Defining and predicting early recurrence in patients with locally advanced rectal cancer treated with neoadjuvant...

The definition of "early recurrence (ER)" after rectal cancer surgery is currently unclear.

To determine an evidence-based cut-off to distinguish early and late recurrence (LR) for patients with rectal cancer and compare the clinicopathological factors between the two groups.

Patients who underwent neoadjuvant chemoradiotherapy (nCRT) and radical resection for locally advanced rectal cancer were included. A minimum p-value approach was used to evaluate the optimal cut-off value of recurrence-free survival to divide the patients into ER and LR groups based on overall survival. A logistic regression model was used to assess risk factors for ER.

A total of 763 patients were included, of which 167 (21.9%) experienced recurrence. The optimal cut-off value of recurrence-free survival to differentiate between ER (n = 125, 74.9%) and LR (n = 42, 25.1%) was 24 months (P = 0.000001). The median postrecurrence survival of ER and LR was 12 months and 22 months, respectively (p = 0.028). The most common recurrent sites in patients with ER and LR were lung metastases, the incidence of liver metastases, however, differed considerably in ER and LR (27.2% vs 9.5%, P = 0.019). Risk factors including elevated preoperative carcinoembryonic antigen (CEA), higher ypTNM stage, positive circumferential resection margin (CRM), and perineural invasion were significantly associated with ER.

A recurrence-free interval of 24 months is the optimal cut-off value for defining ER versus LR. Elevated preoperative CEA, higher ypTNM staging, positive CRM, and perineural invasion were associated with ER of locally advanced rectal cancer.