Jim Zhong, MD
In a propensity-matched analysis of a national cancer registry, patients who underwent radical cystectomy or concurrent chemoradiation to treat urothelial cancer demonstrated statistically similar median overall survival (OS) rates. Further analysis suggests that the hazard ratios between the two treatments change over time and is hypothesis generating, said Jim Zhong, MD, of Winship Cancer Institute at Emory University.
Those patients with American Joint Committee on Cancer tumor stages II-III, N0, M0, urothelial cancer were included in the analysis. Zhong said only patients treated with definitive intent with either radical cystectomy or concurrent chemotherapy and radiation were included. In addition, only patients who underwent a maximal tumor resection, followed by a definitive course of radiation with chemotherapy were included in the bladder preservation chemoradiation arm.
The primary endpoint of the study was OS and researchers used Kaplan-Meier, log-rank test, and multivariable Cox proportional hazards regression for statistical analysis. Zhong reported that propensity score matching was used to reduce treatment selection bias. Also, a time-varying covariate analysis was generated to evaluate interactions between treatment and survival time.
Zhong noted striking differences in patient characteristics at baseline (P
<.001). For example, the median age at diagnosis for the bladder preservation arm was much older than the cohort who had surgery. “For the T-stage score, there were approximately 2% more patients with T3 tumors in the surgery group compared with the bladder preservation group, and patients treated with bladder preservation tended to live closer to the treatment center,” he said.
Given these large baseline differences that favored the surgery group, it is not surprising that there was a significant OS advantage for the surgery group over the bladder preservation group, said Zhong. Thus propensity score matching was used to match these cohorts based on age, facility type, Charleson-Deyo score, primary tumor site, and stage of the tumor.
After propensity matching, there were approximately 1000 patients per cohort. “We can see that median overall survival in the chemoradiation arm was 2.7 years (2.4, 3) and 3 years in the surgery arm (2.4, 3). These were statistically similar with overlapping confidence intervals,” he said. “These data point to the fact that these cohorts might have statistically similar outcomes.”
OS in the propensity-matched analysis demonstrated that the hazard ratio is above 1 for surgery, which is favorable to the chemoradiation arm. Over time, the hazard ratio for chemoradiation falls below 1, which favors the surgery arm.
Hazard ratios between the two cohorts may change over time, so the researchers employed a time-varying hazard ratio model to explore the possibility of changing hazard ratios. They found that the hazard ratio for surgery decreases by a factor of 0.85 (0.79-0.92) per year for surviving patients.
“However, with the limited amount of data available, all I can say is that this is hypothesis generating,” added Zhong.
On multivariable analysis, the chemoradiation arm was not significantly associated with OS. The surgery arm had an HR of 0.93 (0.85-1.02; P
= .105). “However, we did find that greater age at diagnosis (HR, 1.03 [1.02-1.03]; P
< .001), more advanced T stage (HR, 0.78 [0.71-0.85]; P
< .001), and higher Charlson-Deyo score was negatively associated with OS as we would expect. We also found that tumors of the bladder wall (HR, .80 [0.74-0.86], P
< .001) and bladder neck (HR, 0.69 [0.57-0.83]; P
< .001) were positively associated with survival.
In addition to the study being a retrospective registry study, Zhong noted other limitations, including the absence of data for local control, cause of death, treatment-related morbidity, a lack of in-depth chemotherapy information, and an inability to clearly ascertain whether patients undergoing bladder preservation received subsequent radical cystectomy.
Zhong J, Switchenko J, Jegadeesh NK, et al. Comparison of outcomes in patients with muscle-invasive bladder cancer treated with radical cystectomy versus bladder preservation: an NCDB analysis. Presented at: 2016 ASTRO Annual Meeting; Boston, Massachusetts, September 25-28, 2016.