Lymph node counts as an indicator of quality at the hospital level in colorectal surgery

Background: Substantial evidence suggests that the number of lymph nodes examined in colorectal cancer (CRC) is a powerful predictor of outcomes. However, the lymph node count as a benchmark of quality in CRC is controversial. We sought to examine the impact of lymph node counts on disease-specific survival (DSS) of CRC patients at the hospital level. Study Design: This study used data obtained between 1994 and 2003 from Region 5 of the California Cancer Registry. Hospitals in Region 5 of the California Cancer Registry were stratified according to the median number of nodes examined and grouped according to the median number of nodes examined, <7, 7 to 9, and <10. These hospital groups were then evaluated for the frequency of meeting the 12-node threshold, frequency of positive lymph nodes, and DSS at the hospital level. Results: Median number of nodes examined in group A was 4 (mean 5.6, SD 5.9), in group B was 8 (mean 9.7, SD 8.5), and in group C was 10 (mean 11.3, SD 9.2). In group A, 13.7%, in group B 32.8%, and in group C, 42.8% met the 12-node threshold. The frequency of N1 and N2 disease for group A was 20.7% and 9.1%, 19. 7% and 11.1% for group B, and 20.1% and 11.3% for group C (p = 0.12). Five-year DSS was 72.7% for group A, 73.7% for group B, and 76.7% for group C (p = 0.002). DSS survival of N0 patients for group A was 78.6%, 81.5% for group B, and 85.1% for group C (p < 0.0001). There was no statistically significant difference in DSS for N1 (p = 0.18) or N2 (p = 0.90) between the 3 groups. Conclusions: Lymph node counts can have value as a benchmark of surgical/pathologic quality in node-negative CRC. These results question the value of lymph node counts as a benchmark of surgical/pathologic quality for node-positive CRC.