Chemoradiotherapy in octogenarians as primary treatment for muscle-invasive bladder cancer
Victor A McPherson1, George Rodrigues2, Glenn Bauman2
1Division of Urology.
Introduction:
While radical cystectomy is the gold standard for muscle-invasive bladder cancer (MIBC), in octogenarians cystectomy results in a higher perioperative mortality rate (6.8-11.1%) than in younger patients (2.2%). Trimodality therapy is a bladder-sparing regimen composed of transurethral resection of bladder tumour (TURBT) and chemoradiotherapy, with intent for salvage cystectomy, and has a 62.5-90% initial complete response rate. In this study, we evaluate TURBT and chemoradiotherapy without salvage cystectomy in medically inoperable octogenarian patients.
Methods:
We identified a retrospective cohort of patients aged 80-89 years with invasive urothelial carcinoma who received combination chemoradiotherapy between 2008 and June 2014. Outcomes were evaluated by Kaplan-Meier (KM) and Cox regression.
Results:
In 40 patients, the mean age was 84.5 years (interquartile range [IQR] 83-86). Seventeen patients received hypofractionated, low-dose radiotherapy (LD) (37.5-40 Gy), while 23 received conventionally fractionated radiotherapy (high-dose [HD]) (50-65 Gy). Mean overall survival (OS) was 20.7 months (IQR 12.75-23.25), while mean recurrence-free survival (RFS) was 13.75 months (IQR 3.75-16.5). Patients receiving HD radiotherapy showed improved OS and local RFS (LRFS) without significant differences in Grade 3-4 toxicities. Univariate Cox regression identified hydronephrosis as a predictor of worse OS and local recurrence and HD radiotherapy as a predictor of improved OS and local recurrence rates. Multivariate Cox regression identified hydronephrosis to be a significant predictor of LRFS.
Conclusions:
Primary chemoradiotherapy for inoperable patients with MIBC resulted in a three-year OS of 54.9% (comparable to cystectomy) and three-year RFS of 42.3%. Superior outcomes were associated with more aggressive chemoradiotherapy treatment. The results of the local control subanalyses in this study are hypothesis-generating due to the limited patient numbers in the cohort.
