BACKGROUND: We demonstrated that circulating tumor cell (CTC) number at baseline and follow-up is an independent prognostic factor in metastatic colorectal cancer (mCRC). This analysis was undertaken to explore whether patient and treatment characteristics impact the prognostic value of CTCs. PATIENTS AND METHODS: CTCs were enumerated with immunomagnetic separation from the blood of 430 patients with mCRC at baseline and on therapy. Patients were stratified into unfavorable and favorable prognostic groups based on CTC levels of > or = 3 or <3 CTCs/7.5 ml, respectively. Subgroups were analyzed by line of treatment, liver involvement, receipt of oxaliplatin, irinotecan, or bevacizumab, age, and Eastern Cooperative Oncology Group performance status (ECOG PS). RESULTS: Seventy-one percent of deaths have occurred. Median follow-up for living patients is 25.8 months. For all patients, progression-free survival (PFS) and overall survival (OS) for unfavorable compared with favorable baseline CTCs is shorter (4.4 versus 7.8 m, P = 0.004 for PFS; 9.4 versus 20.6 m, P < 0.0001 for OS). In all patient subgroups, unfavorable baseline CTC was associated with inferior OS (P < 0.001). In patients receiving first- or second-line therapy (P = 0.003), irinotecan (P = 0.0001), having liver involvement (P = 0.002), >/=65 years (P = 0.0007), and ECOG PS of zero (P = 0.04), unfavorable baseline CTC was associated with inferior PFS. CONCLUSION: Baseline CTC count is an important prognostic factor within specific subgroups defined by treatment or patient characteristics.