Circulating Epithelial Cell Characterization and Correlation with Remission and Survival in Patients with Thyroid Cancer

Abstract

Background:

Thyroid cancer is the most common endocrine tumor and generally has relatively good clinical outcomes. However, 15–20% of patients ultimately develop recurrence or disease-related death. The appropriate prognostic factors for thyroid cancer are still elusive. This study evaluated whether the number of circulating tumor cells/circulating epithelial cells (CECs) expressing either epithelial cell adhesion molecule (EpCAM), podoplanin (PDPN), or thyrotropin receptor (TSHR) is related to remission and disease-specific mortality (DSM) of patients with thyroid cancer.

Methods:

Blood samples were collected from patients (n = 128) after thyroidectomy or radioactive iodide therapy. CECs were enriched by lysis of red blood cells and depletion of leukocytes. Subtyping and quantification of the enriched cells were performed with immunofluorescence staining using antibodies against EpCAM, TSHR, and PDPN, respectively. Whether the number of a specific subtype of CECs is related to remission and DSM of patients was determined by univariate and multivariate analyses.

Results:

The EpCAM⁺-CECs, TSHR⁺-CECs, and PDPN⁺-CECs counts for patients in the non-remission group (n = 43) were significantly higher when compared to the remission group (n = 85; p < 0.001). Receiver operating characteristic analysis showed that the number of EpCAM⁺-CECs, TSHR⁺-CECs, and PDPN⁺-CECs was able to distinguish the status of remission from non-remission. The cutoff point for EpCAM⁺-CECs, TSHR⁺-CECs, and PDPN⁺-CECs was 40, 47, and 14 (cells/mL), with the accuracy of the assay equivalent to 80.4%, 76.6%, and 77.3%, respectively. On the other hand, the number of EpCAM⁺-CECs (p < 0.001), PDPN⁺-CECs (p = 0.013), and TSHR⁺-CECs (p < 0.001) for patients in the DSM group (n = 17) was significantly higher when compared to the patients who survived (n = 111). Receiver operating characteristic analysis showed that EpCAM⁺-CECs, TSHR⁺-CECs, and PDPN⁺-CECs counts were able to distinguish mortality from survival status. The cutoff point for EpCAM⁺-CECs, TSHR⁺-CECs, and PDPN⁺-CECs was 27, 25, and 9 (cells/mL), with the accuracy of the assay equivalent to 69.5%, 67.2%, and 68.5%, respectively.

Conclusions:

CEC testing is a useful tool for analysis of overall survival and remission status of patients with thyroid cancer. Implementation of CEC testing into routine clinical test may be worthy to consider for patient clinical care.

Introduction

Thyroid cancer is the most common endocrine tumor and has an increasing incidence in both developed and developing countries (1 –3). Papillary and follicular thyroid cancer account for 80–90% of the cases and originate from well-differentiated thyroid follicular cells. On the other hand, medullary thyroid cancer of parafollicular cell origin and anaplastic or poorly differentiated thyroid cancer only account for <10% of all thyroid carcinomas (4). Most patients with well-differentiated thyroid cancer have a good prognosis. A 90% disease-free survival rate can be achieved after receiving appropriate treatment, including thyroidectomy and postoperative radioactive iodide (¹³¹I) therapy (5). However, recurrence occurs in 15–20% of patients with well-differentiated thyroid cancer during the follow-up period (6,7). The occurrence of distant metastasis and locoregional recurrences are poor prognostic signs (8). Early detection of disease recurrence is needed to increase patient survival.

Testing of serum level thyroglobulin (Tg) and medical imaging analysis are the two major modalities for monitoring disease recurrence in current clinical practice. Tg testing can be limited by the presence of anti-Tg antibodies (anti-TgAb) that interfere with the assay, while radioactive iodine is required for scintigraphic imaging analysis. Searching for new methods for monitoring disease status is warranted. The clinical use of liquid biopsies, including circulating tumor cells (CTCs)/circulating epithelial cells (CECs), circulating tumor DNA, and exosomes, in cancer diagnosis and prognosis has grown exponentially (9). In particular, the capacity of CTCs/CECs to provide biological information to assess the status of the patient underlines its potential as an independent biomarker to improve therapeutic outcomes (10 –12). Angiogenesis usually occurs in solid tumors when the tumor size is >2–3 mm³ (13,14). During angiogenesis, tumor cells of epithelial origin expressing epithelial cell adhesion molecule (EpCAM) may shed into the circulation to form CECs, which could serve as a surrogate marker for the angiogenic activity of a tumor and the degree of metastasis (15 –18). Recent studies from the authors' laboratory and others have revealed the potential applications of CEC testing in the early detection and disease monitoring of patients with thyroid cancer (11,19,20). CEC subtypes expressing different cellular markers have also been demonstrated. For example, CECs expressing podoplanin (PDPN) and thyrotropin receptor (TSHR) have been identified in the peripheral blood of patients with thyroid cancer (21). PDPN expression in cancer cells or cancer-associated fibroblasts is an independent unfavorable prognostic indicator in patients with various cancers (22). The TSHR is mainly expressed in thyroid cells, although cells from other tissues have been demonstrated to express TSHR (23). CECs with PDPN or TSHR expression have been shown to facilitate differentiating disease status of patients with papillary thyroid carcinoma (PTC).

CTCs/CECs characterization has also been shown to serve as a prognostic factor for various types of cancer, including lung, osteoblasts, and germ-cell tumors (22,24,25). However, little is known whether the CEC count in the bloodstream is related to the survival and remission status of patients with thyroid cancer. This issue was addressed in this long-term follow-up study. The potential applications of CEC testing in the clinical care of patients with thyroid cancer are discussed.

Methods

Study subjects

Patients were prospectively enrolled and cytopathologically diagnosed with thyroid cancer. After surgery, they underwent long-term follow-up and treatment at the Chang Gung Memorial Hospital in Linkou. This study was approved by the Institutional Review Board of Chang Gung Memorial Hospital (approval IDs 101-3497B, 102-4838C, 102-3433B, and 104-3901B). The enrollment criteria were: (i) aged >18 years, (ii) histopathologically diagnosed with thyroid cancer, (iii) able to comprehend fully and sign an informed consent form, and (iv) able to participate in a long-term follow-up study. Since 2015, 128 patients have been enrolled in the study, with a mean age of 46.1 ± 15.2 years (median 46 years), of whom 96 are women (Table 1).

Table 1.

Baseline Clinical and Pathological Characteristics of 128 Patients with Thyroid Cancer at the Time of Diagnosis

Clinical characteristics	Values
Mean age (years)	46.1 ± 15.2
Median (range)	46 (11–84)
<45	61 (47.7)
45–55	34 (26.6)
≥55	33 (25.8)
Sex (female)	96 (75.0)
Pathology
Papillary	119 (93.0)
Follicular	5 (3.9)
Hürthle cell	2 (1.6)
Poorly differentiated	2 (1.6)
Primary tumor size (cm)^a	2.5 ± 1.6
≤1	24 (20.5)
1–2	31 (26.5)
2.1–4	40 (34.2)
>4	22 (18.8)
Microscopic extrathyroidal extension	25 (19.5)
Gross extrathyroidal extension	22 (17.2)
Cervical lymph node metastasis	41 (32.0)
Level VI	23 (18.0)
Level I–V	18 (14.1)
Distant metastasis	22 (17.2)
Initial surgical extent
Total thyroidectomy	108 (84.4)
Less than total thyroidectomy	20 (15.6)
Accumulated RAI treatment dose (mCi)
≤30	21 (16.4)
30–100	36 (28.1)
100–200	21 (16.4)
>200	42 (32.8)
Pre-ablation Tg (ng/mL)	299.0 ± 1517.7
Median (range)	8.0 (0.0–14944.0)
≤10	67 (52.3)
>10	61 (47.7)
Follow-up (year)	7.8 ± 5.8
Median (range)	6.8 (0.1–28.8)
Follow-up after enrolled in study (year)	2.2 ± 1.0
Median (range)	2.4 (0.8–3.7)

Values are shown as the mean ± standard deviation or n (%).

Eleven cases were without tumor size.

RAI, radioactive iodine; Tg, thyroglobulin.

Patients underwent thyroidectomy at the authors' center after a diagnosis of thyroid cancer. All patients were staged in accordance with the Union for International Cancer Control tumor-node-metastasis (TNM) criteria (sixth edition) (26). All thyroid cancer tissues were pathologically classified according to the World Health Organization criteria (4). For cases with high or intermediate risk (27), ¹³¹I remnant ablation was performed. The dose of ¹³¹I ablation was 30–100 mCi (1.1–3.7 GBq) for most patients. In Taiwan, admission for isolation is required for patients receiving >30 mCi of ¹³¹I ablation. To balance the needs of patient care and the inconvenience of hospitalization for patients, post-thyroidectomy patients are subject to risk assessment for the aggressiveness of thyroid cancer to determine the administered activity of ¹³¹I. Patients who are considered high risk are typically treated with 100 mCi of ¹³¹I with admission for isolation, unless the patients refuse. On the other hand, patients who are considered as intermediate risk are recommended to receive 30 mCi for remnant ablation without admission for isolation. A post-therapy whole-body scan (WBS) was performed one week after ¹³¹I administration using a dual-head gamma camera (Siemens Medical Solutions USA, Inc., Malvern, PA), equipped with a high-energy collimator. The whole-body image was taken using continuous mode scanning at a speed of 5 cm/min. In addition, thyroid scintigraphy was performed using a pinhole collimator with a 4-mm aperture placed 7 cm above the neck for a total of 50 K counts or 30 min. Levothyroxine (LT4) treatment was initiated to decrease thyrotropin (TSH) levels without inducing clinical thyrotoxicosis. Cases in which the foci of ¹³¹I uptake extended beyond the thyroid bed were classified as either persistent disease or metastatic disease. Such patients received increased therapeutic activities of 100–200 mCi (3.7–7.4 GBq). WBS was performed two weeks after administering a higher therapeutic dose of ¹³¹I.

Neck ultrasonography was done 6–12 months after the initial thyroidectomy to exclude the possibility of local recurrence. Serum Tg was measured using a highly sensitive Tg access assay (Beckman Coulter, Brea, CA) when patients were on LT4 treatment and was used as an indicator for monitoring the disease status of patients. The blood samples used for CEC testing, serum Tg measurements, and other thyroid-related assays were collected simultaneously in one blood drawing. The serum levels of anti-TgAb were measured using a competitive radioimmunoassay (Biocode Hycel, Liege, Belgium) with an analytical sensitivity of 6 IU/mL. An anti-TgAb level <115 IU/mL was considered as not interfering with serum Tg testing. At the end of 2016, the enrolled patients were categorized into remission and non-remission groups. Non-remission is defined as biochemical and structural evidence for thyroid cancer, even after receiving multimodality treatments.

Cell culture and transfection

OECM-1 and HepG2 cells were cultured in Rosewell Park Memorial Institute 1640 medium supplemented with 10% fetal bovine serum (FBS). 293T cells were cultured in Dulbecco's modified eagle medium (DMEM) supplemented with 10% FBS and were transfected with plasmid DNA using the Lipofectamine 2000 (LF2000) reagent, as described previously (28,29). Briefly, 293T cells were plated at a density of 2 × 10⁶ cells/six-well culture dish. Twenty-four hours after cell plating, the cells were transfected with the pCMV-SPORT6 expression plasmids (6 μg) in the presence of 10 μL LF2000. The culture medium was replaced with fresh DMEM five hours post transfection. Forty-eight hours post infection, the cells were re-suspended by trypsinization for immunofluorescence staining.

Enrichment and isolation of CECs

A negative selection-based PowerMag system was used for CEC enrichment and isolation, as previously described (11,19,20). Fresh blood samples obtained from patients four to six weeks after thyroidectomy or radioactive iodide therapy (¹³¹I) were processed through lysis of red blood cells (RBC) and depletion of CD45⁺ leukocytes using a magnetic chamber. This method has been used to detect EpCAM-positive and EpCAM-negative CECs/CTCs from patients with various cancer types (11,19).

Immunofluorescence staining and CEC characterization

During immunofluorescence staining, the leukocyte-depleted cell filtrates were separated into two aliquots. One of the aliquots was incubated with an anti-TSHR antibody (Abcam, Cambridge, United Kingdom) and DNA staining Hoechst 33342 dye (Invitrogen, Carlsbad, CA) at room temperature for one hour. The other aliquot was incubated with an anti-EpCAM antibody (Abcam) and anti-PDPN antibody (Angiobio, San Diego, CA) in the presence of DNA staining Hoechst 33342 dye at room temperature for one hour. After several washes and centrifugation to remove the supernatants, the cell pellets were re-suspended, and the Alexa Fluor 488-conjugated donkey anti-mouse antibody (for TSHR and EpCAM) and Alexa Fluor 555-conjugated goat anti-rat antibody (for PDPN) were added to the cell suspension. The unbound antibody was removed after incubation in the dark for 30 minutes. The complete cell aliquot was placed on a slide, and the full immunofluorescence images were captured using fluorescence microscopy with an automated slide scanning platform (Zeiss Axiovert 200M) followed by image analysis using the IN Cell Analyzer 1000 Cellular Imaging and Analysis System (GE Healthcare Life Sciences, Pittsburgh, PA). EpCAM⁺-CECs, TSHR⁺-CECs, and PDPN⁺-CECs were the Hoechst-positive cells that were positive for EpCAM, TSHR, and PDPN, respectively.

Statistical analysis

Categorical data were compared using Pearson's chi-square test or Fisher's exact test. All data are expressed as the mean ± standard error of the mean. Univariate analysis and multivariate analysis by Cox proportional hazards regression model were performed to determine the significance of various factors (30). A p-value of <0.05 was considered statistically significant. Survival analysis was performed by using the Kaplan–Meier method and was compared using the Breslow and Mantel–Cox tests. Receiver operating characteristic (ROC) curve analysis was used to define the best cutoff value for remission and disease-specific mortality (DSM).

Results

Of the 128 enrolled cases, 67 (52.4%) patients were aged >44 years. A total of 119 (93%) patients were histologically diagnosed with PTC and another nine with follicular, Hürthle cell, and poorly differentiated thyroid carcinomas (Table 1). The mean tumor size was 2.5 ± 1.6 cm, including 35 (27.3%) cases with microcarcinomas (≤1.0 cm). A total of 108 patients underwent total thyroidectomy with or without lymph node dissection. Surgical pathological features illustrated 25 (19.5%) cases with microscopic extension to the perithyroidal tissue and 22 (17.2%) cases with gross extrathyroidal extension. After surgery and thyroid remnant ablation with ¹³¹I, 22 (17.2%) cases were diagnosed with distant metastases, and 43 (33.6%) cases were in non-remission status. A total of 42 (32.8%) cases received cumulative ¹³¹I activities >200 mCi. In addition, 61 (47.7%) cases had a postoperative Tg level of >10 ng/mL.

Experiments were performed to validate the antibodies that were used for identifying subpopulation of CEC expressing EpCAM, PDPN, or TSHR. First, immunofluorescence staining of white blood cells and cancer cell lines expressing EpCAM (OECM1 cells), PDPN (OECM1 cells), and TSHR (HepG2 cells) were performed by using the respective antibodies (Fig. 1A–D). Second, 293T cells were transfected with control vector or with the expression plasmids encoding EpCAM, PDPN, and TSHR followed by immunofluorescence staining using the respective antibodies (Fig. 2A–C). Both sets of data indicate that the antibodies used in this study specifically recognize EpCAM, PDPN, and TSHR.

FIG. 1.

Immunofluorescence staining of leukocytes and cancer cell lines expressing epithelial cell adhesion molecule (EpCAM), podoplanin (PDPN), and thyrotropin receptor (TSHR). (A–C) Leukocytes and OECM1 and HepG2 cells were subject to immunofluorescence staining for CD45 (A), EpCAM (B), PDPN (C), and TSHR (D) using respective antibodies. Positive staining with Hoechst 33342 (A–D) indicates the presence of intact nucleated cells.

FIG. 2.

Immunofluorescence staining of HEK293T cells expressing EpCAM, PDPN, and TSHR. (A–C) HEK293T cells were transfected with the control vector (pCMV-SPORT6) (A–C) and the expression plasmids encoding EpCAM (A), PDPN (B), and TSHR (C) by using LF2000 reagent. Forty-eight hours post transfection, the cells were analyzed by immunofluorescence staining using the respective antibodies. Positive staining with Hoechst 33342 (A–C) indicates the presence of intact nucleated cells.

Fresh blood samples from patients four to six weeks after thyroidectomy or radioactive iodide therapy (¹³¹I) were processed through lysis of RBC and depletion of CD45⁺ leukocytes using the PowerMag system. The enriched cells were analyzed through immunofluorescence staining using the anti-EpCAM, PDPN, and TSHR antibodies (31,32). Hoechst 33342 was used to identify nucleated cells. Fluorescence microscopy analysis was performed to identify the CECs expressing EpCAM, PDPN and TSHR. Representative CECs positive for EpCAM, PDPN, or TSHR are shown in Figure 3A–C.

FIG. 3.

Isolation and characterization of circulating epithelial cells (CECs) from patients with thyroid cancer. (A–C) CECs were isolated using the PowerMag system and analyzed using immunofluorescence staining, as described in the Methods. At least three CEC populations positive for EpCAM (A), PDPN (B), and TSHR (C) were defined. Positive staining with Hoechst 33342 (A–C) indicates the presence of intact nucleated cells.

At the end of follow-up, patients were categorized into non-remission and remission groups (Table 2). A total of 43 cases comprise the non-remission group, including 17 cases with DSM. Prognostic factor analysis revealed that age, tumor size, tissue with gross extrathyroidal extension, distant metastasis, TNM stages I and IV, Tg level, and CEC count (EpCAM⁺, TSHR⁺, and PDPN⁺) were statistically different between the remission and non-remission groups. In addition, DSM analysis demonstrated that age, tumor size, tissue with gross extrathyroidal extension, distant metastasis, TNM stages I and IV, and CEC count (EpCAM⁺, PDPN⁺, and TSHR⁺) were statistically different between the DSM and survival groups (Table 3). Multivariate analysis illustrated that age, sex, EpCAM⁺-CECs, PDPN⁺-CECs, and TSHR⁺-CECs were statistically different between the survival and DSM groups (Table 4). Furthermore, analysis was performed of patients who had papillary thyroid microcarcinomas (n = 24). No statistical difference was found when the number of CECs was compared between the non-remission (n = 2) and the remission (n = 22) group, and the DSM (n = 2) and the survival (n = 22) group for patients with papillary thyroid microcarcinomas (data not shown). It is likely that the total number of patients with papillary thyroid microcarcinomas is not sufficient for analysis. In particular, only two patients were in the non-remission and DSM groups. Further studies focusing on recruiting more patients with PTC <1 cm is required to address this important issue.

Table 2.

Clinical Features of Patients with Thyroid Cancer in Remission or Non-Remission at Final Follow-Up Status

Clinical characteristic	Non-remission (n = 43)	Remission (n = 85)	p-Value
Age (years)	53.9 ± 18.1	42.2 ± 11.7	<0.001
Median (range)	54 (11–84)	42 (18–78)
<45	14 (32.6)	47 (55.3)	0.001
45–55	9 (20.9)	25 (29.4)
>55	20 (46.5)	13 (15.3)
Sex, female	28 (65.1)	68 (80.0)	0.066
Pathology
Papillary	41 (95.3)	78 (91.8)	0.454
Follicular	—	5 (5.9)	0.105
Hürthle cell	—	2 (2.4)	0.311
Poorly differentiated	2 (4.7)	—	0.045
Primary tumor size (cm)^a	3.2 ± 1.6	2.1 ± 1.4	<0.001
≤1	2 (5.3)	22 (27.8)	<0.005
1–2	7 (15.8)	24 (31.6)
2–4	17 (44.7)	23 (29.1)
>4	12 (34.2)	10 (11.4)
Microscopic extrathyroidal extension	12 (27.9)	13 (15.3)	0.089
Gross extrathyroidal extension	14 (32.6)	8 (9.4)	0.001
Multifocality	15 (34.9)	21 (24.7)	0.226
Cervical LN metastasis	19 (44.2)	22 (25.9)	0.036
Level VI	10 (23.3)	13 (15.3)	0.268
Level I-V	9 (20.9)	9 (10.6)	0.112
Distant metastasis	13 (30.2)	9 (10.6)	0.005
Initial surgical extent
Total thyroidectomy	38 (88.4)	70 (82.4)	0.376
Less than total thyroidectomy	5 (11.6)	15 (17.6)
TNM stage
Stage I	10 (23.3)	58 (68.2)	<0.001
Stage II	6 (14.0)	13 (15.3)	0.840
Stage III	3 (7.0)	7 (8.2)	0.802
Stage IV	24 (55.8)	7 (8.2)	<0.001
RAI treatment (mCi)^b	492.7 ± 401.0	139.8 ± 179.6	<0.001
≤30	1 (2.3)	20 (23.5)	<0.001
30–100	3 (7.0)	33 (38.8)
100–200	8 (18.6)	13 (15.3)
>200	28 (65.1)	14 (16.5)
Pre-ablation Tg (ng/mL)	827.7 ± 2556.4	37.8 ± 80.9	0.006
Median (range)	100.0 (0.0–14,944.0)	3.9 (0.0–470.0)
≤10	12 (27.9)	55 (64.7)	<0.001
>10	31 (72.1)	30 (35.3)
Follow-up (year)	7.9 ± 6.3	7.8 ± 5.5	0.901
Median (range)	7.7 (0.1–24.6)	6.4 (0.1–28.8)
Follow-up (year)	2.4 ± 0.9	2.1 ± 1.0	0.080
Median (range)	2.5 (0.8–3.7)	2.4 (0.8–3.6)
Second primary cancer	1 (2.3)	4 (4.7)	0.511
Disease specific mortality (n)	17 (39.5)	—	<0.001
CEC test
EpCAM⁺-CECs (cells/mL)	87 ± 83	19 ± 20	<0.001
PDPN⁺-CECs (cells/mL)	36 ± 48	8 ± 11	<0.001
TSHR⁺-CECs (cells/mL)	101 ± 178	25 ± 54	<0.001

Values are shown as the mean ± standard deviation or n (%).

Eleven cases were without tumor size.

Postoperative ¹³¹I accumulative dose.

LN, lymph node; TNM, tumor-node-metastasis; CEC, circulating epithelial cell; EpCAM, epithelial cell adhesion molecule; PDPN, podoplanin; TSHR, thyrotropin receptor.

Table 3.

Clinical Features of Patients with Thyroid Cancer in the Status of Mortality or Survival

Clinical characteristic	Mortality (n = 17)	Survival (n = 111)	p-Value
Age (years)	66.8 ± 12.2	43.0 ± 13.0	<0.001
Median (range)	66 (41–84)	43 (11–82)
<45	2 (11.8)	59 (53.2)	<0.001
45–55	1 (5.9)	33 (29.7)
≥55	14 (82.4)	19 (17.1)
Sex, female	11 (64.7)	85 (76.6)	0.293
Pathology
Papillary	15 (88.2)	104 (93.7)	0.412
Follicular	—	5 (4.5)	0.372
Hürthle cell	—	2 (1.8)	0.577
Poorly differentiated	2 (11.8)	—	<0.001
Primary tumor size (cm)^a	3.4 ± 1.8	2.3 ± 1.5	0.013
≤1	2 (13.3)	22 (21.6)	0.032
1–2	—	31 (30.4)
2–4	8 (53.3)	32 (31.4)
>4	5 (33.3)	17 (13.7)
Microscopic extrathyroidal extension	5 (29.4)	20 (18.0)	0.270
Gross extrathyroidal extension	7 (41.2)	15 (13.5)	0.005
Multifocality	7 (41.2)	29 (26.1)	0.199
Cervical LN metastasis	6 (35.3)	35 (31.5)	0.757
Level VI	1 (5.9)	22 (19.8)	0.163
Level I–V	5 (29.4)	13 (11.7)	0.051
Distant metastasis	6 (35.3)	16 (14.4)	0.034
Initial surgical extent
Total thyroidectomy	12 (70.6)	96 (86.5)	0.093
Less than total thyroidectomy	5 (29.4)	15 (13.5)
TNM stage
Stage I	—	68 (61.3)	<0.001
Stage II	3 (17.6)	16 (14.4)	0.727
Stage III	2 (11.8)	8 (7.2)	0.514
Stage IV	12 (70.6)	19 (17.1)	<0.001
RAI treatment (mCi)b	550.8 ± 496.8	215.5 ± 261.1	<0.001
≤30	1 (5.9)	20 (18.0)	0.009
30–100	1 (5.9)	35 (31.5)
100–200	2 (11.8)	19 (17.1)
>200	11 (64.7)	31 (27.9)
Pre-ablation Tg (ng/mL)	243.9 ± 265.3	307.1 ± 1622.9	0.878
Median (range)	137.5 (1.4–1023.0)	4.8 (0.0–14,944.0)
≤10	4 (23.5)	63 (56.8)	0.011
>10	13 (76.5)	48 (43.2)
Follow-up (year)	2.6 ± 0.8	2.2 ± 1.0	0.083
Median (range)	2.8 (1.0–3.7)	2.4 (0.8–3.7)
Second primary cancer	1 (5.9)	4 (3.6)	0.652
CEC test
EpCAM⁺-CECs (cells/mL)	96 ± 80	33 ± 51	<0.001
PDPN⁺-CECs (cells/mL)	35 ± 48	15 ± 27	0.013
TSHR⁺-CECs (cells/mL)	145 ± 255	37 ± 65	<0.001

Values are shown as the mean ± standard deviation or n (%).

Eleven cases were without tumor size.

Postoperative ¹³¹I accumulative dose.

Table 4.

Multivariate Analysis by Cox Proportional Hazards Regression Model for Survival and Mortality Analysis

	β coefficient	Hazard ratio	p-Value	Confidence interval
	β coefficient	Hazard ratio	p-Value	Lower	Upper
Age	0.153	1.165	0.0003	1.073	1.265
Sex	3.675	39.437	0.0014	4.146	375.1
Tumor size	0.048	1.049	0.8204	0.695	1.582
EpCAM⁺-CECs (cells/mL)	0.022	1.022	0.0018	1.008	1.036
PDPN⁺-CECs (cells/mL)	–0.022	0.979	0.0234	0.961	0.997
TSHR⁺-CECs (cells/mL)	0.011	1.011	0.0317	1.001	1.022

Dependent variable: survival vs. mortality.

ROC analysis showed that the number of EpCAM⁺-CECs, TSHR⁺-CECs, and PDPN⁺-CECs distinguished the remission from non-remission status with an area under the curve (AUC) equivalent to 0.782 (p < 0.0001), 0.688 (p = 0.0005), and 0.761 (p < 0.0001; Fig. 4A). The cutoff point for EpCAM⁺-CECs, TSHR⁺-CECs, and PDPN⁺-CECs was 40, 47, and 14 (cell/mL), with the accuracy of the assay equivalent to 80.4%, 76.6%, and 77.3%, respectively. In addition, ROC analysis showed that the number of EpCAM⁺-CECs, TSHR⁺-CECs, and PDPN⁺-CECs distinguished mortality from survival status with an AUC equivalent to 0.760 (p = 0.0006), 0.733 (p = 0.0020), and 0.748 (p = 0.0010), respectively (Fig. 4B). The cutoff point for EpCAM⁺-CECs, TSHR⁺-CECs, and PDPN⁺-CECs was 27, 25, and 9 (cell/mL), with the accuracy of the assay equivalent to 69.5%, 67.2%, and 68.5%, respectively. Analysis using cutoff points showed that the number of CECs with the three markers in the DSM group is statistically higher than when compared to the survival group (p < 0.0001; Fig. 5A–C).

FIG. 4.

Receiver operating characteristic (ROC) analyses of CEC counts from patients with thyroid cancer. (A and B) ROC analyses of EpCAM⁺-CEC, TSHR⁺-CEC, and PDPN⁺-CEC counts from thyroid cancer patients comparing remission and non-remission groups (A), and survival and DSM groups (B). The parameters used to perform the ROC analysis are also shown.

FIG. 5.

Kaplan–Meier survival curves of patients with thyroid cancer. (A–C) Thyroid cancer patients were divided into two groups according to the cutoff points of EpCAM⁺-CEC, TSHR⁺-CEC, and PDPN⁺-CEC as defined by ROC curve analysis. The analysis of overall survival was performed by using the Kaplan–Meier method, and the survival curves were plotted.

Discussion

The presence of CECs in the bloodstream is considered a prognostic factor for poor survival in different cancer types (33 –35). Patients with well-differentiated thyroid cancer tend to have excellent prognosis after treatment. However, distant metastases may occur and threaten the life of patients (6). To date, the correlation of the number of CECs with overall survival of patients with thyroid cancer is still unknown. This study reports for the first time that the number of EpCAM⁺-CECs, TSHR⁺-CECs, and PDPN⁺-CECs is statistically higher in the non-remission group than in the remission group. The CEC count is also higher in the DSM group than in the survival group. This study highlights that the CEC count may be a prognostic factor for predicting the overall survival of patients with thyroid cancer.

CEC characterization has been shown to facilitate the identification of a patient with early-stage PTC and follow-up of disease status post thyroidectomy (19), define and monitor the therapeutic outcome of patients with PTC (20), and supplement serum Tg testing and medical imaging analysis for distinguishing disease status of patients with PTC (11). This study continues the authors' long-term effort and explores the clinical applications of CEC characterization in the management of patients with thyroid cancer. It is demonstrated that the overall survival of patients with thyroid cancer is closely related to the levels of CECs in the bloodstream.

At present, the prognostic relevance of CECs/CTCs in cancer patients is still controversial. A large cohort study of patients with breast cancer at the time of primary diagnosis and from the prospectively randomized multicenter SUCCESS-A trial (EudraCT2005000490-21) did not demonstrate prognostic values of CTCs, as quantified by an immunocytochemistry method (36). On the other hand, a meta-analysis of patients with triple-negative breast cancer provided strong evidence that CTCs in the peripheral blood is an independent prognostic factor of poor survival outcomes (37). Using the CellSearch system for CTC enrichment and characterization of blood from 18 patients with distantly metastatic medullary thyroid cancer indicated that the presence of five or more CTCs is associated with worse overall survival of patients (18). However, the conclusion is limited by the number of patients enrolled in the study. By analyzing a large cohort of 128 patients with thyroid cancer, the current data provide strong evidence that in addition to age and sex, the number of CECs were statistically different between the remission group and the non-remission group. In addition, patients with a high CEC count above a specific cutoff value usually have a significant shorter overall survival, regardless of whether EpCAM, PDPN, or TSHR is used as the CEC marker. Similar results were observed when CECs with simultaneous expression of EpCAM and PDPN (i.e., EpCAM⁺PDPN⁺-CECs) were analyzed (data not shown). CEC testing thereby could be considered as a potential clinical test for monitoring treatment response and prognosis of patients with thyroid cancer.

Different subtypes of CECs are present in the bloodstream. In addition to the common epithelial cell marker EpCAM, CECs expressing PDPN and TSHR were identified in the peripheral blood of patients with thyroid cancer. PDPN is a 38 kDa mucin-type I transmembrane glycoprotein known as a marker of lymphatic endothelial cells (38). In a recent study, PDPN expression was evaluated in patients with PTC and thyroid cancer cell lines derived from papillary (TPC1 and BcPAP) and follicular (FTC133 and CGTH-W-1) thyroid carcinomas using reverse transcription polymerase chain reaction, Western blotting, immunofluorescence staining, and immunohistochemistry (21). Results of this study demonstrated that PDPN mediates the invasiveness of cells derived from PTC, consistent with the current data demonstrating that an increase in the number of CECs expressing PDPN is a prognostic factor for poor survival of patients with thyroid cancer. On the other hand, thyroid follicular cells are the major cell types to express TSHR, although TSHR can also be found in nonthyroidal tissues (23). Although both PDPN and TSHR are not only expressed in thyroid cells, there is a minimal impact on their use as prognostic markers for patients who are diagnosed with thyroid cancer. These markers could also have prognostic value by supplementing other clinical factors or markers such as Tg in clinical surveillance of disease status.

This large prospective study was designed to perform CEC testing after thyroidectomy with or without radioactive iodine therapy (¹³¹I). The procedure of thyroidectomy may trigger the release of cancer cells into the peripheral blood (12,39) and cause an increase in the number of CECs. Hence, the time interval between surgery and blood collection for CEC testing must be taken into consideration. In this study, blood sampling was usually performed four to six weeks after surgical intervention in order to minimize its potential effect on the results of CEC testing. This time interval is also appropriate for Tg testing, and both tests can be performed simultaneously with a single blood draw. Combined analysis of CECs and Tg testing may assist in the clinical management of patients with thyroid cancer.

In conclusion, this study, together with previous reports (11,20), demonstrates that CEC testing is a useful tool for detecting residual disease, monitoring treatment response, and predicting overall survival of patients with thyroid cancer. The value of CEC testing needs to be corroborated in larger prospective studies in order to assess its value as a clinical tool that may complement Tg measurements combined with imaging studies.

Footnotes

Acknowledgments

This work was supported by the Chang Gung Memorial Hospital (grant numbers CMRPG3E1901, CMRPD1C0551-3, CMRPD1E0181-3, CMRPD1F0611-2, and BMRP466); the Ministry of Science and Technology (grant numbers 102-2628-B-182-009-MY3, 103-2314-B-182-018-MY3, 105-2320-B-182-030, 105-2320-B-182-029-MY3, 106-2320-B-182-027-MY3, and 106-2314-B-182-042); and the Chang Gung Molecular Medicine Research Center (grant number EMRPD1F0121) to J.-D.L. and C.-P.T.

Author Disclosure Statement

No competing financial interests exist.

References

Fidler

, Soerjomataram

, Bray

. 2016. A global view on cancer incidence and national levels of the human development index. Int J Cancer, 139:2436–2446.

Chen

, Jemal

, Ward

. 2009. Increasing incidence of differentiated thyroid cancer in the United States, 1988–2005. Cancer, 115:3801–3807.

Ceresini

, Corcione

, Michiara

, Sgargi

, Teresi

, Gilli

, Usberti

, Silini

, Ceda

. 2012. Thyroid cancer incidence by histological type and related variants in a mildly iodine-deficient area of Northern Italy, 1998 to 2009. Cancer, 118:5473–5480.

DeLellis

, Lloyd

, Heitx

, Eng

(eds) 2004. Pathology and Genetics of Tumors of Endocrine Organs. World Health Organization Classification of Tumours. IARC Press, Lyon, France, pp 73–76.

Dal Maso

, Tavilla

, Pacini

, Serraino

, van Dijk

BAC

, Chirlaque

, Capocaccia

, Larrañaga

, Colonna

, Agius

, Ardanaz

, Rubió-Casadevall

, Kowalska

, Virdone

, Mallone

, Amash

, De Angelis

; EUROCARE-5 Working Group. 2017. Survival of 86,690 patients with thyroid cancer: a population-based study in 29 European countries from EUROCARE-5. Eur J Cancer, 77:140–152.

Lin

, Hsueh

, Chao

. 2015. Long-term follow-up of the therapeutic outcomes for papillary thyroid carcinoma with distant metastasis. Medicine, 94:e1063.

Wang

, Cheung

, Farrokhyar

, Roman

, Sosa

. 2013. A meta-analysis of the effect of prophylactic central compartment neck dissection on locoregional recurrence rates in patients with papillary thyroid cancer. Ann Surg Oncol, 20:3477–3483.

Ito

, Masuoka

, Fukushima

, Inoue

, Kihara

, Tomoda

, Higashiyama

, Takamura

, Kobayashi

, Miya

, Miyauchi

. 2010. Prognosis and prognostic factors of patients with papillary carcinoma showing distant metastasis at surgery (M1 patients) in Japan. Endocr J, 57:523–531.

Cree

, Uttley

, Buckley Woods

, Kikuchi

, Reiman

, Harnan

, Whiteman

, Philips

, Messenger

, Cox

, Teare

, Sheils

, Shaw

; UK Early Cancer Detection Consortium. 2017. The evidence base for circulating tumour DNA blood-based biomarkers for the early detection of cancer: a systematic mapping review. BMC Cancer, 17:697.

10.

Sorg

, Pachmann

, Brede-Hekimian

, Freesmeyer

, Winkens

. 2015. Determining tissue origin of circulating epithelial cells (CEC) in patients with differentiated thyroid cancer by real-time PCR using thyroid mRNA probes. Cancer Lett, 356:491–495.

11.

Lin

, Liou

, Hsu

, Hsieh

, Chen

, Tseng

, Lin

. 2016. Combined analysis of circulating epithelial cells and serum thyroglobulin for distinguishing disease status of the patients with papillary thyroid carcinoma. Oncotarget, 7:17242–17253.

12.

Winkens

, Pachmann

, Freesmeyer

. 2014. The influence of radioiodine therapy on the number of circulating epithelial cells (CEC) in patients with differentiated thyroid carcinoma—a pilot study. Exp Clin Endocrinol Diabetes, 122:246–253.

13.

Hanahan

, Weinberg

. 2000. The hallmarks of cancer. Cell, 100:57–70.

14.

Hanahan

, Weinberg

. 2011. Hallmarks of cancer: the next generation. Cell, 144:646–674.

15.

Rahbari

, Aigner

, Thorlund

, Mollberg

, Motschall

, Jensen

, Diener

, Büchler

, Koch

, Weitz

. 2010. Meta-analysis shows that detection of circulating tumor cells indicates poor prognosis in patients with colorectal cancer. Gastroenterology, 138:1714–1726.

16.

Najjar

, Alammar

, Al-Massarani

, Almalla

, Aljapawe

, Ikhtiar

. 2017. Circulating endothelial cells and microparticles for prediction of tumor progression and outcomes in advanced non-small cell lung cancer. Cancer Biomark, 20:333–343.

17.

Sun

, Liao

, Deng

, Li

, Feng

, Lei

, Yuan

, Zou

, Mao

, Shao

, Wu

, Zhang

. 2017. The diagnostic value of assays for circulating tumor cells in hepatocellular carcinoma: a meta-analysis. Medicine, 96:e7513.

18.

, Handy

, Michaelis

, Waguespack

, Hu

, Busaidy

, Jimenez

, Cabanillas

, Fritsche

Jr , Cote

, Sherman

. 2016. Detection and prognostic significance of circulating tumor cells in patients with metastatic thyroid cancer. J Clin Endocrinol Metab, 101:4461–4467.

19.

Hsieh

, Lin

, Hsueh

, Hsu

, Wang

, Wu

, Tseng

, Lin

. 2016. Circulating epithelial cell enumeration facilitates the identification and follow-up of a patient with early stage papillary thyroid microcarcinoma: a case report. Clin Chim Acta, 454:107–111.

20.

, Tseng

, Hsu

, Lin

, Chen

, Liou

, Lin

. 2018. Circulating epithelial cells as potential biomarkers for detection of recurrence in patients of papillary thyroid carcinoma with positive serum anti-thyroglobulin antibody. Clin Chim Acta, 477:74–80.

21.

Rudzińska

, Gaweł

, Sikorska

, Karpińska

, Kiedrowski

, Stępień

, Marchlewska

, Czarnocka

. 2014. The role of podoplanin in the biology of differentiated thyroid cancers. PLoS One, 9:e96541.

22.

Yurugi

, Wakahara

, Matsuoka

, Sakabe

, Kubouchi

, Haruki

, Nosaka

, Miwa

, Araki

, Taniguchi

, Shiomi

, Nakamura

, Umekita

. 2017. Podoplanin expression in cancer-associated fibroblasts predicts poor prognosis in patients with squamous cell carcinoma of the lung. Anticancer Res, 37:207–213.

23.

Williams

. 2011. Extrathyroidal expression of TSH receptor. Ann Endocrinol, 72:68–73.

24.

Wetterwald

, Hofstetter

, Cecchini

, Lanske

, Wagner

, Fleish

, Atkinson

. 1996. Characterization and cloning of the E11 antigen, a marker expressed by rat osteoblasts and osteocytes. Bone, 18:125–132.

25.

Schacht

, Dadras

, Johnson

, Jackson

, Hong

, Detmar

. 2005. Up-regulation of the lymphatic marker podoplanin, a mucin-type transmembrane glycoprotein, in human squamous cell carcinomas and germ cell tumors. Am J Pathol, 166:913–921.

26.

Sobin

, Wittekind

(eds) 2002. TNM Classification of Malignant Tumors. Sixth edition. Wiley-Liss, New York, NY, pp 52–56.

27.

Cooper

, Doherty

, Haugen

, Kloos

, Lee

, Mandel

, Mazzaferri

, McIver

, Pacini

, Schlumberger

, Sherman

, Steward

, Tuttle

. 2009. Revised American Thyroid Association management guidelines for patients with thyroid nodules and differentiated thyroid cancer. American Thyroid Association (ATA) guidelines task force on thyroid nodules and differentiated thyroid cancer. Thyroid, 19:1167–1214.

28.

Tsai

, Chien

, Liao

, Shih

, Lin

, Chang

, Cheng

, Tseng

. 2017. Functional links between Disabled-2 Ser723 phosphorylation and thrombin signaling in human platelets. J Thromb Haemost, 15:2029–2044.

29.

Hung

, Ling

, Cheng

, Chang

, Tseng

. 2016. Disabled-2 is a negative immune regulator of lipopolysaccharide-stimulated Toll-like receptor 4 internalization and signaling. Sci Rep, 6:35343.

30.

Zhang

, Zhou

, Freeman

Jr , Freeman

. 2002. A non-parametric method for the comparison of partial areas under ROC curves and its application to large health care data sets. Stat Med, 21:701–715.

31.

Litvinov

, Velders

, Bakker

, Fleuren

, Warnaar

. 1994. EpCAM: a human epithelial antigen is a homophilic cell–cell adhesion molecule. J Cell Biol, 125:437–446.

32.

Farid

, Szkudlinski

. 2004. Structural and functional evolution of the thyrotropin receptor. Endocrinology, 145:4048–4057.

33.

Abrahamsson

, Aaltonen

, Engilbertsson

, Liedberg

, Patschan

, Rydén

, Sjödahl

, Gudjonsson

. 2017. Circulating tumor cells in patients with advanced urothelial carcinoma of the bladder: association with tumor stage, lymph node metastases FDG-PET findings, and survival. Urol Oncol, 35:606.e9–606.e16.

34.

Liu

, Ling

, Qi

, Lan

, Zhu

, Zhang

, Bao

, Zhang

. 2017. Prognostic value of circulating tumor cells in advanced gastric cancer patients receiving chemotherapy. Mol Clin Oncol, 6:235–242.

35.

Bulfoni

, Gerratana

, Del Ben

, Marzinotto

, Sorrentino

, Turetta

, Schoes

, Toffoletto

, Isola

, Beltrami

, Di Loreto

, Beltrami

, Puglisi

, Cesselli

. 2016. In patients with metastatic breast cancer the identification of circulating tumor cells in epithelial-to-mesenchymal transition is associated with a poor prognosis. Breast Cancer Res, 18:30.

36.

Jueckstock

, Rack

, Friedl

, Scholz

, Steidl

, Trapp

, Tesch

, Forstbauer

, Lorenz

, Rezai

, Häberle

, Alunni-Fabbroni

, Schneeweiss

, Beckmann

, Lichtenegger

, Fasching

, Pantel

, Janni

, SUCCESS Study Group. 2016. Detection of circulating tumor cells using manually performed immunocytochemistry (MICC) does not correlate with outcome in patients with early breast cancer—results of the German SUCCESS-A- trial. BMC Cancer, 16:401.

37.

, Wang

, Peng

, Zhu

, Shen

. 2016. The significant prognostic value of circulating tumor cells in triple-negative breast cancer: a meta-analysis. Oncotarget, 7:37361–37369.

38.

Breiteneder-Geleff

, Soleiman

, Kowalski

, Horvat

, Amann

, Kriehuber

, Diem

, Weninger

, Tschachler

, Alitalo

, Kerjaschki

. 1999. Angiosarcomas express mixed endothelial phenotypes of blood and lymphatic capillaries: podoplanin as a specific marker for lymphatic endothelium. Am J Pathol, 154:385–394.

39.

Winkens

, Pachmann

, Freesmeyer

. 2013. Circulating epithelial cells in patients with thyroid carcinoma. Can they be identified in the blood?. Nuklearmedizin, 52:7–13.