HSR&D Citation Abstract
Search | Search by Center | Search by Source | Keywords in Title
Comparative effectiveness of non-cisplatin (cis)-based first-line (1L) regimens in patients with metastatic urothelial carcinoma (mUC): Veterans Affairs control cohorts vs. IMvigor210
Vander VN, Guerin A, Ionescu-Ittu R, Shi S, Wu E, Lin S, Hsu L, Kai-Uwe S, de Ducia S, Wang J, Li S, Derleth C, Liu S, Shi L, Leppert JT. Comparative effectiveness of non-cisplatin (cis)-based first-line (1L) regimens in patients with metastatic urothelial carcinoma (mUC): Veterans Affairs control cohorts vs. IMvigor210. [Abstract]. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2018 Feb 6; 36(6):abstract 496.
Background: For mUC patients (pts) ineligible for cis, 1L carboplatin (carbo)-based regimens and checkpoint inhibitors are recommended, but comparative data are lacking. We assessed OS in pts given 1L atezolizumab (atezo; anti-PD-L1) from Ph 2 study IMvigor210 and Veterans Affairs (VA) mUC pts treated with carbo standards. Methods: Carbo-based control cohorts were derived from the VA Health Care Network databases (diagnosed 2006-2017), with treatment (tx) defined per NCCN guidelines V2.2005. IMvigor210 eligibility criteria were used to refine non-cis VA cohorts. Prespecified OS predictors were balanced between VA and IMvigor210 pts using entropy (e)-balance weighting. OS was evaluated by Kaplan-Meier (KM) and Cox regression (weighted). Sensitivity analyses used stabilized inverse probability weighting (sIPW) and statistical adjustments. Results: Of 2056 VA pts included, 926 received carbo-based tx, of whom 380 received carbo+gemcitabine [CarboGem]; after IMvigor210 exclusion criteria, the carbo-based cohort included 282 pts, of whom 120 had CarboGem. Non-proportional hazards were seen with KM curves crossing at 5 (CarboGem) and 9 mo (carbo based). OS data are in the table. Weighted analyses showed OS benefit for atezo vs CarboGem and delayed benefit vs carbo-based tx. HRs from sensitivity analyses were consistent in direction and significance. Conclusions: We compared OS in mUC pts receiving 1L non-cis regimens in VA (carbo) and IMvigor210 (atezo) mUC cohorts. A lower risk of death was seen for atezo vs carbo, manifesting after a few months of tx. Randomized studies may validate these 1L mUC tx trends. Clinical trial information: NCT02951767.
OS in VA and IMvigor210 mUC cohorts by 1L tx.
Pts, n (events, %)282 (68%)120 (71%)110 (51%)
Median (95% CI), mo12.1 (9.6, 14.7)8.7 (5.7, 15.3)15.0 (9.3, 16.5)
Milestone (95% CI), %
6 mo75 (67, 82)64 (45, 78)68 (58, 76)
12 mo50 (41, 58)39 (25, 53)56 (46, 64)
24 mo27 (19, 35)22 (11, 34)45 (35, 55)
Cox model HR (95% CI); P value
E-balance weighting (main)0.83 (0.62, 1.12); 0.220.67 (0.48, 0.94); 0.02-
sIPW (sensitivity)0.78 (0.58, 1.06); 0.110.73 (0.52, 1.02); 0.06
a Excludes locally advanced UC.