Impact of gastric cancer treatment pathways on patient outcomes in a community oncology practice setting.

Abstract

299 Background: Gastric cancer (GC, including gastroesophageal junction adenocarcinoma) pathways have been implemented and refined since 2010 in the US Oncology Network (USON), a community-practice-based network. This study was designed to evaluate the impact of 4 pathway periods (PP): pre-pathway: pre-Aug ‘10; Level 1: Sept ‘10-Nov ‘14; Clear Value, Dec ‘14-Feb ‘17; Value: After Mar ‘17, on treatment heterogeneity, treatment duration, and overall survival (OS). Methods: Adult patients were eligible if they were treated at a participating USON site and were diagnosed with and treated for GC; follow up was through Mar 2019. Heterogeneity was measured by the Herfindahl-Hirschman Index (HHI), which is evaluated from 0.0 to 1.0 (complete to no heterogeneity). Time to treatment failure (TTF) was defined as initiation of the line of therapy until start of the subsequent line of therapy or death. OS was estimated from start of first-line (1L) therapy. Time-to-event outcomes were estimated using Kaplan-Meier. Results: Of 3191 eligible patients, 2297 received treatment for advanced/metastatic disease. Of these, patient median age was 65.3 years, 60% were male, 70% were initially diagnosed with stage IV disease. Pre-pathway, common 1L regimens were single-agent fluorouracil (15%) and docetaxel-cisplatin-fluorouracil (14%); FOLFOX (45%) was most frequent during the value PP, 941 (41%) received second-line (2L) therapy. During Level 1 PP, single-agent irinotecan (11%) was most common during 2L, whereas in Value PP, ramucirumab+paclitaxel (43%) was most common. HHI and TTF are presented in the table. Median OS was 12.6 months (95% confidence interval, CI: 11.9, 13.5), which did not change significantly over PPs. Conclusions: 1L and 2L heterogeneity was initially high, and was reduced over time; TTF showed modest increase. OS data are limited by high levels of censoring in the latter PPs, which reduces the ability to evaluate changes in more recent years. [Table: see text]