Abstract
Background: Despite evidence that liver resection improves survival in patients with colorectal cancer liver metastases (CRCLM) and may be potentially curative, there are no population-level data examining utilization and predictors of liver resection in the United States. Methods: This is a population-based cross-sectional study. We abstracted data on patients with synchronous CRCLM using California Cancer Registry from 2000 to 2012 and linked the records to the Office of Statewide Health Planning Inpatient Database. Quantum Geographic Information System (QGIS) was used to map liver resection rates to California counties. Patient- and hospital-level predictors were determined using mixed-effects logistic regression. Results: Of the 24 828 patients diagnosed with stage-IV colorectal cancer, 16 382 (70%) had synchronous CRCLM. Overall liver resection rate for synchronous CRCLM was 10% (county resection rates ranging from 0% to 33%) with no improvement over time. There was no correlation between county incidence of synchronous CRCLM and rate of resection (R2 =.0005). On multivariable analysis, sociodemographic and treatment-initiating-facility characteristics were independently associated with receipt of liver resection after controlling for patient disease- and comorbidity-related factors. For instance, odds of liver resection decreased in patients with black race (OR 0.75 vs white) and Medicaid insurance (OR 0.62 vs private/PPO); but increased with initial treatment at NCI hospital (OR 1.69 vs Non-NCI hospital), or a high volume (10 + cases/year) (OR 1.40 vs low volume) liver surgery hospital. Conclusion: In this population-based study, only 10% of patients with liver metastases underwent liver resection. Furthermore, the study identifies wide variations and significant population-level disparities in the utilization of liver resection for CRCLM in California. © 2020 The Authors. Cancer Medicine published by John Wiley & Sons Ltd