Clinico-pathologic characteristics of sub-epithelial tumor (SET) like gastric cancers

Abstract

It has been known that sub-epithelial tumor (SET)-like gastric cancer (GC) is a special type of gastric cancer with a relatively low prevalence of approximately 1%. Most of the studies are old and would be underestimated. Recently, the application of endoscopic ultrasound has increased, so its frequency is expected to increase. A consecutive study was undertaken from March 2020 to March 2024, and we analyzed 380 GC tissues (194 early GCs; 186 advanced GCs) from the patients who had received gastrectomy or endoscopic submucosal dissection. Through picture archiving and communication system, the endoscopic findings at the time of diagnosis were analyzed. After removal, molecular biological characteristics were examined. The prevalence of SET like GC was 5.8% (22/ 380) endoscopically. There were no differences between the SET like GC and the other group with respect to gender, smoking, drinking history, and Helicobacter pylori infection. SET like GC patients were statistically older (P = 0.04). The frequency of advanced GC (P = 0.03) and the frequency of microsatellite instability (MSI) (P < 0.01) were statistically higher in SET like GC. Exceptional cases included Epstein-Barr virus-associated lymphoepithelioma-like carcinoma, undifferentiated carcinoma with rhabdoid features, mixed neuroendocrine tumor, case with gastritis cystica profunda, and cases that used anti-ulcer drugs or responded to chemotherapy. It also appeared in gastric cancer with confirmed genetic predisposition. SET like GC had many advanced cases and MSI-positive GCs. If we investigate various exceptional cases, it is likely that their frequency will continue to increase. We urgently need measures to address this.

Introduction

In general, gastric cancers (GCs) originate from the mucosa, usually grow laterally along the mucosa, and invade the submucosa or deeper. There are rare cancers that invade deeply while maintaining the mucosal structure and have a special growth pattern that takes the form of a sub-epithelial tumor (SET). GCs presenting as a SET so called as “SET-like GCs” is extremely rare. The first case in the English literature was reported by Ohara et al. in 1997¹. Currently, the clinic-pathologic characteristics of SET-like GCs are unclear, because of the limited number of cases reported. Most prevalence studies are from about 20 years ago, and there are no multi-institutional or multi-national studies, so it is difficult to confirm the exact prevalence of SET like GCs. Looking at the studies reported to date, the prevalence according to the definition of SET like GCs were very low, at 0.24 ~ 1.27% of all GCs^2–4. However, most studies were conducted in Japan. Although the definition of the disease is endoscopically diagnosed, many cases were confirmed based on pathologic findings, so some may have been underestimated. Furthermore, since endoscopic ultrasound has become a commonly performed test for the invasion of GCs, an approach from this perspective seems necessary.

The definition of SET-like GCs used in Japan refers to the appearance of a SET in radiological and endoscopic findings, and it is known that the long axis of the exposed surface of the cancer is less than 1/3 of the long axis of the tumor in the macroscopic and histological findings of fixed specimens (Fig. 1)³. Although the findings of endoscopic ultrasound are not absolute in diagnosis, they may be of some help. Gastrointestinal stromal tumors (GISTs), a malignant subepithelial tumor, is usually suspected when it is larger than 3–5 cm, grows rapidly, has irregular echogenicity, and has irregular margins. Sometimes, ultrasound-guided fine needle aspiration can be attempted when a pathologic diagnosis cannot be made on mucosal biopsy. In some cases, rare cases have been reported in which endoscopic mucosal resection was applied. In actual clinical practice, when lesions are observed in the form of SET and have erosion or ulceration in the center, they can be diagnosed as carcinoid tumors (neuroendocrine tumors), and in addition, it is not uncommon for them to be lymphomas including MALToma. In addition, some metastatic gastric cancers that have spread upward may appear as raised erosions or SET like forms on endoscopic examination. These appear as SETs because they spread through the blood vessels of the submucosal layer of the stomach through hematogenous metastasis. In Korea, SET-like GCs with mucinous adenocarcinoma tend to be reported frequently (Fig. 2). Therefore, the aim of this study was to investigate the prevalence, clinicopathological characteristics, and molecular biological features of SET-like gastric adenocarcinomas, and to identify factors independently associated with this morphological phenotype.

Fig. 1.

SET-like gastric cancer (GC) is defined when subepithelial tumor (SET) appearance is present on both radiological and endoscopic findings, with the long axis of the exposed tumor surface being less than 1/3 of the total long axis of tumor.

Fig. 2.

Endoscopic and abdominal CT findings of SET-like GC with mucinous adenocarcinoma.

Materials and methods

The study was on the patients with gastric cancer who we have treated with gastrectomy or endoscopic submucosal dissection from 2020 to 2024. This study protocol was approved by the Institutional Research Ethics Board of the Catholic University of Korea (VC25RISI0057) and adhered to the tenets of the Declaration of Helsinki. The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required from the participants and/or their legal guardian(s) in accordance with institutional requirements. Clinical data such as age, gender, weight, and height at the time of surgery were collected, and histopathologic features - depth of invasion, and histologic type were analyzed. Pathological cancer stage was interpreted according to the evaluation protocol recommended by 7th edition of American Joint Committee on Cancer (AJCC) TNM classification. Early gastric cancer (EGC) defined as tumor cells were confined to mucosa and submucosa regardless of lymph node metastasis.

SET-like gastric cancer was defined as a lesion showing subepithelial tumor appearance on endoscopy, with the long axis of the exposed mucosal surface being less than one-third of the total tumor long axis, as previously described. The endoscopic appearance prior to surgery was confirmed through picture archiving and communication system. Two experienced endoscopists (WCC and NRL) independently reviewed the endoscopic images while blinded to histopathological and molecular results, and final classification was determined by consensus.

For advanced gastric cancer (AGC), the Borrmann type classification was followed as previously known. On the other hand, for EGC, the Paris classification was used⁵. In the Paris classification, type 0 is divided into three categories: protruding (0 - I), non-protruding and nonexcavated (0 - II), and excavated (0 - III). Type I was defined as a height more than twice the thickness of the mucosal layer (> 2.5 mm or more than the height of the biopsy forcep). The classification of diffuse atrophic gastritis was defined as open type if it included gastric cardia according to the Kimura-Takemoto classification. All patients underwent Helicobacter pylori (H. pylori) testing and molecular biological testing (MSI, EBV, HER2, EGFR), and patients with gastric cancer that showed SET endoscopically underwent endoscopic ultrasound examination before resection.

Evaluation of microsatellite instability (MSI) state

We used the PCR- fragment analysis to analyze whether there is MSI. This test is a test that separates genomic DNA from normal and cancerous tissues, amplifies five microsatellite markers (BAT-25, BAT-26, D2S123, D5S346, and D17S250) using the PCR method, and then analyzes instability using the fragment analysis method using the ABI3500xL Dx from Applied biosystems. According to the National Cancer Institute Bethesda panel criteria, if instability is shown in two or more of the five markers, it is defined as MSI-high, if instability is shown in one marker, it is MSI-low, and if instability is not observed in any markers, it is defined as microsatellite stable.

Epstein barr virus (EBV)

In all cases, the presence of EBV in cancer cells was previously assessed by in situ hybridization for EBV-encoded RNA (EBER) using either automated systems or manual staining methods.

Human epidermal growth factor receptor-2 (HER2) and epidermal growth factor receptor (EGFR) over-expression

Immunohistochemical analysis for HER2 and EGFR was performed on tumor sections. We performed immunohistochemical staining using tissue microarray sections and a BenchMark XT autostainer (Ventana Medical Systems, Tucson, AZ, USA), according to the manufacturer’s protocol. The commercially available antibodies used for immunohistochemistry of EGFR (1:50; BD Biosciences, Rockville, MD, USA). HER2 was detected using mouse monoclonal antibody, clone 45MI (Novocastra, Newcastle upon Tyne, England). The immunohistochemical results were scored semi-quantitatively using a four-point scale: 0, no immunoreaction; 1+, faint or equivocal immunoreaction in < 10% of cells; 2+, unequivocal, strong immunoreaction in < 30% of cells; and 3+, unequivocal, strong immunoreaction in > 30% of cells. Tumors with 1+, 2 + or 3 + expression were interpreted as positive, while tumors with no expression were interpreted as negative. HER2 overexpression positivity was defined as 2 + or 3 + in immunohistochemistry, and EGFR overexpression was defined as 3 + in immunohistochemistry.

Statistics

All statistical analyses were performed with the SPSS program (SPSS version 25.0; IBM, Chicago, IL, USA). Values are expressed as mean ± standard deviation. Categorical variables were tested by a Chi-square test or a logistic regression model. Firth’s penalized logistic regression was used to address potential small-sample bias due to the limited number of SET-like cases. P-value < 0.05 was considered significant.

Results

Characteristics of SET like GC compared with gastric cancer without SET like feature

A total 380 gastric cancer patients were evaluated. Complete data were available for all patients. MSI-H and MSI-L were analyzed separately given their distinct clinical implications. The prevalence of SET like GC was 5.8% (22/ 380) endoscopically. There were no differences between the SET like GC and the other group with respect to gender, smoking, drinking history, H. pylori infection, tumor location, and the presence of diffuse atrophy. The frequency of poorly differentiated gastric cancer, including undifferentiated gastric cancer, did not differ between the two groups. SET like GC patients were statistically older (P = 0.04). The frequency of advanced GC (P = 0.03) and the frequency of microsatellite instability (P < 0.01) were statistically higher in SET like GC (Table 1). Besides, there was no difference in the frequency of detection, HER2 and EGFR over-expression.

Table 1.

Baseline demographics of the patients.

	SET like Gastric cancers (N = 22)	Gastric cancers without SET features (N = 358)	P value
Age (years)	71.8213.23	66.3112.07	0.04*
Male sex (%)	14	273	0.46
Current smoker	10	205	0.40
Alcohol	12	230	0.68
H. pylori infection	15	197	0.22
Diffuse atrophy (open type)	15	196	0.23
Location (body- cardia)	4	108	0.23
Poorly differentiated cancer	9	158	0.75
Early gastric cancer	7	187	0.03*
Advanced gastric cancer	15	171
Microsatellite instable tumor (MSI)	MSI-H 10 MSI-L 3	MSI-H 59 MSI-L 11	< 0.01*
Epstein Barr virus (EBV) association	2	25	0.71
HER2 over-expression	3	49	0.98
EGFR over-expression	12	176	0.64
Endoscopic treatment	6	90	0.83

Values are presented as mean ± standard deviation or number (%).

SET; sub-epithelial tumor.

H. pylori; Helicobacter pylori,

Closed, Open; classified by Kimura Takemoto endoscopic classification.

HER2; human epidermal growth factor receptor 2.

EGFR; epidermal growth factor receptor.

Endoscopic features of MSI GCs

We compared 83 MSI-positive gastric cancer patients (MSI-H, n = 69;MSI-L, n = 14) with 297 MSS controls. The overall frequency was 21.8%, and there was no difference in the frequency of EGC. As previously known, MSI-H gastric cancer patients were relatively older, more frequently female, had more frequent diffuse gastritis, and were often located in the antrum (Table 2). In EGC, no prominent morphology of MSI-H gastric cancer was identified, but in the group showing 0-II, the IIa-predominant morphology was statistically more prevalent. In AGC, B-II (ulcerative) lesions were statistically prominent (Fig. 3).

Table 2.

Morphologic characteristics of microsatellite instable gastric cancer (MSI GC) compared to microsatellite stable gastric cancers (MSS GC).

	MSI gastric cancer (N = 83)	MSI-H gastric cancer (N = 69)	MSS gastric cancer (N = 297)	P value
Age (years)	72.64 ± 9.68	73.33 ± 9.82	65.04 ± 12.28	< 0.01*
Male sex	44	36	216	< 0.01*
H. pylori infection	43	36	159	0.84
Diffuse atrophy (open type)	67	58	142	< 0.01*
Location (antrum)	68	64	161	< 0.01*
EGC frequency (vs. AGC)	45	41	158	0.35
EGC type 0-I	7	7	4	0.01*
0-II	34	32	148	0.61
(0-IIa dominant)	30	26	57	< 0.01*
0- III	4	2	6	0.65
AGC	38	28	139
Borrmann type I	2	1	9	0.69
type II	22	18	46	0.04*
type III	12	9	66	0.09
type IV	2	0	18	0.03*

Values are presented as mean ± standard deviation or number.

H. pylori; Helicobacter pylori,

Closed, Open; classified by Kimura Takemoto endoscopic classification.

Paris classification of early gastric cancers (Endoscopy 2005;37:570-8). Type 0 is divided into three categories: protruding (0 - I), non-protruding and nonexcavated (0 - II), and excavated (0 - III).

Borrmann type of advanced gastric cancers (Japanese Classification of Gastric Carcinoma: 3rd English edition. Gastric Cancer 2011, 14(2):101 − 12).

Fig. 3.

Endoscopic findings of SET-like gastric cancers in patients with microsatellite instability (MSI). (A) A smooth, elevated lesion with central ulceration in the gastric antrum. (B) A protruding mass with central ulcer in the antrum, demonstrating typical SET-like appearance. Cases were confirmed as MSI-high on molecular analysis.

In multivariable logistic regression analysis, AGC (OR 0.30, 95% CI 0.11–0.81, P = 0.012) and MSI-H status (OR 3.36, 95% CI 1.26–8.96, P = 0.016) were independently associated with SET-like morphology after adjusting for age, tumor location, and histological differentiation (Table 3).

Table 3.

Results of a multivariate logistic regression analysis with SET-like morphology.

	Odds Ratio (OR)	95% confidence interval (CI) Low Upper	P value
Age (years)	1.022	0.983 ~ 1.065	0.278
EGC (vs. AGC)	0.311	0.112 ~ 0.802	0.015*
Location (antrum)	1.304	0.445 ~ 4.530	0.641
Poorly differentiated cancer	0.691	0.263 ~ 1.759	0.439
MSI-H	3.265	1.248 ~ 8.570	0.016*

Variables in univariate analysis were included in the Firth’s penalized logistic regression model.

OR, odds ratio; CI, confidence interval.

* statistically significant.

Discussion

According to previous reports for SET like GCs, the male-to-female ratio is approximately 3:2, and approximately 30% of cases were not diagnosed by preoperative endoscopic biopsy. A notable result is that undifferentiated GC is approximately twice as common as differentiated GC. In this study, there was no difference in the frequency of undifferentiated gastric cancer. However, the frequency of advanced GC that progressed in SET-like GC was high, and the frequency of MSI-positive GC was more than 50%. The Cancer Genome Atlas (TCGA) project divided GCs into four molecular subtypes: EBV, MSI, genomic stability, and chromosomal instability. Among them, GCs with MSI is a specific molecular subtype of GCs. H. pylori infection contributes to occur MSI in early steps of gastric carcinogenesis. In general, the frequency of MSI-positive GCs is reported in 15–20% of all GCs, and MSI- positive GCs are mainly related to age, female sex, antrum tumor location, and intestinal type⁶. According to a previous report, 81.2% of MSI- positive GCs were located in the gastric antrum⁷. In particular, MSI- positive GCs were associated with distinctive clinicopathologic feature of synchronous or metachronous lesions⁸. Endoscopically, patients with MSI- positive GCs are more often diagnosed at an earlier disease stage (TNM stage I or II) and classified as Borrmann type I or II⁹. Although previously unreported, the results of this study showed that SET type frequently occurs in MSI-positive GCs. This finding may be associated with the potentially higher proliferation rate observed in MSI-positive tumor, although the underlying mechanism requires further investigation. A previous study has reported a high frequency of EGFR overexpression in MSI-positive gastric cancer¹⁰. This molecular biological feature appears to be reflected in endoscopic findings.

The endoscopic ultrasound findings of SET-like GCs vary depending on the mechanism of occurrence⁴. First, in the case of lymphocytic infiltrating GCs with lymphoid stroma, a relatively clear-margined uniform hypoechoic or mixed-echoic mass is observed mainly from the second to third layers. On the other hand, in the case of mucinous adenocarcinoma, depending on the degree of cancer progression, it shows a cystic appearance, invading the third and fourth layers, and sometimes even the fifth layer, and shows an anechoic or mixed-echoic finding depending on the viscosity of the mucus, and sometimes a septum is observed in the middle of lesion. Histologic features of surgical specimens include minimal intramucosal spread, early submucosal invasion, and proliferation with clear margins deeper than the submucosal layer. In addition to typical gastric adenocarcinoma, cell types include medullary carcinoma with lymphatic stroma¹¹, carcinoma with solid growth¹², focal carcinoma with severe fibrosis¹, mucinous adenocarcinoma¹³, and endocrine cell carcinoma¹⁴. Several distinct clinical and pathological conditions have been associated with SET-like morphology in gastric cancers.

Special conditions

EBV associated gastric cancer; lymphoepithelial like carcinoma

EBV is a double-stranded DNA virus belonging to the herpes virus, and it is mainly transmitted through saliva while proliferating in the epithelial cells of the oropharynx. Primary infection passes without any particular symptoms in childhood, and after that, it shows an asymptomatic latent infection form in B lymphocytes, and it is known that more than 90% of the adult population worldwide shows a positive serological reaction. EBV associated gastric cancer is mainly a depressed type with a protrusion such as SET in the upper part of the stomach near the atrophic border of the gastric mucosa¹⁵.

Undifferentiated carcinoma of the stomach

Undifferentiated carcinoma of the stomach occurs rarely, but Immunologically, it is a very aggressive neoplasm that often shows rhabdoid findings. Especially in cases with immunological rhabdoid features, it has been called by various names such as giant cell carcinoma, pleomorphic carcinoma, malignant rhabdoid tumor, adenocarcinoma with rhabdoid phenotype, and dysplastic carcinoma. The mechanism of carcinogenesis is an inactivating mutation in the SMARCA4 gene of the SWI/SNF chromatin remodeling complex¹⁶. These tumors are characterized by loss of SMARCA4 protein expression and often exhibit a solid, undifferentiated structure. Immunohistochemically, the rhabdoid tumor cells were strongly positive for cytokeratins and vimentin. Considering the literature reports published so far, it had a highly aggressive clinical course¹⁷. SET type of stomach cancer is presumed to be caused by early infiltration and rapid growth of undifferentiated carcinoma. Additionally, if severe fibrosis accompanied by mucus or edema occurs, it will more clearly take the form of SET. The poor response to conventional chemotherapy has made complete surgical resection and the development of alternative treatment methods important. They may respond to immunotherapy.

Mixed adenocarcinoma-neuroendocrine tumor

Mixed tumors of neuroendocrine tumors and gastric adenocarcinomas that occur above are very rare and are often diagnosed after surgery because it is difficult to distinguish between the two types of cancer macroscopically. It is generally assumed that the origin of occurrence is that neuroendocrine tumor cells transform during the development of adenocarcinoma, and this appears to rapidly develop into the submucosal and deeper layers to form other tumors¹⁴.

Gastric cancer with gastritis cystica profunda (GCP)

GCP is a disease in which hyperplasia and cystic expansion of the gastric epithelium form cysts in the deep and submucosal layers of the gastric mucosa. It usually occurs in patients who have undergone gastrectomy, but in rare cases, it has also been reported in cases without a history of gastric surgery. There is much controversy regarding the malignant tendency of GCP or its association with malignancy. It was said that it is not a precancerous lesion itself but a peri-cancer lesion, and the prevailing view to date has been that it is a benign lesion¹⁸. The characteristic endoscopic ultrasound finding is thickening of the gastric wall due to submucosal hyperplasia containing cysts. Macroscopic findings include thickened mucosal folds or polypoid or ulcerative lesions and SETs.

Recurrence of endoscopic procedure

The therapeutic indications of endoscopic submucosal dissection (ESD) are limited to early gastric cancer without lymph node metastasis. It is limited to early gastric cancer that is less than 2.0 cm in size, has no ulcer, is differentiated, and mucosal cancer without lymph node metastasis, but its scope has been expanded recently. However, in cases where there is a positive deep margin or submucosal invasion or lympho-vascular invasion, recurrence in the submucosal or extragastric area is more common than recurrence in the mucosa¹⁹, so it is difficult to detect local recurrence even if a gastric mucosal biopsy is frequently performed near the ESD scar during follow-up endoscopy. In this case, recurrence occurs below the surface, and after the tumor has grown to a certain extent, it appears as an ulcer and recurs on the surface in the form of an ulcer.

Administration of anti-ulcer drugs

It is known that even malignant ulcers in stomach cancer can be partially cured²⁰. Particularly, in the case of depressed type of early gastric cancer, depending on the drug treatment or observation period, a benign ulcer develops in the cancer tissue in the presence of acid and pepsin. This goes through a “malignant cycle of early gastric cancer” in which it appears as a benign ulcer as it repeats healing and relapses, and accordingly, it can appear in various forms, including the SET. Therefore, even if it is diagnosed as a benign gastric ulcer at the initial diagnosis, it is very important to confirm the healing of the gastric ulcer through an endoscopic examination after treatment, and if the ulcer is not healed, a tissue biopsy should be performed again to differentiate it from a malignant ulcer.

Response after systemic chemotherapy

For resectable gastric cancer, perioperative chemotherapy has become the standard treatment. More recently, the potential benefits of targeted therapy or immunotherapy in the perioperative or adjuvant setting have been explored. For metastatic disease, there have been notable recent advances in immunotherapy and biomarker-directed therapy. These developments are thought to have improved the response to chemotherapy for gastric cancer compared to before, and it has been confirmed that some mucosal lesions, including metastatic lesions, are cured partially.

Genetically predisposing conditions

The characteristics of hereditary tumors are that they generally occur at a younger age and occur multiple times in the developing organs compared to non-hereditary tumors. Judging from reports that the incidence of family members of gastric cancer patients is higher, it can be expected that genetic factors are also related to the cause of gastric cancer, and it is judged that genetic factors also play an important role in the gastric cancer carcinogenesis process in some gastric cancers. Hereditary gastric cancer is known to account for 3–5% of all gastric cancers in the West²¹. Hereditary tumor syndromes related to gastric cancer include hereditary nonpolyposis colorectal cancer, Li-fraumeni syndrome, familial adenomatous polyposis, and Peutz-Jeghers syndrome. In addition, Duchenne muscular dystrophy (DMD) and Prader-Willi syndrome have also been reported^22,23. Mutation of the DMD gene involves in the pathogenesis of several cancers including gastric carcinomas. In cases where genetic mutations are involved, the pathological malignancy is poor and rapid growth is common, and in such cases, the form of a SET may be observed.

Among special conditions of SET like GC, the part about the use of anti-ulcer drugs is noteworthy. Ulcers due to stomach cancer can be treated in the early stages, but they are often diagnosed at a later stage when symptoms begin, which can be clinically problematic. Recently, potent acid suppressants such as proton pump inhibitors have been shown to mask symptoms of gastric cancer and partially improve endoscopic findings of malignant ulcers²⁴. Since only the mucosal lesion can be partially healed and observed as a SET, even if the biopsy result is negative once, if it is suspected endoscopically, the biopsy should be performed continuously, and if possible, care should be taken to discontinue antiulcer medication for more than 2 weeks before the test. With the development of ESD, it has become possible to perform lump-sum resection regardless of tumor size, location, or shape, and the expansion of endoscopic treatment indications has been suggested along with the argument that the existing indications are too limited. From this perspective, it is expected that recurrence due to microscopic cancer cells in the surgical site will be frequent. Although recurrence through metastasis to surrounding lymph nodes is a common finding, caution is required because recurrence in the form of submucosal tumors is also possible. Therefore, the possibility of microscopic metastasis of cancer cells remaining below the submucosal layer after treatment cannot be completely ruled out, so regular endoscopic examinations and follow-up examinations using computed tomography, etc. are required.

This study investigated the clinical and endoscopic molecular characteristics of patients with GCs, who had undergone resection and showed a subepithelial tumor morphology. Considering the various circumstances, it seems that the number of patients with this type of GCs will increase in the future and special guidelines for this type of gastric cancer will be needed.

This study has several limitations. First, this was a retrospective, single-center study, which may limit the generalizability of our findings. Second, formal interobserver agreement analysis (e.g., kappa statistics) was not performed for the classification of SET-like morphology, although independent blinded review with consensus was employed to minimize bias. Third, the number of SET-like gastric cancer cases was relatively small (n = 22), which limited the statistical power for subgroup analyses and may have affected the stability of effect estimates in the multivariable regression model. Fourth, outcome or survival analysis was not performed because the cohort included a high proportion of early gastric cancers, and disease progression-related mortality was infrequent during the follow-up period, precluding meaningful outcome analysis. Finally, although the association between SET-like morphology and MSI was statistically significant, the retrospective design and limited sample size preclude definitive conclusions regarding causality or underlying biological mechanisms.

Author contributions

Chung WC conceptualized the study, designed the methodology, performed the formal analysis, revised and edited the sequential versions of the manuscript, and supervised the project; Lim NR performed the data curation and investigation, and drafted the original manuscript; and all authors thoroughly reviewed and endorsed the final manuscript.

Funding

The author received no funding for this work.

Data availability

The data that support the findings of this study are not openly available due to reasons of sensitivity and are available from the corresponding author upon reasonable request. Data are located in controlled access data storage at the Catholic University of Korea, St. Vincent’s hospital.

Declarations

Competing interests

The authors declare no competing interests.