Clinical lymph node staging by imaging in colorectal cancer: A flip of the coin?

e15160 Background: Clinical lymph node staging by MRI and CT is important in stratification for neoadjuvant therapy in colorectal cancer. Overstaging may result in unnecessary neoadjuvant therapy, but understaging may refrain patients from adequate preoperative treatment. This study aims to provide insight in current daily practice in clinical lymph node staging in CRC in the Netherlands. Methods: All patients with primary CRC, diagnosed between 2003-2014, who underwent lymph node dissection were selected from the nationwide population-based Netherlands Cancer Registry (n=100,211). Trends in patient- and tumor-characteristics, and lymph node staging were analyzed. For the years 2011-2014, sensitivity, specificity, positive (PPV) and negative predictive value (NPV) were calculated for clinical lymph node staging, with histology as the gold standard. Only patients without preoperative treatment were analyzed. Since prospective studies have shown that 5x5 Gy radiotherapy (RT) followed by total mesorectal excision within 10 days does not lead to nodal downstaging, an additional analysis was performed in this group. Results: The proportion clinically positive lymph nodes increased significantly between 2003-2014; from 7% to 22% for colon cancer and from 7% to 53% for rectal cancer. The proportion histological positive lymph nodes remained fairly stable over time (±35% colon, ±33% rectum). During 2011-2014, clinical lymph node staging was available in the registry in 86% of colon cancer patients, 92% of rectal cancer patients without neoadjuvant treatment and 95% of rectal cancer patients with 5x5 Gy RT. The parameters based on data from this period are presented in table 1. Conclusions: With a sensitivity and PPV of approximately 50%, clinical lymph node staging is about as accurate as flipping a coin. This leads to overtreatment in patients with rectal cancer with neoadjuvant RT. Acceptable specificity and NPV limit the risk of undertreatment. [Table: see text]