Liver-directed therapies for colorectal cancer liver metastasis (CLRM): A Surveillance, Epidemiology, and End Results (SEER)-Medicare analysis.

577 Background: Oligometastatic CLRM comprises a distinct subset of stage IV colorectal cancer. Liver directed therapies (LDT) including surgery, ablation, radiation, and transarterial chemotherapy/embolization have been shown to improve cancer outcomes in smaller series. We sought to evaluate utilization of LDT and their impact on survival in patients with CLRM in a population-based database. Methods: We analyzed linked SEER-Medicare data. Eligible patients were ≥ 66 years, diagnosed between 1992-2009, carry a code for secondary malignancy of the liver (ICD-9 197.7) and survived ≥ 30 days after diagnosis. LDT (yes vs. no) were defined by ICD-9 and CPT codes of surgery (hepatectomy), ablation (e.g. radiofrequency and cryoablation), radiation (e.g., stereotactic surgery and brachytherapy), and transarterial/embolization (e.g., radioembolization). Treatment and non-treatment groups were matched using a propensity score comprised of a comprehensive set of patient and tumor characteristics including: age, sex, race, marital status, Medicaid status, and tumor histology, grade, and location. Cox Proportional Hazard models were used to compare the impact of the LDT on overall survival among propensity-matched pairs of treated and untreated patients. Results: LDT were performed in 12.7% (n = 1,793) of all 14,150 patients. There were 13.7% of patients over 85 years. Females were 52.1% and 83 % White. Off all, 5.4% had surgery, 3.9% had ablation, 4.6% had radiation, and 1.6% had transarterial chemotherapy/embolization. Unadjusted 5 year overall-survival (OS) was 26.9% for those underdoing LDT vs. 7.5% who did not (HR = 0.44, (CI 0.41,0.46)). Cox modeling demonstrated a survival benefit for each LDT with HR of 0.36 (0.33,0.39) for surgery, 0.35 (0.32,0.39) for ablation, 0.78 (0.72, 0.85) for radiation and 0.42 (0.36,0.49) for transarterial/embolization. Conclusions: While use of LDT for CLRM in this national sample of Medicare patients were low, those who received treatment had markedly improved survival compared to matched patients who did not. Surgery and ablation were the most effective therapies. These findings require evaluation in a prospective clinical study.