Prognostic impact of circulating tumor cells in patients with ampullary cancer

Circulating tumor cells (CTCs) are an important topic of investigation for both basic and clinical cancer research. In this prospective study, we evaluated the clinical role of CTCs in ampullary cancer. We analyzed blood samples from 62 consecutively diagnosed patients with ampullary adenocarcinoma and 24 healthy controls for their CTC content. Combined data from immunostaining of CD45, 4′,6‐diamidino‐2‐phenylindole (DAPI), and fluorescence in situ hybridization with a chromosome 8 centromere (CEP8) probe were used to identify CTCs; cells that were CD45‐/DAPI+/CEP8>2 were considered CTCs. The Cox proportional hazards model was used to assess the relationship between CTCs, clinical characteristics, and patient outcomes. We detected ≥2 CTCs/3.2 ml whole blood in 43 of 62 patients (69.4%), as well as ≥5 CTCs/3.2 ml in 16 of these patients (25.8%). A CTC cutoff value of 2 cells/3.2 ml achieved 69.4% sensitivity and 95.8% specificity as a diagnostic tool; CTCs were associated with tumor burden. CTC levels ≥3/3.2 ml (hazard ratio [HR]: 2.5, 95% confidence interval [CI]: (1.2–5.2), p = 0.014) and ≥5/3.2 ml (HR: 3.5, 95% CI: 1.7–7.3, p < 0.001) were both associated with shorter disease‐free survival. Moreover, ≥3 CTCs/3.2 ml (HR: 2.7, 95% CI: 1.2–6.3, p = 0.019) and ≥5 CTCs/3.2 ml (HR: 3.8, 95% CI: 1.8–8.5, p < 0.001) were predictive of shorter overall survival. CTC assessment may help identify patients with ampullary cancer who are at high risk of an unfavorable outcome.

major efforts in recent decades, patients with ampullary cancer continue to experience poor prognosis and a high death rate. It is imperative to identify the potential biological markers for early diagnosis, novel therapeutic strategies, and prognostic predictors in patients with ampullary cancer.
Circulating tumor cells (CTCs) are minute numbers of tumor cells that escape from primary tumors and metastatic foci and enter the bloodstream (Cristofanilli, 2006). CTCs have been discovered in most major cancers, although at varying rates (Allard et al., 2004). Clinical studies have suggested that CTCs are correlated with poor progression-free survival (PFS) and overall survival (OS) in patients with breast cancer, colon cancer, and cholangiocarcinoma, as well as other malignancies (Cohen et al., 2008;Rack et al., 2014;Sonpavde & Antonarakis, 2017;Yang et al., 2016). These cells have also been utilized for guiding clinical management, evaluating curative efficacy, and monitoring tumor recurrence (Plaks, Koopman, & Werb, 2013).
However, the clinical significance of CTCs in ampullary cancer has not been explored as thoroughly as it has for other epithelial cancers.
As CTCs are scarce in the blood stream, a number of novel methodologies have been developed to improve the efficiency and precision of their detection. The CellSearch system is the only CTC detection technology approved by the Food and Drug Administration (Allard et al., 2004); the procedure relies on cytokeratin staining of cells in blood samples. However, the CellSearch system is limited in that epithelial cell adhesion molecules (EpCAMs) and cytokeratin are not expressed in all CTCs (Nagrath, Jack, Sahai, & Simeone, 2016).
Aneuploidy, a common manifestation of chromosome instability, is a hallmark of malignant solid tumors (Danielsen, Pradhan, & Novelli, 2016). As chromosome numbers are reflected by the chromosome 8 centromere (CEP8) that can be detected using fluorescence in situ hybridization (FISH), aneuploidy detection in peripheral blood provides a fresh avenue for CTC detection (Pecot et al., 2011). The feasibility of this approach has been demonstrated since the vast majority of circulating epithelial cells have been found to be aneuploid cells derived from primary tumors (Fehm et al., 2002). This method shows improved isolation efficiency in solid tumors compared to conventional cytokeratin-based methods (Ning et al., 2014).
In this study, we detected and characterized CTCs in 86 blood samples from 62 consecutive patients diagnosed with ampullary adenocarcinoma, as well as from 24 healthy control subjects using a method that combines immunostaining of CD45, 4′,6-diamidino-2phenylindole (DAPI), and CEP8-FISH. We further explored the relationships between CTCs and clinical parameters and outcomes.  (Edge & Compton, 2010). Treatments that patients received after enrollment were reviewed, and patients were followed until February 28, 2017. DFS was defined as the interval between the initial surgery and documented manifestation of recurrent disease (Eisenhauer et al., 2009). OS was calculated from the date of initial surgery to the date of death or last contact.

| Patients and clinical information
Follow-up data were available for all included patients. This study was conducted with the approval of the Ethics Committee of Fudan University-Affiliated Huashan Hospital in adherence with the principles of the Declaration of Helsinki.

| Sample collection and polyploidy detection
Whole peripheral blood samples were obtained via venipuncture from healthy subjects and from patients with ampullary cancer before their undergoing invasive tests and therapeutic measures. Tubes containing acid citrate dextrose-anticoagulant (Becton Dickinson, NJ) were used to obtain 3.2 ml whole blood samples for CTC detection. All blood samples were processed within 48 hr of collection.
Strategies for CTC isolation and detection were similar to those described previously (Ning et al., 2014). In brief, 4 ml of collected blood (including 0.8 ml of anticoagulant) was centrifuged to separate cells from whole blood. Next, red blood cells were dissolved using hypotonic hemolysis, and residual cell particles were resuspended in phosphate buffer solution. CTCs were cultured with anti-CD45 monoclonal antibody-coated magnetic beads (Life Technologies, Carlsbad, CA). Subsequently, the magnetic beads (which were mostly bound to leukocytes) were separated from the cell suspension by a magnetic stand (Promega, Madison, WI), and the leukocytes were smeared onto slides and fixed for identification. Specimens were dipped in 2× saline-sodium citrate (SSC) buffer at 37°C for 15 min, then dehydrated by sequential exposure to 75%, 85%, and 100% ethanol for 3 min each. Spectrum Orange-labeled CEP8 was added (Abbott Molecular Diagnostics, Des Plaines, IL). Specimens were denatured at 76°C for 5 min and hybridized at 37°C for 1.5 hr, followed by soaking in 50% formamide for 15 min and washing in 2× SSC twice for 5 min. Specimens were then dehydrated in a series of 75%, 85%, and 100% ethanol baths for 3 min each, incubated with Alexa Fluor 594-conjugated anti-human CD45 at room temperature for 1 hr, and washed twice with 0.2% bovine serum albumin. Following the addition of DAPI, slides were observed along the "S" track using a microscope (Nikon). CTCs were identified by their staining pattern of CD45-/DAPI+/CEP8>2. Theoretically, the size and strength of CEP8 signals in the nucleus of the same specimen should be consistent. Close hybridization signals in single cells were referred to as 1. Small-scale and weak hybridization signals may have been non-specific and were thus not considered.

| Patient characteristics
A total of 86 participants comprising 62 patients with ampullary cancer and 24 healthy controls were investigated in this study; their clinical characteristics and pathological information are shown in Table 1. Twenty-nine of the patients with ampullary cancer (46.8%) were female and 33 (53.2%) were male; their median age was 62 years (range, 34-78 years). Forty-three patients (69.4%) were diagnosed with well-or moderately differentiated disease, while the remaining 19 (30.6%) were diagnosed with poorly differentiated disease. As for pathological disease stage, 52 patients (83.9%) were stage I-II and 10 (15.2%) were stage III-IV. Twenty-three patients

| CTC identification
CD45, DAPI, and CEP8 staining data as obtained under fluorescence microscopy were combined to identify CTCs ( Figure 1). Immunostaining for CD45 was performed to discriminate blood cells from tumor cells. Cell nuclei were observed using DAPI, while CEP8 was used to count the chromosome 8 copy number. Cells with CD45 staining ( Figure 1a) and/or no more than 2 CEP8 signals ( Figure 1b) were not considered CTCs, while those that were CD45-/DAPI +/CEP>2 were considered to be CTCs. Those with 3 CEP8 signals were classified as triploid CTCs (Figure 1c), those with 4 CEP8 signals tetraploid CTCs (Figure 1d), and those with 5 or more CEP8 signals multiploid CTCs (Figure 1e).
We further constructed ROC curves to assess the utility of CTCs as a diagnostic tool for ampullary cancer (Figure 2b). The value of the area under the ROC curve reached 0.854 (95% confidence interval [CI] = 0.777-0.931, p < 0.001). Youden's index was then used to select the optimum cutoff value with maximum sensitivity and specificity; this value was 2 CTCs/3.2 ml of whole blood, yielding a sensitivity of 69.4%, and specificity of 95.8%. One out of 24 healthy controls (4.2%) was found to be 'CTC'-positive, indicating that our method can yield false positive results.

| CTCs and tumor characteristics
We analyzed the relationships between CTCs and clinicopathologic parameters using cutoffs of ≥2, ≥3, and ≥5 CTCs/3.2 ml. Neither ≥2 nor ≥3 CTCs/3.2 ml was associated with most clinicopathologic factors, including sex, age, tumor differentiation, tumor stage, T stage, lymph node invasion, and distant metastasis. However, ≥5 CTCs/3.2 ml was significantly associated with advanced TNM stage (p = 0.002) and tended to be associated with higher tumor size (p = 0.078), lymph node invasion (p = 0.106), and distant metastasis (p = 0.083). These results demonstrated that higher CTC levels appear to correlate with a larger tumor burden. CTCs were not associated with elevated carcinoembryonic antigen (CEA) levels. This is consistent with a previous study that found aneuploidy not to be associated with CEA levels (Laubert et al., 2013).

| DISCUSSION
In this study, CEP8, CD45, and DAPI were combined to detect CTCs in patients with ampullary cancer and healthy controls. Our method, which used a cutoff of 2 CTCs/3.2 ml whole blood, achieved a sensitivity of 69.4%, and specificity of 95.8%. CTC may serve as a potential prognostic factor for patients with ampullary cancer both in terms of DFS and OS.
The detection rate of CTC depends on the method. The CellSearch system relies on EpCAMs and cytokeratin for cell identification and enumeration, respectively. The CTC detection rates range from 11% to 57% in different tumors types (Allard et al., 2004). As EpCAM and cytokeratin are not expressed in a fair proportion of CTCs, this method has had to be continuously improved. Isolation by size of epithelial tumor cells (ISET) allows highly sensitive enrichment, morphological identification, and enumeration of CTCs (Vona et al., 2000). In one study based on the ISET method, CTCs were found in 18 of 29 patients patients with pancreatic cancer, and found that ISET detects a higher number of CTCs than the CellSearch System. Polymerase chain reaction (PCR)-based methods are also commonly used; CTCs were detected in 33.8-84% of patients by using Nested PCR with CK20 mRNA as a CTC marker in different studies (Soeth et al., 2005;Zhou et al., 2011). The method of combined immunostaining for CD45, DAPI, and CEP8-FISH has shown great potential for CTC detection; this method yielded sensitivities of 83.33% and 76.25% for detecting CTCs in lung and ovarian cancers, respectively (Ning et al., 2014). In our study, combined CD45, DAPI, and CEP8-FISH staining yielded a sensitivity of 69.4% and specificity of 95.8% for ampullary cancer CTC detection. This detection rate was marginally lower than those in lung and ovarian cancers, which may be related to the different biological characteristics of these tumors. To improve the detection rate, immunostaining for cytokeratin was used to detect CTCs; immunostaining of CK, CD45, DAPI, and CEP8/FISH yielded sensitivities and specificities of 68.2% and 94.9% in one pancreatic cancer study (Zhang et al., 2015) and 88% and 90% in another, respectively (Gao et al., 2016). Another technique to improve the detection rate was to capture circulating cells with aneuploidy based on both CEP8 and CEP7 staining. This method yielded a sensitivity of 84% and specificity of 97.6% in patients with lung cancer, which were slightly better than the sensitivity and specificity rates of 76% and 100%, respectively, yielded when using CEP8 alone (Chen & Xu, 2014).
The prognostic significance of CTC has been widely investigated in tumors. A prospective study of 302 patients with nonmetastatic breast cancer showed that ≥1 CTC/7.5 ml was a predictor of poorer PFS and OS (Lucci et al., 2012). A similar result was obtained in a large prospective trial using the CellSearch system in patients with breast cancer (Rack et al., 2014). Studies have also evaluated the prognostic significance of CTC in hepatocellular carcinoma (HCC); a study using the CellSearch system showed that ≥1 CTC per sample was associated with shorter OS on univariate analysis of 59 patients with HCC (p = 0.02) (Schulze et al., 2013). A landmark study of CTC using the CellSearch system showed that CTCs were independent predictors of OS in patients with cholangiocarcinoma (Yang et al., 2016). CTCs based on the immunostaining of CD45, DAPI, CK, and FISH-CEP8 have also been found to be associated with OS; a concentration of ≥3 CTCs/7.5 ml was an independent prognostic factor in patients with pancreatic cancer (Gao et al., 2016). CTCs identified based on FISH-CEP8 were associated with a curative effect following paclitaxel and cisplatin treatment in patients with advanced gastric cancer (Li et al., 2014). Patients with colon cancer who had CTCs detected by CK20 using real-time PCR had substantially worse OS rates compared to those without detectable CTCs (Soeth et al., 2005). At the same time, several groups have shown that CTCs are not a prognostic factor for OS in several cancers. There was no statistically significant association between CTCs detected using the CellSearch system and OS in 90 patients with locally advanced rectal cancer who underwent neoadjuvant chemoradiotherapy (Magni et al., 2014). CTCs detected by ISET Hence, further studies are required for additional validation in an independent validation cohort. As we did not draw blood from patients after treatment, we did not determine the impact of surgery and chemotherapy on CTCs. Overall, the relationship between CTC trends and prognosis requires further study.

| CONSLUSIONS
In summary, our data indicate a potential clinical value for CTCs based on their detection via combination CD45, DAPI, and CEP8-FISH staining. CTC levels were shown to be associated with tumor extent,

ETHICS APPROVAL AND CONSENT TO PARTICIPATE
This study was conducted with the approval of the Ethics Committee of Fudan University-Affiliated Huashan Hospital.

CONSENT FOR PUBLICATION
Written informed consent was obtained from all subjects.

AVAILABILITY OF DATA AND MATERIAL
The datasets analyzed during the current study are available from the corresponding author on reasonable request.

CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.