC) ranges 15%-70%; 30% of these patients are diagnosed with local recurrence (LR). Cisplatin-based chemotherapy (CHT) is the standard treatment for LR, whilst radiotherapy (RT) and surgery have marginal roles. However, salvage treatments are rarely curative. The aims of the present study are: to describe patterns of LR; to identify the most effective salvage treatments and any predictors of CSS and OS.
Methods: We identified 145 patients who underwent RC in a tertiary referral center between 1980-2017 and experienced LR during follow-up. Exclusion criteria were: incomplete data; cM1 at RC; non-urothelial histology at RC; rare recurrence sites. LR were grouped as: recurrences in pouch/neobladder, bladder fossa/pelvis/vaginal cuff, regional nodes, >2 of the previous sites. Descriptive statistics were used to show baseline data. Multivariable Cox regression analysis was used to identify any predictors of CSS and OS.
Results: 71 (49%) patients had non organ confined disease; 24 (16.6%) had nodal involvement at RC. Median time to recurrence was 10.7 (5.9-24.6) months; median follow-up time was 31 (15.9-72.9) months. All-cause and cancer-specific death rates were 86.9% (126 pts) and 71% (103). The most common site of LR was bladder fossa/pelvis/vaginal cuff (82, 56.6%). 21.4% of patients had surgery, 21.4% CHT, 17.2% CHT+RT as salvage treatments. At multivariable analysis, higher age at LR (HR 1.02, p .011), bladder fossa/pelvis/vaginal cuff recurrences (HR 3.2, p.003) and multiple recurrence sites (HR 2.4, p.03) were associated to shorter CSS. CHT (HR .45, p.015), CHT+RT (HR .39, p.006) and miscellaneous multimodal treatments (HR .45, p.037) were significantly associated to longer CSS, without a significant difference among them. Higher age at LR (HR 1.03, p < .0001) and bladder fossa/pelvis non nodal/vaginal cuff recurrence (HR 2.4, p.012) were associated to shorter OS. All treatments except for RT were associated to longer OS (all p < .005), but none demonstrated a significant better effect.
Conclusions: LR are associated to poor prognosis, but outcomes differ according to LR patterns. A multimodal approach including CHT may achieve the best effect in prolonging CSS and OS.