**3.1. Sensitivity and specificity**

#### *3.1.1. Sensitivity*

It is demonstrated that the sensitivity of the method of CTCs detection by using multiparameter flow cytometry was 0.001%, or 10-5, according to the serial dilutions test, which is shown in **Figure 1a**–**d**. It was highly reproducible on recovery and linearity across three separate experiments (**Figure 1e**). The recovery of the tumor events was quite correlated with the tumor events expected based on the serial dilutions (*R*<sup>2</sup>  = 0.997). The recovery of the tumor cells was not significantly different from the tumor cells expected according to the serial dilutions (*P* >  0.6, Mann‐Whitney rank sum test).

#### *3.1.2. Specificity*

The detection of CTCs by multiparameter flow cytometry contributed a higher specificity compared with RT‐PCR. We detected the expression of Ep‐CAM in three typical advanced breast cancer (ABC) patient (CTCs ≥ 5) and three limited breast cancer (LBC) patients (CTCs < 5). The result showed that the expression of Ep‐CAM was positive for both patients with ABC and LBC as shown in **Figure 2**. It was hard to distinguish ABC with LBC by RT‐PCR, while the flow cytometry could distinguish them obviously and quantitatively.

**Figure 1.** The ability to detect human tumor cells SKBR‐3 cells in normal blood by cytomentry is titratable down to a sensitivity of 0.001%.Human tumor cells concentration was normalized adding to the leukocyte count, and serial dilu‐ tions (0.0001, 0.001%, 0.005%, 0.05%) using normal mononucleocytes as the diluent. Samples were lysed, incubated with anti‐CD45‐PerCP (20ul), EpCAM (20ul) and Cytokeratin8, 18, 19 (20ul) at 4°C for 30 minutes, and resuspended in 250ul PBS before multiparameter flow cytometric analysis. For samples with normal blood cells, up to 1000,000 total events were collected. Events that fell within up right region were counted as meeting the criteria for SKBR‐3 tumor cells (CD45‐EpCAM+CK+). Representative SKBR‐3 cells are shown for **(a)** 0.0001% **(b)** 0.001% **(c)** 0.005%, **(d)** 0.05%. **(e)** Correlation and regression analysis of recovered versus expected number of positive tumor events at dilutions signifi‐ cant at the 0.01 level (2‐tailed, R<sup>2</sup> =0.997, three separate experiments.). The percentage of tumor cells recovered was not significantly different from the percentage of tumor cells expected based on the serial dilutions (P>0.6, Mann‐Whitney rank sum test).

Detection of Circulating Tumor Cells and Circulating Tumor Stem Cells in Breast Cancer by Using Flow Cytometry http://dx.doi.org/10.5772/63423 201

**Figure 1a**–**d**. It was highly reproducible on recovery and linearity across three separate experiments (**Figure 1e**). The recovery of the tumor events was quite correlated with the tumor

not significantly different from the tumor cells expected according to the serial dilutions (*P* > 

The detection of CTCs by multiparameter flow cytometry contributed a higher specificity compared with RT‐PCR. We detected the expression of Ep‐CAM in three typical advanced breast cancer (ABC) patient (CTCs ≥ 5) and three limited breast cancer (LBC) patients (CTCs < 5). The result showed that the expression of Ep‐CAM was positive for both patients with ABC and LBC as shown in **Figure 2**. It was hard to distinguish ABC with LBC by RT‐PCR,

**Figure 1.** The ability to detect human tumor cells SKBR‐3 cells in normal blood by cytomentry is titratable down to a sensitivity of 0.001%.Human tumor cells concentration was normalized adding to the leukocyte count, and serial dilu‐ tions (0.0001, 0.001%, 0.005%, 0.05%) using normal mononucleocytes as the diluent. Samples were lysed, incubated with anti‐CD45‐PerCP (20ul), EpCAM (20ul) and Cytokeratin8, 18, 19 (20ul) at 4°C for 30 minutes, and resuspended in 250ul PBS before multiparameter flow cytometric analysis. For samples with normal blood cells, up to 1000,000 total events were collected. Events that fell within up right region were counted as meeting the criteria for SKBR‐3 tumor cells (CD45‐EpCAM+CK+). Representative SKBR‐3 cells are shown for **(a)** 0.0001% **(b)** 0.001% **(c)** 0.005%, **(d)** 0.05%. **(e)** Correlation and regression analysis of recovered versus expected number of positive tumor events at dilutions signifi‐ cant at the 0.01 level (2‐tailed, R<sup>2</sup> =0.997, three separate experiments.). The percentage of tumor cells recovered was not significantly different from the percentage of tumor cells expected based on the serial dilutions (P>0.6, Mann‐Whitney

while the flow cytometry could distinguish them obviously and quantitatively.

 = 0.997). The recovery of the tumor cells was

events expected based on the serial dilutions (*R*<sup>2</sup>

0.6, Mann‐Whitney rank sum test).

*3.1.2. Specificity*

200 Tumor Metastasis

rank sum test).

**Figure 2.** RT‐PCR assay for EpCAM mRNA **(a)** and GAPDH mRNA **(b)**. 1. DNA ladder, 2. negative control (H2O), 3. positive control (SKBR‐3 cells), 4.5.6. ABC samples, 7.8.9. LBC samples. All were positive. Size of EpCAM and GAPDH is 186bp and 177bp, respectively.


**Table 2.** Analysis of EPCAM and CD44 gene expression in patients with BC.

We also compared breast cancer (BC) patients with healthy volunteers by QRT‐PCR. The expression of Ep‐CAM was increased statistically higher in patients with BC (24.29 ± 44.10 vs. 0.14 ± 0.02 × 10-5, *P* = 0.000) than in healthy volunteers, which was shown in **Table 2** and it was calculated by 2-Δ*<sup>t</sup>* method. However, there were no obvious differences between BC and health on the expression of CD44. Therefore, we confirmed that Ep‐CAM and CD44 cannot be identified by QRT‐PCR but can be identified by multiparameter flow cytometry.

#### **3.2. Patient characteristics**

Forty‐five patients were identified and included in this analysis (detail is shown in **Table 1**); 27 (60.0%) patients had CTCs levels <5 and 18 (40.0%) patients had CTCs levels ≥5, respectively. The age of all the 45 patients ranged from 32 to 74, while the median age of CTCs levels <5 group was 49 years and the median age of CTCs ≥5 was 51 years. There were 25 (55.6%) patients in metastasis including 15 (33.3%) patients had CTCs ≥5. At the same time, there were 20 (44.4%) patients with no metastasis including 17 patients had CTCs <5. It showed that there was statistically significantly differences (*P* = 0.002) on CTCs level between the metastasis group and no metastasis group by chi‐square test analyses. And the statistical differences of CTCs level also exist in different TNM stage (*P* = 0.033).

**Figure 3.** Kaplan‐Meier Plots of overall survival are shown for all patients during the follow‐up. 27 patients in CTCs <5 group and the median survival is estimated to be 95 weeks. 18 patients in CTCs≥5 group and the median survival is 65.5 weeks. *P* value is 2‐tailed. Logrank indicates the *P* value, p=0.004.

#### **3.3. Survival analysis**

There were 17 (37.8%) patients who died during the follow‐up period including 11 (24.5%) patients in CTCs ≥5 group. The median survival among CTCs <5 group was 95 weeks (standard deviation, 18.67 weeks) and the median survival among CTCs ≥5 was 65.5 weeks (standard deviation, 30.0 weeks). Axillary lymph node dissection (ALND) had correlation with CTCs level (*P* = 0.143) and 45.7% ALND patients got CTCs ≥5. During the follow‐up, 11 (24.4%) patients got lost contact with seven (15.6%) in < 5 CTCs and four (8.9%) in ≥ five CTCs group. The results of the patients survival are shown in **Figure 3** by Kaplan‐Meier with logrank *P*=0.004 and Breslow *P*=0.003, which confirm that the survival of CTCs < 5 and CTCs ≥ 5 group were different statistically. And the overall survival (OS) of CTCs < 5 group was higher than CTCs ≥ 5 group.


**Table 3.** Cox regression analysis results.

We also compared breast cancer (BC) patients with healthy volunteers by QRT‐PCR. The expression of Ep‐CAM was increased statistically higher in patients with BC (24.29 ± 44.10 vs. 0.14 ± 0.02 × 10-5, *P* = 0.000) than in healthy volunteers, which was shown in **Table 2** and it was

on the expression of CD44. Therefore, we confirmed that Ep‐CAM and CD44 cannot be

Forty‐five patients were identified and included in this analysis (detail is shown in **Table 1**); 27 (60.0%) patients had CTCs levels <5 and 18 (40.0%) patients had CTCs levels ≥5, respectively. The age of all the 45 patients ranged from 32 to 74, while the median age of CTCs levels <5 group was 49 years and the median age of CTCs ≥5 was 51 years. There were 25 (55.6%) patients in metastasis including 15 (33.3%) patients had CTCs ≥5. At the same time, there were 20 (44.4%) patients with no metastasis including 17 patients had CTCs <5. It showed that there was statistically significantly differences (*P* = 0.002) on CTCs level between the metastasis group and no metastasis group by chi‐square test analyses. And the statistical differences of

**Figure 3.** Kaplan‐Meier Plots of overall survival are shown for all patients during the follow‐up. 27 patients in CTCs <5 group and the median survival is estimated to be 95 weeks. 18 patients in CTCs≥5 group and the median survival is

There were 17 (37.8%) patients who died during the follow‐up period including 11 (24.5%) patients in CTCs ≥5 group. The median survival among CTCs <5 group was 95 weeks (standard deviation, 18.67 weeks) and the median survival among CTCs ≥5 was 65.5 weeks (standard deviation, 30.0 weeks). Axillary lymph node dissection (ALND) had correlation with CTCs

identified by QRT‐PCR but can be identified by multiparameter flow cytometry.

CTCs level also exist in different TNM stage (*P* = 0.033).

65.5 weeks. *P* value is 2‐tailed. Logrank indicates the *P* value, p=0.004.

**3.3. Survival analysis**

method. However, there were no obvious differences between BC and health

calculated by 2-Δ*<sup>t</sup>*

202 Tumor Metastasis

**3.2. Patient characteristics**

In addition, we found that the prognosis of patients with breast cancer was statistical signifi‐ cant in CTCs level (*P* = 0.041), age (*P* = 0.001) and metastasis (*P* = 0.002) base on the Cox regression analysis for the follow‐up in 45 patients as shown in **Table 3**.

#### **3.4. Circulating tumor stem cells**

We also analyzed the expression of CTSCs and the characteristic of clinical data was shown in **Table 4**. There were 21 patients had at least one CTSCs expression among 45 patients with breast cancer. The CTSCs level in different TNM stages was statistically different (*P* = 0.020). It was obvious that stage III and IV patients contributed more CTSCs expression than stage I and II.



*CIS*, carcinoma *in situ.*

a There are missing values.

\* *P*<0.05.

**Table 4.** Patient characteristics in different CTC and CTSC levels.

The CTSC positive was also related with RLNM status (RLNM 0, 20.8%; RLNM 1–2, 66.7%; RLNM >3, 83.3%) and the *P*‐value was 0.001. There were no statistical differences in regard to age, diameter of tumor, clinical pathology, histology stage, ER status, PR status and Her‐2 status in different CTSC groups.

#### **3.5. Percentages of CD45‐cells, CTCs, CTSCs and their clinical relevance**

**Patient characteristics Number (%) CTC (%)** *P***‐value CTSC (%)** *P***‐value**

Tumor size 0.155 0.165

RLNM 0.075 0.0012\*

Clinical pathology 0.264 0.098

Histology stagea 0.002\* 0.919

ER 0.0182\* 0.482

PR 0.0052\* 0.771

Her‐2 0.405 0.262

I 4 (10.5) 3 (75.0) 1 (25.0) 2 (50.0) 2 (50.0) II 17 (44.7) 14 (82.4) 3 (17.6) 8 (47.1) 9 (52.9) III 17 (44.7) 4 (23.5) 13 (76.5) 7 (41.2) 10 (58.8)



1+ 14 (31.1) 10 (71.4) 4 (28.6) 10 (71.4) 4 (28.6) 2+ 18 (40.0) 11 (61.1) 7 (38.9) 8 (44.4) 10 (55.6) 3+ 13 (28.9) 6 (46.2) 7 (53.8) 6 (46.2) 7 (53.8)

*CIS*, carcinoma *in situ.*

204 Tumor Metastasis

There are missing values.

**Table 4.** Patient characteristics in different CTC and CTSC levels.

a

\* *P*<0.05.

Tis 4 (8.9) 4 (100.0) 0 (0.0) 4 (100.0) 0 (0.0) T1 10 (22.2) 7 (70.0) 3 (30.0) 7 (70.0) 4 (30.0) T2 20 (44.4) 12 (60.0) 8 (40.0) 8 (45.0) 11 (55.0) T3 5 (11.1) 1 (20.0) 4 (80.0) 2 (40.0) 3 (60.0) T4 6 (13.3) 3 (50.0) 3 (50.0) 2 (33.3) 4 (66.7)

0 24 (53.3) 16 (66.7) 8 (33.3) 19 (79.2) 5 (20.8) 1–3 9 (20.0) 7 (77.8) 2 (22.2) 3 (33.3) 6 (66.7) >3 12 (26.7) 4 (33.3) 8 (66.7) 2 (16.7) 10 (83.3)

CIS 4 (8.9) 4 (100.0) 0 (0.0) 4 (100.0) 0 (0.0) Infiltrating duct 36 (80.0) 21 (58.3) 15 (41.7) 16 (44.4) 20 (55.6) Mucous 2 (4.4) 1 (50.0) 1 (50.0) 2 (100.0) 0 (0.0) Infiltrating lobular 1 (2.2) 1 (100.0) 0 (0.0) 1 (100.0) 0 (0.0) Medullar 1 (2.2) 0 (0.0) 1 (100.0) 1 (100.0) 0 (0.0) Papillary 1 (2.2) 0 (0.0) 1 (100.0) 0 (0.0) 1 (100.0)

**<50 ≥50 Negative Positive**

The expression of CD45, CTCs and CTSCs was further explored in different BC groups (**Table 5**). The expression of CTSCs on CD45‐C showed a rising tendency in different TNM stage (*P* = 0.034) and also in different RLNM status (*P* = 0.001).



**Table 5.** Percentage of CD45 cells, CTC and CTSC, and their clinical relevance.

The percentages of CTCs on CD45‐C with TNM and histology stage increasing. In addition, we found that the percentage of CTCs on CD45‐C in ER and PR- groups was higher than that in ER<sup>+</sup> and PR<sup>+</sup> groups. And so was on CTSCs. Above all, we found the relationship between the percentage of CTC on CD45‐C and clinical pathology. Then, it was good for evaluate TNM stage and RLNM status according to the percentage of CTSC on CD45‐C.

Therefore, multiparameter flow cytometry technique is capable enough to identify patients with breast cancer and assess the progression of disease.
