Treatment patterns and outcomes in metastatic colorectal cancer (mCRC).

699 Background: Treatment (tx) selection for mCRC is multifactorial and affected by patient (pt) and disease factors, comorbidities, and previous tx. Few agents existed for therapy after second progression (i.e. third line (3L) at the time of this study. We aimed to capture mCRC tx patterns up to 3L and associated clinical and economic outcomes. Methods: A retrospective observational study of mCRC pts initiating 1L tx from 1/1/07 to 6/30/11 and 3L tx before 6/30/12. Data was extracted from The US Oncology Network (USON)/McKesson Specialty Health EHR and claims database. Pts on clinical trial or with another cancer diagnosis were excluded. Tx was organized into tx backbone categories and lines of therapy (LOTs), and adherence to NCCN guidelines and USON pathways assessed. The Kaplan-Meier method was used to estimate overall survival (OS). A multivariable generalized linear regression model was used to evaluate total costs during 3L tx as a function of 3L covariates by Medicare allowable. Clinical outcomes ...