The prognostic significance of circulating (CTCs) and disseminated tumor cells in patients with colorectal cancer (CRC) is controversial. We performed a meta-analysis of available studies to assess whether the detection of tumor cells in the blood and bone marrow (BM) of patients diagnosed with primary CRC can be used as a prognostic factor.

We searched the Medline, Biosis, Science Citation Index, and Embase databases and reference lists of relevant articles (including review articles) for studies that assessed the prognostic relevance of tumor cell detection in the peripheral blood (PB), mesenteric/portal blood (MPB), or BM of patients with CRC. Meta-analyses were performed using a random effects model, with hazard ratio (HR) and 95% confidence intervals (95% CIs) as effect measures.

A total of 36 studies, including 3094 patients, were eligible for final analyses. Pooled analyses that combined all sampling sites (PB, MPB, and BM) associated the detection of tumor cells with poor recurrence-free survival (RFS) (HR = 3.24 [95% CI: 2.06–5.10], n = 26, I² = 77%) and overall survival (OS) (2.28 [1.55–3.38], n = 21, I² = 66%). Stratification by sampling site showed that detection of tumor cells in the PB compartment was a statistically significant prognostic factor (RFS: 3.06 [1.74–5.38], n = 19, I² = 78%; OS: 2.70 [1.74–4.20], n = 16, I² = 59%) but not in the MPB (RFS: 4.12 [1.01–16.83], n = 8, I² = 75%; OS: 4.80 [0.81–28.32], n = 5, I² = 82%) or in the BM (RFS: 2.17 [0.94–5.03], n = 4, I² = 78%; OS: 1.50 [0.52–4.32], n = 3, I² = 84%).

Detection of CTCs in the PB indicates poor prognosis in patients with primary CRC.