Introduction
Although improvements have been made in surveillance and clinical treatment, the prognosis of colorectal cancer (CRC) remains very poor due to the high incidence of recurrence and metastasis. Approximately 20%–45% of patients who undergo curative resection subsequently develop local tumor recurrence or metastasis at distant sites.1 The lack of effective methods for timely diagnosis and monitoring response to anticancer therapy is the main obstacle for improving overall survival (OS) of patients with CRC.
Traditional clinicopathological parameters and serologic tumor markers offer limited information for CRC diagnosis, prognosis prediction, and monitoring the therapeutic response in a real-time manner. Therefore, it is urgent to develop a reliable and versatile method to identify high-risk factors of recurrent patients and make continuous surveillance for antitumor treatment response.2
The distribution of circulating tumor cells (CTCs) in the blood plays an important role in the initiation of metastases and tumor recurrence after surgery.3 The clinical relevance of CTCs as a prognostic and/or surrogate marker of treatment response has been established in several cancer types, such as breast cancer,4 CRC,5 and prostate cancer.6
A multicenter prospective study which included 456 patients with metastatic CRC (mCRC) demonstrated that CTCs level before treatment was an independent prognostic factor for progression-free survival (PFS) and OS.7 A meta-analysis based on 12 relevant studies demonstrated that detection of CTCs is an independent prognostic factor for survival.8 These studies confirm the association between CTCs and worse PFS and OS in patients with metastatic disease.
Most of these studies focus on the correlation of CTCs enumeration with prognosis. However, recent studies showed that only CTCs enumeration was not enough to reflect the heterogeneous condition of CRC.3,9
CTCs disseminate from primary tumors by undergoing phenotypic changes that allow the cells to penetrate blood vessels.10 These changes are accompanied by a process described as epithelial–mesenchymal transition (EMT), a complicated process that plays an essential role in metastasis.11
Some recent reports have provided evidence that CTCs exhibit dynamic changes in epithelial and mesenchymal composition.12–14 Mesenchymal CTCs (mCTCs) are associated with metastasis of tumor and resistance to chemotherapy. These results encourage future studies to regard the association between the expression of EMT-related markers in CTCs and cancer progression.
The family with sequence similarity 172, member A (FAM172A) was first identified in human aortic endothelial cells in 2009. Then, several researchers studied its function with cancer. Feng et al.15 found that FAM172A was downregulated among patients of hepatocellular carcinoma. It plays an important role in cell cycle control and tumor cell proliferation. G1/S phase arrest may be mediated by the Notch 3 signaling pathway. In our previous studies, the protein expression of FAM172A is significantly lower in colorectal cancerous tissues than that in adjacent tissues. It suppressed the proliferation but promoted the apoptosis and invasive potentials of colon cancer cells, which might regulate endoplasmic reticulum stress (ERS) through PERK-elF2α and ATF6-XBP1-GRP78 signal pathway.16,17 However, in papillary thyroid carcinoma (PTC), one research found that FAM172A expression was significantly higher in cancerous tissues than that in carcinoma adjacent tissues and normal thyroid tissues. FAM172A accelerated PTC cell proliferation via activation of the p38 mitogen-activated protein kinase (MAPK) signaling pathway.18
As FAM172A is closely related to CRC proliferation and invasion, it would be highly interesting to detect FAM172A expression in CTCs to get a deeper understanding of the role of FAM172A in EMT process.
The aim of this study was to discriminate different metastasis potentials of CTCs, explore the FAM172A expression in individual CTCs, and find out the correlation of CTCs subgroups and FAM172A expression in CTCs with the commonest clinical and morphological variables of CRC patients.
Methods
Patient samples and blood collection
Patients were recruited by Zhujiang Hospital from March 2015 to December 2015. This study was approved by the Ethical Committee of Guangzhou Zhujiang Hospital. All patients had given informed consent to be included in this study.
Blood samples were collected before surgery or adjuvant chemotherapy from patients with early stage and before palliative chemotherapy from those with advanced disease. Blood samples (5 mL) were drawn into heparinized tubes and stored at 4°C within 4 h.
CTCs identification
Erythrocytes were removed using a red blood cell lysis buffer containing ammonium chloride (NH4Cl) and then were transferred to the filtration tube and filtered with the help of a pump valve. CTCs were isolated using a calibrated membrane with 8-µm diameter pores.19
Four-color fluorescent imaging is used to differentiate CTC types. The cells on the membrane were hybridized for 2 h, and un-bound probes were washed for three times with phosphate-buffered saline (PBS). Subsequently, samples were incubated with preamplifier solution for 20 min and then incubated with amplifier solution (three types of fluorescently labeled probes, which had been conjugated with the fluorescent dyes: Alexa Fluor 594 (EpCAM and CK8/18/19), Alexa Fluor 488 (vimentin and twist), and Alexa Fluor 647 (CD45)). Finally, the cells were stained with 4′,6-diamidino-2-phenylindole (DAPI) for 5 min and then analyzed with a fluorescence microscope.12,13
The leukocytes were characterized as CD45+ DAPI+ cells. CTCs were classified into three subgroups: (1) epithelial marker–positive CD45− DAPI+ cells (epithelial CTCs); (2) biophenotypic epithelial/mesenchymal marker–positive CD45− DAPI+ cells (biophenotypic CTCs); and (3) mesenchymal marker–positive CD45− DAPI+ cells (mCTCs).
Statistical methods
Correlation of CTCs with clinical variables was done by contingency table analysis using the chi-square test. Continuous data were compared using nonparametric tests (Mann–Whitney test for comparison between two groups and Kruskal–Wallis test for comparison among three or more groups). All analyses were conducted by SPSS 20.0. For all the analyses, p < 0.05 was considered statistically significant.
Results
Patient demographics
Blood samples for CTCs assessment were collected in 45 consecutive patients with CRC. Clinical and morphological characteristics of the patients are summarized in Table 1. The median number of CTCs isolated was 4 (range = 0–31).
Table 1.Demographics of patients included in the study (n = 45).
Characteristics | n | % |
---|---|---|
Age (years) | ||
≤60 | 26 | 57.8 |
>60 | 19 | 42.2 |
Gender | ||
Male | 18 | 40 |
Female | 27 | 60 |
Tumor location | ||
Colon | 30 | 66.7 |
Rectal | 15 | 33.3 |
Tumor size (cm) | ||
≤5 | 32 | 71.1 |
>5 | 13 | 28.9 |
Tumor grade | ||
Low | 8 | 17.8 |
Moderate | 37 | 82.2 |
Vascular invasion | ||
No | 32 | 71.1 |
Yes | 13 | 28.9 |
Depth of invasion | ||
T1–T3 | 21 | 46.7 |
T4 | 24 | 53.3 |
Lymphatic metastasis | ||
No | 20 | 44.4 |
Yes | 25 | 55.6 |
Distant metastasis | ||
No | 39 | 86.7 |
Yes | 6 | 13.3 |
TNM stage | ||
I | 5 | 11.1 |
II | 20 | 44.4 |
III | 14 | 31.1 |
IV | 6 | 13.3 |
CEA (ng/mL) | ||
≤5 | 24 | 53.3 |
>5 | 21 | 46.7 |
Ki-67 | ||
≤60 | 24 | 53.3 |
>60 | 21 | 46.7 |
CTCs counts | ||
≥1 CTCs/5 mL | 34 | 75.6 |
≥3 CTCs/5 mL | 28 | 62.2 |
≥5 CTCs/5 mL | 20 | 44.4 |
CEA: carcinoembryonic antigen; Ki-67: nuclear-associated antigen Ki-67; CTCs: circulating tumor cells; TNM: tumor–node–metastasis.
mCTCs closely related to hematogenous metastasis
The CTCs could be classified into three subgroups according to the expressed EMT markers, including epithelial CTCs, biophenotypic epithelial/mCTCs, and mCTCs. Typical photographs are shown in Figure 1.
Figure 1.
Representative images of three subgroups of CTCs isolated from patients with colorectal cancer based on RNA-ISH staining of E (red dots) and M (green dots) markers. (The scale bar is 10 m. E: epithelial; M: mesenchymal.)
[Figure omitted. See PDF]
Overall, ≥3 CTCs/5 mL were detected in 28 of 45 patients (62.2%), which was defined as CTCs positive. mCTCs were found in 26 enrolled patients; ≥1 mCTCs/5 mL was defined as mCTCs positive. Correlations between typical clinical/pathological variables and the presence of CTCs in blood were analyzed by chi-square test,20 which are shown in Table 2.
Table 2.Correlation among CTCs and clinical/morphological variables (n = 45).
Characteristics | n | ≥3 CTCs/5 mL | p | ≥1 mCTCs/5 mL | p |
---|---|---|---|---|---|
Tumor location | |||||
Colon | 30 | 20 | 0.384 | 18 | 0.670 |
Rectal | 15 | 8 | 8 | ||
Tumor size (cm) | |||||
≤5 | 32 | 21 | 0.460 | 16 | 0.097 |
>5 | 13 | 7 | 10 | ||
Tumor grade | |||||
Low | 8 | 6 | 0.411 | 7 | 0.113 |
Moderate | 37 | 22 | 19 | ||
Vascular invasion | |||||
No | 32 | 18 | 0.195 | 15 | 0.020 |
Yes | 13 | 10 | 11 | ||
Depth of invasion | |||||
T1–T3 | 21 | 10 | 0.203 | 8 | 0.058 |
T4 | 24 | 18 | 18 | ||
Lymphatic metastasis | |||||
No | 20 | 13 | 0.615 | 10 | 0.434 |
Yes | 25 | 15 | 16 | ||
Distant metastasis | |||||
No | 39 | 23 | 0.252 | 20 | 0.024 |
Yes | 6 | 5 | 6 | ||
TNM stage | |||||
I and II | 25 | 13 | 0.114 | 12 | 0.138 |
III and IV | 20 | 15 | 14 | ||
CEA (ng/mL) | |||||
≤5 | 24 | 12 | 0.071 | 11 | 0.083 |
>5 | 21 | 16 | 15 | ||
Ki-67 | |||||
≤60 | 22 | 11 | 0.299 | 11 | 0.302 |
>60 | 23 | 17 | 15 |
CTCs: circulating tumor cells; mCTCs: mesenchymal circulating tumor cells; CEA: carcinoembryonic antigen; Ki-67: nuclear-associated antigen Ki-67; TNM: tumor–node–metastasis.
There was no correlation between positive CTCs and most of the clinicopathological features. Only stage (52.0% in stages I and II and 75.0% in stages III and IV, p = 0.114) and carcinoembryonic antigen (CEA) level (50.0% in CEA ≤5 ng/mL and 76.2% in CEA >5 ng/mL, p = 0.071) were correlated with positive CTCs, but did not reach statistical significances (Table 2).
In CRC patients, mCTCs percentage was significantly increased along with tumor progression. We observed a significant association between mCTCs positivity and development of distant metastases in CRC patients. mCTCs were detected in all the patients with distant metastasis, which were significantly higher than those in patients who did not develop distant metastasis (100% vs 51.3%, p = 0.024). In addition, mCTCs were also closely related to vascular invasion. Our study showed that mCTCs were more common in patients with vascular invasion (84.6% vs 46.9%, p = 0.020). There was also a clear association between the presence of mCTCs and the depth of invasion and/or tumor–node–metastasis (TNM) stage, but did not reach statistical significances (Table 2).
The significantly higher percentage of mCTCs in the more aggressive disease status prompts us to hypothesize whether mCTCs can be a surrogate marker of tumor aggressiveness.
FAM172A gene expression in CTCs correlated with tumorous aggressivity
Furthermore, in our platform, CTCs can be captured and used for further analysis of gene expression. This is attractive because we can obtain genome information from the cancers via CTCs without invasive procedures and detect genetic changes in real time, which has the potential to provide predictive information to guide the selection of therapy.
A recent report showed that FAM172A suppressed the proliferative and invasive potentials of CRC cell line.17 Nevertheless, how FAM172A was expressed in CTCs and their clinical values were still unknown.
In all, 28 patients with CTCs positive (≥3 CTCs/mL) were enrolled for evaluating biomarker expression. FAM172A+ CTCs were detected in 20/28 patients (71.4%). Photographs of CTCs with FAM172A expression are shown in Figure 2.
Figure 2.
Representative images of FAM172A expression in three subgroups of CTCs isolated from patients with colorectal cancer based on RNA-ISH staining of E (red dots), M (green dots) markers, and FAM172A (purple dots) markers. (The scale bar is 10 m. E: epithelial; M: mesenchymal; eCTCs: epithelial circulating tumor cells; bCTCs: biophenotypic circulating tumor cells; mCTCs: mesenchymal circulating tumor cells.)
[Figure omitted. See PDF]
Previous studies have found wide molecular and cellular heterogeneity of CTCs from the same cancer type and even from the same patient. Our study found that the overall expression rate of FAM172A in CTCs was 60.7%, with 56.3% in epithelial CTCs, 58.6% in biophenotypic CTCs, and 68.8% in mCTCs. The FAM172A expression was significantly higher in mCTCs than that in epithelial CTCs, which meant that FAM172A may correlate with malignant degree of tumor and promote cancer cell metastasis and invasion.
The hypothesis was supported by the analyses of the relationship between FAM172A gene expression and characteristics of CRC patients, which are shown in Table 3. We observed a significant association between FAM172A expression and depth of invasion in CRC patients (68.1% in T1–T3 vs 51.3% in T4, p = 0.024). Besides, higher nuclear-associated antigen Ki-67 (Ki-67) value was associated with higher FAM172A expression rate (71.3% in Ki-67 ≤ 60 vs 48.8% in Ki-67 > 60, p = 0.003). In addition, the FAM172A expression rate in mCTCs was closely correlated with metastasis-associated clinicopathological features, such as vascular invasion (78.9% vs 37.5%, p = 0.007) and depth of invasion (77.3% in T1–T3 vs 33.3% in T4, p = 0.004) in CRC patients. This meant that combining CTCs subgroups with FAM172A gene expression may enhance clinical prediction of CRC metastasis.
Table 3.Demographics of patients with CTCs ≥ 3 used for analysis of FAM172A expression rate with clinicopathological features (n = 28).
Characteristic | FAM172A expression in CTCs (%) | p | FAM172A expression in mCTCs (%) | p |
---|---|---|---|---|
Tumor location | ||||
Colon | 64.0 | 0.151 | 56.0 | 0.977 |
Rectal | 52.6 | 55.6 | ||
Tumor size (cm) | ||||
≤5 | 56.6 | 0.164 | 51.9 | 0.497 |
>5 | 68.0 | 62.5 | ||
Tumor grade | ||||
Low | 75.0 | 0.523 | 60.0 | 0.841 |
Moderate | 69.3 | 55.3 | ||
Vascular invasion | ||||
No | 58.7 | 0.619 | 37.5 | 0.007 |
Yes | 62.7 | 78.9 | ||
Depth of invasion | ||||
T1–T3 | 51.3 | 0.024 | 33.3 | 0.004 |
T4 | 68.1 | 77.3 | ||
Lymphatic metastasis | ||||
No | 64.5 | 0.250 | 52.6 | 0.708 |
Yes | 55.8 | 58.3 | ||
Distant metastasis | ||||
No | 58.1 | 0.297 | 50.0 | 0.190 |
Yes | 69.6 | 72.7 | ||
TNM stage | ||||
I and II | 63.8 | 0.378 | 52.4 | 0.658 |
III and IV | 57.1 | 59.1 | ||
CEA (ng/mL) | ||||
≤5 | 60.4 | 0.822 | 55.6 | 0.780 |
>5 | 58.8 | 60.0 | ||
Ki-67 | ||||
≤60 | 48.8 | 0.003 | 52.0 | 0.553 |
>60 | 71.3 | 61.1 |
CTCs: circulating tumor cells; mCTCs: mesenchymal circulating tumor cells; CEA: carcinoembryonic antigen; Ki-67: nuclear-associated antigen Ki-67; TNM: tumor–node–metastasis.
CTCs/FAM172A detection may predict high-risk subgroups in stage II CRC
Correlations between CTCs and prognostic subgroups were analyzed for patients of stage II CRC. CTCs detection would be an easy and reproducible method to select high-risk stage II patient candidates for adjuvant chemotherapy. At present, high-risk stage II patient is defined by clinical/pathological prognostic factors such as T4, perforation, acute bowel obstruction, undifferentiated tumors, high preoperative CEA levels, or <12 lymph nodes removed.20
The correlations of CTCs and FAM172A detection with prognostic subgroups in stage II CRC are shown in Table 4. We found that mCTCs positive rate (66.7% vs 37.5%, p = 0.199) and FAM172A expression positive rate (54.5% vs 22.2%, p = 0.142) were both higher in high-risk groups than those in low-risk groups, although they did not reach statistical significances.
Table 4.CTCs/FAM172A detection and prognostic subgroups in stage II colorectal cancer.
Prognostic subgroups | CTCs positive (%) | p | mCTCs positive (%) | p | FAM172A expression positive (%) | p |
---|---|---|---|---|---|---|
Low risk | 5 (50.0) | 0.653 | 3 (37.5) | 0.199 | 2 (22.2) | 0.142 |
High risk | 6 (60.0) | 8 (66.7) | 6 (54.5) |
CTCs: circulating tumor cells; mCTCs: mesenchymal circulating tumor cells.
Limited by sample size, there is not enough evidence for our study to discriminate high/low risk of prognostic subgroups. However, CTCs have potential clinical value in selection of patients in high-risk groups, as it is controversial on the selection of chemotherapy method based on clinicopathological criteria.
Discussion
Blood sampling is a widely accepted method with less trauma and less pain and is easy to perform. CTCs might be used as a surrogate marker in evaluating therapeutic interventions for CRC.
CellSearch™ CTC Test is Food and Drug Administration (FDA) approved as an aid in the monitoring of patients with mCRC.5 The presence of ≥3 CTCs is associated with decreased PFS and OS in CRC patients, as well as the prognosis, no matter whether therapy was used.
This enrichment approach involves the attachment of magnetic particles to EpCAM expression on the cell surface for separation of CTCs from the sample using magnetic fields. Although it is frequently used, CellSearch System needs to be interpreted with caution.21
The presence of non-tumor epithelial cells within the bloodstream may contribute to the false-positive results. It has been noted that patients with benign colon disease exhibit “tumor cells” as detected by the CellSearch System (11.3%).22
Besides, this approach might miss CTCs that have low levels of EpCAM expression and fail to detect the most aggressive CTCs subpopulation which may have undergone EMT.23 For example, a previous study has demonstrated the rarity of CTCs in early CRC, in which 20 consecutive patients underwent curative resection for stages I–III CRC.24 The detection rate was 5% in the preoperative samples using CellSearch System, using 2 CTCs/7.5 mL as the cutoff value. Although the cascades of cancer metastasis formation are not fully understood, the EMT process is believed to have a great role in these cascades.25
In this article, we have taken all the above-mentioned limitations into consideration. In order to minimize CTCs losses, we isolated CTCs via a filter-based method, which could entrap non-blood-derived cells because of their bigger size and inflexibility. Afterward, an RNA in situ hybridization (RNA-ISH) method based on the branched DNA signal amplification technology was used to classify the CTCs according to EMT markers: an antibody cocktail consisting of EpCAM, CK8/18/19, vimentin, and twist. Hematopoietic cells were excluded using CD45 markers.13
Classifying CTCs by EMT markers helps to identify the more aggressive CTCs subpopulation and provides useful evidence for determining an appropriate clinical approach.26 Therefore, this study has the potential to provide better prognostic information on the probability of metastasis for cancer patients in the early stage.
CRC is the third leading cause of cancer death in China.1 In the last few years, there was a significant expansion in the number of available systemic therapies for mCRC.27 However, the increasing treatment options lead to greater complexity in decision making. Use of a biomarker to guide therapy in a noninvasive manner would thus be of great potential clinical utility.
As liquid biopsies, CTCs have great potential to prognosticate disease and guide treatment in CRC. Moreover, they have important role in determining the genome information of tumor metastasis, providing comprehensive biomarker detection for targeted therapies and determining drug resistance.28 Because EMT can be used as a potential biomarker of cancer metastasis and therapeutic resistance, the classification of CTCs according to their EMT phenotype can help to identify the most aggressive CTCs subpopulation and provide data for clinical applications.29 In addition, as the FAM172A expression was associated with the malignant degree of tumor, it can be used to identify patients with more serious disease stage and make appropriate treatment accordingly as early as possible.
Although this is an exploratory study, and the number of available samples is small, it highlights the potential value of CTCs isolation as a source for molecular analysis. In this study, the sample size of follow-up was not enough to confirm the prognostic value of CTCs detection at the same cutoff, and this will be the objective of future analysis. According to our study, CTCs positive rate was 60.0% and 50.0% in high-risk group and low-risk group, respectively, and FAM172A expression positive rate was 54.5% and 22.2% in high-risk group and low-risk group, respectively. We used the following parameters to calculate the sample size: α = 0.05, β = 0.10, and two-sided test. The results show that a sample size of 1038 and 92 was needed when using CTCs and FAM 172, respectively, to predict high-risk CRC in the future follow-up study.
Declaration of conflicting interestsThe author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
FundingThe author(s) received no financial support for the research, authorship, and/or publication of this article.
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
© 2017. This work is published under http://creativecommons.org/licenses/by-nc/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Abstract
Previous studies used to enumerate circulating tumor cells to predict prognosis and therapeutic effect of colorectal cancer. However, increasing studies have shown that only circulating tumor cells enumeration was not enough to reflect the heterogeneous condition of tumor. In this study, we classified different metastatic-potential circulating tumor cells from colorectal cancer patients and measured FAM172A expression in circulating tumor cells to improve accuracy of clinical diagnosis and treatment of colorectal cancer. Blood samples were collected from 45 primary colorectal cancer patients. Circulating tumor cells were enriched by blood filtration using isolation by size of epithelial tumor cells, and in situ hybridization with RNA method was used to identify and discriminate subgroups of circulating tumor cells. Afterwards, FAM172A expression in individual circulating tumor cells was measured. Three circulating tumor cell subgroups (epithelial/biophenotypic/mesenchymal circulating tumor cells) were identified using epithelial–mesenchymal transition markers. In our research, mesenchymal circulating tumor cells significantly increased along with tumor progression, development of distant metastasis, and vascular invasion. Furthermore, FAM172A expression rate in mesenchymal circulating tumor cells was significantly higher than that in epithelial circulating tumor cells, which suggested that FAM172A may correlate with malignant degree of tumor. This hypothesis was further verified by FAM172A expression in mesenchymal circulating tumor cells, which was strictly related to tumor aggressiveness factors. Mesenchymal circulating tumor cells and FAM172A detection may predict highrisk stage II colorectal cancer. Our research proved that circulating tumor cells were feasible surrogate samples to detect gene expression and could serve as a predictive biomarker for tumor evaluation.
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
Details
1 Department of General Surgery, Zhujiang Hospital, Southern Medical University, Guangzhou, China