**4. Discussion**

#### **4.1. The significances of CTCs in breast cancer**

The detection of CTCs by using multiparameter flow cytometry relied on the epithelial‐specific marker, which was expressed on epithelial cells but not on leukocytes [2, 27, 28]. Some of the CTCs that had higher metastatic potential may lose the expression of epithelial‐specific markers during the migration process [29–31]. We also target another epithelial‐specific marker—Ep‐CAM (epithelial‐cell adhesion molecule) to avoid possible false negative [32]. In order to detect the tumor cells that come from epithelium tissue, a monoclonal antibody directed against CD45 for negative selection of leukocytes [9–11]. Therefore, we targeted the dual‐positive cells (CD45‐EpCAM<sup>+</sup> CK+ ) as a surrogate marker for CTCs. And the serial dilution test was demonstrated to confirm the sensitivity of the assay by adding SKBR‐3 into healthy sample. At the same time, we also verified the higher specificity of multiparameter flow cytometry by comparing with RT‐PCR, although RT‐PCR had a higher sensitivity [10, 11, 33, 34]. However, the detection of CTCs by nucleic acid techniques may overestimate the sensi‐ tivity, which resulted from the membrane fragments or nucleic acid of markers because of the crack of tumor cells in circle. CTC detection should be performed on cell level. On the other hand, we came out the result that RT‐PCR could not identify the ABC patient (CTCs ≥ 5) from LBC patients (CTCs < 5) because of the high sensitivity of RT‐PCR technology. We chose the multiparameter flow cytometry as the way to detect CTCs under considering the balance of sensitivity and specificity. It was important to discard the first few milliliters of sampled blood to make sure that there was no false positive because of the epithelial cell fell off from the skin when punctured.

Some researchers had tried the immunomagnetic combining flow cytometry technique to detect CTCs, which had a higher false negative losing amount of target cells. This technique cost expensive and not brief enough to apply in clinical.

It was common to meet the problem in clinical that it was hard to distinguish the tumor from inflammation by imaging, while we found that there were no more than two CTCs in stage I. Budd et al. [35] proposed that it was more accurately to evaluate the development of disease by CTCs detection than imaging. Both tissue biopsy and marrow biopsy were traumatic, but CTCs detection by multiparameter flow cytomety was atraumatic that can be achieved in clinical. Above all, multiparameter flow cytometry was the appropriate technique for the detection of CTCs to monitor the progression of breast cancer.

We demonstrated that the retrospective study to confirm the CTCs detection by multipara‐ meter flow cytometry technique was a value method to apply in clinical. It suggested the patients who had CTCs ≥5 prompting a poorer median overall survival (65.5 weeks vs. 95 weeks; *P <* 0.05). We also found that the prognosis of the breast cancer was related with CTCs level, age and metastasis but not the clinical pathology and diameter of tumor. Cristofanilli et al. [36] also proposed that the patients with ≥5 CTCs/7.5 ml common had poorer prognosis compared with patients with <5 CTCs. CTCs detection by multiparameter flow cytometry should be a significant method for the evaluation of development and prognosis of cancer, and it also helped to estimate the treatment of target therapy for patients.
