KRAS testing of mCRC patients in a community-based oncology setting.

Abstract

364 Background: In late 2008, treatment guidelines were updated to recommend KRAS testing for patients at diagnosis (Dx) of metastatic colorectal cancer (mCRC). This study explored KRAS testing rates over time and compared characteristics of those mCRC patients who were tested and those not tested in a community-based setting. Methods: Patients ≥18 years old, with a mCRC Dx between 1/1/2008 - 12/31/2011, no other primary cancer, and 2+ outpatient visits were selected from the ACORN Data Warehouse (Memphis, TN). Text mining of progress notes and random forests modeling predicted if KRAS testing was done. If tested, test date and result were abstracted from charts. The proportions of those tested overall and at Dx (pre and ≤ 60 days post-mCRC Dx) were calculated by year. Chi-square and t-tests were used to compare demographic and cancer characteristics at Dx and clinical history of those tested and not. Results: Among 1,363 patients identified, 47.5% (n=648) were KRAS tested. Overall 48.1% were KRAS WT, 42.3% KRAS mutant, and 9.6% unknown. Proportion of eligible patients who underwent KRAS testing in pre-2008, 2008, 2009, 2010, and 2011 was 5.8%, 15.9%, 29.1%, 26.9%, and 28.0% respectively. Of those tested, 6.2% were tested at Dx pre-2008 and 27.4%, 65.4%, 73.2%, and 78.4% were at Dx in 2008, 2009, 2010, and 2011 respectively. KRAS tested patients were younger (59.5% vs. 51.5% <65 years, p=0.004), had more baseline comorbid conditions (3.7±4.0 vs. 2.7±3.4, p<0.001), were more likely to have bone (12.1% vs. 8.5%), lung (36.6% vs. 26.2%) and liver (73.5% vs. 66.2%) metastases, were less likely to have peritoneal (3.7% vs. 10.1%) metastases (all p<0.05), and had poorer performance status (PS) (ECOG >1: 41.9% vs. 34.0%, p=0.042). No significant differences were found by cancer location (colon, rectum, both), if the cancer was newly diagnosed or a recurrence, or in tumor histologic grade. Conclusions: While KRAS testing increased in 2008-2009 corresponding to changes in treatment guidelines and anti-EGFR product labels, the majority of patients were not tested. The proportion tested at diagnosis continued to increase annually from 2008-2011. Patients with greater comorbidities and poorer PS are more likely to undergo KRAS testing than healthier patients.