Variation in Lymph Node Assessment for Colon Cancer at the Tumor, Surgeon, and Hospital Level

Michael E Egger,Yana Feygin,Maiying Kong,Triparna Poddar,Indranil Ghosh,Qian Xu,Ryan M McCabe,Kelly M McMasters,C Tyler Ellis

DOI: https://doi.org/10.1097/xcs.0000000000000963

2024-03-15

Journal of the American College of Surgeons

Abstract:BACKGROUND: We hypothesized that tumor- and hospital-level factors, compared with surgeon characteristics, are associated with the majority of variation in the 12 or more lymph nodes (LNs) examined quality standard for resected colon cancer. STUDY DESIGN: A dataset containing an anonymized surgeon identifier was obtained from the National Cancer Database for stage I to III colon cancers from 2010 to 2017. Multilevel logistic regression models were built to assign a proportion of variance in achievement of the 12 LNs standard among the following: (1) tumor factors (demographic and pathologic characteristics), (2) surgeon factors (volume, approach, and margin status), and (3) facility factors (volume and facility type). RESULTS: There were 283,192 unique patient records with 15,358 unique surgeons across 1,258 facilities in our cohort. Achievement of the 12 LNs standard was high (90.3%). Achievement of the 12 LNs standard by surgeon volume was 88.1% and 90.7% in the lowest and highest quartiles, and 86.8% and 91.6% at the facility level for high and low annual volume quartiles, respectively. In multivariate analysis, the following tumor factors were associated with meeting the 12 LNs standard: age, sex, primary tumor site, tumor grade, T stage, and comorbidities (all p < 0.001). Tumor factors were responsible for 71% of the variation in 12 LNs yield, whereas surgeon and facility characteristics contributed 17% and 12%, respectively. CONCLUSIONS: Twenty-nine percent of the variation in the 12 LNs standard is linked to modifiable factors. The majority of variation in this quality metric is associated with non-modifiable tumor-level factors.

surgery