Background: Most circulating tumor cell (CTC) platforms rely on EpCAM for capture and cytokeratin (CK) for detection. However, an important population of cells that display an epithelial-mesenchymal transition (EMT) phenotype will be missed. We report a new strategy to efficiently isolate a more heterogeneous population of CTCs using an antibody cocktail. Methods: In the first prospective study, blood (20 mL) was collected from 23 patients diagnosed with various late stage metastatic/recurrent cancer (breast, CRC, lung, prostate) following IRB approval. PBMCs were incubated with either EpCAM alone or a mixture of 10 capture antibodies to target both epithelial and mesenchymal cells. CTCs were subsequently captured in the CEE channels and detected with cytokeratin (CK) and CD45. A second prospective IRB approved study involving 54 patients diagnosed with late stage metastatic/recurrent breast cancer was performed using similar detection strategies (CK cocktail mixture and anti-CD45) with the addition of HER2 FISH to determine amplification status among captured CK+/CD45- and CK-/CD45- cells (presumable EMT cells). Results: In the first study, overall detection of CK+ cells was 83% with EpCAM alone and 93% with antibody cocktail. In addition, a median of 0.4 CK+ cells/mL and 1.0 CK+ cells/mL was observed using EpCAM and antibody cocktail, respectively. In the second study, CK+/CD45- cells were detected in 43 of 54 cases (80%). Among the 43 cases in which CK+/CD45- cells were detected, high concordance (93%) in HER2 status between primary tumor and CTCs was observed with Her2 amplification noted in both CK+/CD45- (50%) and CK-/CD45- (50%) cells. Conclusions: We have developed a novel and robust method for CTC enumeration that utilizes a cocktail of antibodies for the detection of a heterogeneous population of CTCs in multiple cancer types. Our findings suggest an important population of CK- cells is being missed by current stain criteria. Data also demonstrate that recovery of CTCs from peripheral blood using the CEE platform is efficient and suitable for FISH-based testing.