Validation of the Social Security Administration Life Tables (2004-2014) in Localized Prostate Cancer Patients within the Surveillance, Epidemiology, and End Results database

Abstract

Backround: Accurate life expectancy estimation is crucial in clinical decision-making including management and treatment of clinically localized prostate cancer (PCa).

Objective: We hypothesized that Social Security Administration (SSA) life tables' derived survival estimates closely follow observed survival of PCa patients. To test this relationship, we examined 10-yr overall survival rates in patients with clinically localized PCa and compared it with survival estimates derived from the SSA life tables.

Patients and methods: Within the Surveillance, Epidemiology, and End Results database (2004), we identified patients aged >50-<90yr. Follow-up was at least 10 yr for patients who did not die of disease or other causes.

Outcome measurements and statistical analysis: Monte Carlo method was used to define individual survival in years, according to the SSA life tables (2004-2014). Subsequently, SSA life tables' predicted survival was compared with observed survival rates in Kaplan-Meier analyses. Subgroup analyses were stratified according to treatment type and D'Amico risk classification.

Results and limitations: Overall, 39191 patients with localized PCa were identified. At 10-yr follow-up, the SSA life tables' predicted survival was 69.5% versus 73.1% according to the observed rate (p<0.0001). The largest differences between estimated versus observed survival rates were recorded for D'Amico low-risk PCa (8.0%), brachytherapy (9.1%), and radical prostatectomy (8.6%) patients. Conversely, the smallest differences were recorded for external beam radiotherapy (1.7%) and unknown treatment type (1.6%) patients.

Conclusions: Overall, SSA life tables' predicted life expectancy closely approximate observed overall survival rates. However, SSA life tables' predicted rates underestimate by as much as 9.1% the survival in brachytherapy patients, as well as in D'Amico low-risk and radical prostatectomy patients. In these patient categories, an adjustment for the degree of underestimation might be required when counseling is provided in clinical practice.

Patient summary: Social Security Administration (SSA) life tables' predicted life expectancy closely approximate observed overall survival rates. However, SSA life tables' predicted rates underestimate by as much as 9.1% the survival in brachytherapy patients, as well as in D'Amico low-risk and radical prostatectomy patients.

Keywords: Life expectancy; Prostate cancer; SEER database; Social Security life tables; Survival.