Clinicopathological characteristics predict lymph node metastases in ypT0-2 rectal cancer after chemoradiotherapy

markdownabstract__Aims:__ Changes in rectal cancer treatment include increasing emphasis on organ preservation. Local excision after chemoradiotherapy (CRT) for rectal cancer with excellent clinical response reduces morbidity and mortality compared to total mesorectal excision, although residual lymph node metastases (LNM) may cause local recurrence. Our aim is to identify clinicopathological factors predicting the presence of residual LNM in rectal cancer patients with ypT0-2 tumours after neoadjuvant CRT. These risk factors may help to select patients who can be spared radical surgery without compromising oncological outcomes. __Methods and results:__ Rectal cancer patients with ypT0-2 tumours after CRT and radical resection from five centres treated between June 1999 and February 2012 were included. Histopathology was reviewed extensively. Clinicopathological characteristics and their association with residual LNM were investigated. Of 657 consecutive CRT-treated rectal cancer patients 210 with ypT0-2 disease were included. Residual nodal disease was found in 44 cases (21.0%). Independent predictors of LNM were clinical nodal involvement (cN+) [odds ratio (OR): 2.79, 95% confidence interval (CI): 1.04-7.48, P = 0.042], high-grade histopathology assessed in the post-CRT resection specimen (OR: 6.46, 95% CI: 1.23-34.02, P = 0.028) and residual tumour diameter (RTD) ≥10 mm (OR: 2.54, 95% CI: 1.06-6.09, P = 0.036). An algorithm combining these factors stratified patients adequately according to LNM risk, independently of ypT category. __Conclusions:__ Clinical nodal involvement, high-grade histopathology and RTD ≥10 mm are strong and independent predictors of residual nodal disease in rectal cancer patients with ypT0-2 tumours after CRT. Risk stratification based on these factors may help to identify patients suitable for organ preserving therapy and should be validated in appropriately selected populations