Background: After patients with stage IV colorectal cancer undergo curative surgical resection, there is a large risk for recurrence. To establish optimal surveillance guidelines, an understanding of the temporal risk factors for recurrence is necessary.
Objective: The primary aim of our study was to determine predictors for early (within 1 year), middle (1-2 years), and late (2 years or later) recurrence following curative resection in patients with stage IV colorectal cancer.
Design: This was a retrospective cohort study.
Settings: The study was conducted at multiple institutions.
Patients: The retrospective cohort study comprised 1070 patients with stage IV colorectal cancer after an R0 resection for the primary and metastatic lesions in 19 institutions from January 1997 to December 2007.
Main outcome measures: Risk factors for early, middle, and late recurrence were determined by logistic regression and Cox proportional hazards models.
Results: The overall recurrence rate was 73% (784/1070). Cancer-specific survival was 29.5 months, and recurrence-free survival was 8.9 months. Early recurrence occurred in 488 (62%), middle recurrence in 184 (24%), and late recurrence in 112 (14%). In multivariable analysis, early recurrence risk factors included rectum site, depth of tumor invasion (T4), increasing N-staging, venous invasion, and liver metastasis. Late recurrence risk factors were tumor size ≤50 mm, and peritoneal dissemination.
Limitations: Because of the retrospective nature of this study, postoperative therapy was not standardized.
Conclusions: Risk factors differ for early, middle, and late recurrences of stage IV colorectal cancer following curative resection. Early (within 1 year) recurrence factors were rectum site, T4, N-staging, venous invasion, and liver metastasis, whereas late (2 years or later) recurrence risk factors were small tumor size and peritoneal dissemination. Our study provides important data to guide a surveillance protocol following stage IV colorectal cancer curative resection. See Video Abstract at http://links.lww.com/DCR/A460.