Extended lymph node dissection in colorectal cancer surgery. Reliability and reproducibility in assessments of operative reports

Stratification of surgical strategies or techniques by operative reports is rarely validated. This study evaluates the assessment of lymph node dissection from operative reports and the possible survival benefit of extended lymph node dissection. The operative reports of 342 colorectal resections (R0) were assessed twice by two surgeons. The lymph node dissection was classified as limited or extended according to a novel scheme. Intraobserver reproducibility and interobserver reliability were evaluated by kappa (kappa) statistics and a Cox model. For colonic resections, the reproducibility of assessments was moderate or substantial (kappa: 0.42-0.75), and the reliability was moderate (kappa: 0.54-0.58). For rectal resections, reproducibility was moderate (kappa: 0.45-0.58), and reliability was fair (kappa: 0.29-0.36; all kappa values: p < 0.001). The 5-year survival rates of colonic cancer patients subject to a limited or extended procedure were 52% (45-60%, 95% confidence interval) and 69% (53-84%; p = 0.034), respectively. The hazard ratios of extended lymph node dissection (limited as reference) were 0.34 (0.14-0.86; p = 0.023) for stage I and II and 1.11 (0.50-2.44) for stage III. In rectal cancer patients, the 5-year survival rates of a limited or extended procedure were 63% (54-72%) and 55% (37-73), respectively. Valid assessment of lymph node dissection can be obtained from operative reports. Extended lymph node dissection improves long-term survival rates of colonic cancer patients.