PD2-12 PATTERNS OF SURVEILLANCE IMAGING AFTER NEPHRECTOMY IN THE MEDICARE POPULATION

Abstract

Additional eligibility criteria for the meta-analysis were (1)malignant tumors â&¤7 cm or benign tumors of any size, and (2)survival analysis performed with adjustment for confounding variables. The pooled estimate of the hazard ratio (HR) was computed using the inverse variance weighting approach with random effects. Heterogeneity was assessed with Cochran†s Q test. RESULTS: A total of 26 studies were eligible for the systematic review. Of these, 16 studies reported only the unadjusted survival analysis despite substantial differences between PN and RN pts in baseline characteristics, or imposed no upper bound on the tumor size, thus including a large number of locally advanced cases, mostly treated with RN. Because these studies had a strong selection bias in favor of PN pts which were younger with less advanced tumors, they were excluded from the meta-analysis. Meta-analysis of the remaining 10 studies is summarized in Table 1. The pooled HR indicated improved OS after PN, although there was substantial disagreement between individual studies in reported HRs with strong evidence of statistical heterogeneity (p1⁄40.006). CONCLUSIONS: A meta-analysis of 10 observational studies resulted in a pooled HR for OS favoring PN, but with substantial evidence of heterogeneity. Because residual confounding could be present in these observational studies despite adjustment for measured covariates, another randomized trial of PN vs. RN may be needed to definitively confirm or refute findings of EROTC 30904.