Clinical characteristics and long-term disease course in patients with Crohn’s disease as diagnosed by video capsule endoscopy: a multicenter retrospective matched case-control study
Article information
Abstract
Background/Aims
Video capsule endoscopy is rarely used to diagnose Crohn’s disease in patients with negative ileocolonoscopy or cross-sectional image findings. We evaluated clinical characteristics and long-term outcomes of these rare cases.
Methods
This multicenter study included patients with Crohn’s disease from 3 tertiary hospitals from January 2007 to October 2022. Patients with normal findings on ileocolonoscopy and computed tomography (CT)/magnetic resonance (MR) enterography but had ulcerations at the small bowel detected by video capsule endoscopy were included. The controls were patients with abnormal findings on endoscopy or CT/MR enterography. Controls were case-matched in a ratio of 3:1 for sex, calendar year of diagnosis, and age at diagnosis.
Results
Among 3,752 patients, 24 (0.6%) were diagnosed with Crohn’s disease using video capsule endoscopy findings. The disease location (P<0.001) and behavior at diagnosis (P=0.013) of the cases significantly differed from that of controls. The perianal fistula modifier (25.0% vs. 33.3%, P=0.446) did not differ significantly between the 2 groups. Initial disease activity and C-reactive protein and fecal calprotectin levels were significantly lower in cases versus controls. The median Lewis score was 838 (interquartile range, 393–1,803). Over 10 years of follow-up, the cases showed significantly lower cumulative risk of complicated behavior, biologics use, Crohn’s disease-related hospitalization, and surgeries (log-rank test P<0.05).
Conclusions
Patients with Crohn’s disease whose lesions were observed only by video capsule endoscopy were rare, and exhibit different clinical characteristics and a more favorable long-term disease course compared to those who were conventionally diagnosed.
INTRODUCTION
Crohn’s disease (CD) is a chronic inflammatory bowel disease (IBD) that can affect the entire gastrointestinal tract [1]. No gold standard has been approved for diagnosing CD; therefore, the currently established diagnosis methods typically integrate the clinical, endoscopic, radiological, and histological features [2,3]. According to previous reports, the small bowel is affected in 30% to 70% of individuals diagnosed with CD, with the terminal ileum being affected in 90% of these patients. The proximal segments of the small bowel might be affected in only 10% of the remaining cases, extending beyond the scope of a colonoscopy [4]. Diagnosing isolated small bowel CD is challenging due to the nonspecific nature of symptoms and the anatomical location of the small bowel.
Video capsule endoscopy (VCE) has markedly transformed gastrointestinal imaging since its approval by the U.S. Food and Drug Administration in 2001. VCE has a high sensitivity, patient-friendly approach, reduces invasiveness, and has become a pivotal advancement in the field. VCE can be used to assess both small bowel and colonic mucosa, particularly in proximal bowel segments, providing enhanced visualization reach within the small bowel [5]. VCE is a helpful modality in diagnosing CD in patients with a normal ileocolonoscopy and/or normal radiological imaging [6]. Given that some manifestations of disease are visible only using this method, VCE has a role as the first diagnostic modality in selected patients with suspected CD [7]. VCE is highly sensitive to small bowel inflammation, especially superficial and proximal lesions, and has greater sensitivity compared with computed tomography enterography (CTE) and magnetic resonance enterography (MRE) [8-10].
It has been reported that CD only visible via VCE is a rare condition and has a more favorable disease course than that of general CD [11]. However, the number of patients included in that study was small, and there was no control group for direct comparison. CD exclusively detectable through VCE may present a distinct natural course and outcomes compared to other manifestations of CD, yet data on this is limited. We aimed to evaluate the clinical characteristics and long-term outcomes in patients whose lesions were evident on VCE only compared with a control group of patients with conventionally diagnosed CD.
METHODS
1. Patients and Study Design
We performed a retrospective cohort study in patients with newly diagnosed CD based on VCE in 3 tertiary hospitals (Asan Medical Center, Kyungpook National University Hospital, and Yeungnam University Medical Center) in Korea from January 2007 to October 2022. We included consecutive adult patients referred to hospitals with symptoms or signs of suspected CD but with negative findings of CD using ileocolonoscopy and cross-sectional imaging (including CTE or MRE) and a final diagnosis of CD using VCE (cases/VCE group) (Fig. 1).
Flowchart of the study. Cases included patients diagnosed with Crohn's disease exclusively detectable on VCE at 3 tertiary hospitals. Controls were case-matched in a ratio of 3:1 from a pool of 2,929 patients with conventionally diagnosed CD who were registered at the IBD Clinic of Asan Medical Center. CD, Crohn’s disease; VCE, video capsule endoscopy; IBD, inflammatory bowel disease.
Eligible patients were matched by sex, calendar year of diagnosis (± 2 years), and age at diagnosis (± 3 years) in a 3:1 ratio to patients conventionally diagnosed with CD. Controls were selected from 2,929 patients with CD who were registered at the IBD Clinic of Asan Medical Center. They had been diagnosed with CD based on conventional clinical, radiological, endoscopic, and histopathological criteria (Fig. 1).
We excluded patients treated with nonsteroidal anti-inflammatory drugs (NSAIDs) within 6 months prior to VCE and patients exposed to immunosuppressive medications typically used in the treatment of IBD within 6 months. All patients included in the analysis had a follow-up period of at least 6 months after diagnosis of CD. Follow-up data were collected until April 2023. The Institutional Review Board of Asan Medical Center approved the study protocol (IRB No. 2023-0747). The informed consent was waived because this design is a retrospective study.
2. Video Capsule Endoscopy
VCE was performed using PillCam SB, SB2 or SB3 (Given Imaging Ltd., Yokneam, Israel). The images were analyzed with RAPID Reader software on a RAPID workstation (software and workstation from Given Imaging Ltd.). Images were reviewed by expert gastroenterologists at the participating institutions. VCE was performed after a 12-hour fast and bowel preparation with 2 L or 4 L of polyethylene glycol and simethicone. Patients were able to resume a clear liquid diet 4 hours after capsule ingestion.
For each VCE, we defined the total small bowel transit time from the beginning of the duodenum to the cecum and divided it into tertiles. The Lewis score was calculated using the RAPID reading program for standardizing CD activity based on the degree and extent of mucosal damage [12]. VCE readers assessed villous edema, ulcers, and stenosis for each tertile. Edema was recorded as focal, patchy, or diffuse. The distribution of edema was reported as short segment (< 10% of segment), long segment (10%–50%), or extensive (> 50%). For ulcers, the number, size, and distribution were reported. The presence/absence of stenosis, presence of ulceration within the stenotic segment, and whether the passage of the capsule was seen was noted. A score of 135–789 was classified as representing mild small bowel Crohn’s inflammation, and a score of ≥ 790 was classified as representing moderate to severe activity [12].
CD was defined based on the following VCE findings: (1) more than 3 ulcerations in the absence of NSAID use, (2) more than 10 aphthous ulcers plus stenotic lesions, or (3) less than 3 ulcerations with diffuse erythema and edema, and a Lewis score > 135 [13,14].
3. Data Collection
We collected baseline data on patient demographics (age, sex, family history of IBD, smoking status, and NSAIDs exposure) and disease characteristics (clinical symptoms or signs and disease extent and behavior); clinical (Crohn’s Disease Activity Index [CDAI]) or biochemical (C-reactive protein [CRP] and fecal calprotectin) activities; and capsule endoscopy findings (mucosal edema, erythema, erosions, ulcerations, stenosis, and Lewis score).
Follow-up data included occurrence of complications (development of strictures, abscesses, fistulas, or perianal lesions), need for immunomodulators (azathioprine or methotrexate) or biologics, and CD-related hospitalization or surgery. The results of follow-up ileocolonoscopy, cross-sectional images, and VCE during the study period were also investigated.
4. Study Outcomes
The primary outcome was a comparison of the long-term disease course between the VCE group and controls. We evaluated the (1) occurrence of complicated behaviors, (2) use of immunomodulators, (3) use of biologics, (4) CD-related hospitalization, and (5) CD-related surgery. The secondary outcomes included a comparison of disease characteristics between the VCE and control groups. Additionally, the capsule endoscopy findings in the VCE group were analyzed.
5. Statistical Analysis
Descriptive statistics were used to analyze the demographic and disease characteristics. Categorical variables are presented as proportions, and continuous variables are presented as means with standard deviation (SD) for normally distributed variables or median with interquartile ranges (IQRs) for non-normally distributed variables. Differences in disease characteristics between cohorts were compared by the Student t-test or Mann-Whitney U-test for continuous variables and the Fisher exact test or the chi-square test for categorical variables where appropriate. The cumulative incidence of clinical outcomes was estimated using the Kaplan-Meier method. A log-rank test was used to compare the incidence of each outcome between the 2 groups. A two-sided P-value < 0.05 was considered statistically significant. All analyses were performed using IBM SPSS statistics 21.0 for Windows (IBP Corp., Armonk, NY, USA).
RESULTS
1. Baseline Characteristics of the Patients and Disease
Out of 3,752 patients diagnosed with CD across 3 tertiary hospitals over a study period of around 16 years, 24 patients (0.6%) were diagnosed based on VCE and were included in the VCE group (Fig. 1). Each patient was matched to 72 patients in the control group. Most patients were males (75.0%) with a mean age of 28.1 years (SD, 9.6) at diagnosis. The mean duration from symptom onset to diagnosis did not differ significantly (23.9 months [SD, 28.9] vs. 31.0 months [SD, 56.2], P=0.555). Marked differences in disease location and behavior in the groups based on the Montreal Classification (Table 1) were noted. In the VCE group, all patients showed small bowel location (ileal or upper GI), whereas most of the control group (62.5%) had ileocolonic location (P<0.001). Non-stricturing and non-penetrating (B1) type was observed more often in VCE than in the control group (91.7% vs. 62.5%), whereas penetrating (B3) type was only observed in the control group (0.0% vs. 26.4%, P=0.013). No significant differences in the perianal fistula modifier were observed (25.0% vs. 33.3%, P=0.446) between the groups.
Demographic and Clinical Characteristics of Included Patients
The disease activity index (CDAI) and biomarkers (CRP and fecal calprotectin) at baseline were significantly lower in the VCE group compared to the control group (Table 1). In the VCE group, 23 patients (95.8%) used 5-aminosalicylic acid (5-ASA) as treatment at the time of initial diagnosis, which was significantly higher compared to the control group (95.8% vs. 62.5%, P=0.002). Demographics and disease characteristics for each group are presented in Table 1.
2. VCE Findings
For indications of VCE, 11 patients (45.8%) underwent the VCE examination due to abdominal pain with diarrhea, 6 patients (25.0%) with perianal disease, and 5 patients (20.8%) with unexplained iron deficiency anemia. The types of CD-associated small bowel capsule endoscopy findings are presented in Table 2 and Fig. 2. Nineteen patients (79.2%) showed 3 or more ulcers. Luminal stenosis was identified in 6 patients (25.0%), 2 of whom experienced capsule retention. The capsules were presumed to have stagnated in the distal ileum and were passed after medical treatment without any particular intervention. In 1 patient, the capsule was excreted about 10 days after the examination, while in the other patient, it took place 8 months after the examination. The ileum was affected in most cases (87.5%), whereas the jejunum and duodenum were involved in 70.8 and 16.7% of cases, respectively. The median Lewis score was 838 (IQR, 393–1,803).
Capsule Endoscopy Findings of Patients in VCE Group
Video capsule endoscopy findings associated with Crohn’s disease. (A) Villous edema. (B) Ulceration. (C) Stenosis.
3. Clinical Outcomes of Included Patients
The mean follow-up duration after diagnosis of CD was 66.3 months (SD, 48.7) in the VCE group and 70.2 months (SD, 46.7) in the control group, with no significant differences between the groups. The occurrence of complicated behavior was less in the VCE group than in the control group (41.7% vs. 66.7%, P=0.030). In the VCE group, 17 patients (70.8%) used immunomodulators, and 4 (16.7%) used biologics during the follow-up period. The hospitalization rate (55.6% vs. 12.5%, P<0.001) and intestinal resection rate (29.2% vs. 4.2%, P=0.012) were significantly higher in the control group than in the VCE group. In contrast, the risk of perianal surgery did not differ between these groups (8.3% vs. 8.3%, P>0.999). Table 3 summarizes the clinical outcomes of the VCE group and control group.
Comparison of Clinical Outcomes between VCE Group and Control Group
In the VCE group, complicated behavior was observed in 10 patients (7 with perianal disease and 3 with strictures). Of the 3 patients with strictures, 2 had luminal stenosis confirmed during the initial VCE. One patient was on 5-ASA and an immunomodulator and remained stable over a follow-up period of approximately 44 months. Another patient, who used an immunomodulator for 41 months, required segmental resection surgery due to the worsening of small bowel obstruction. The remaining patient initially showed no signs of strictures at diagnosis, but began using adalimumab at 135 months post-diagnosis upon discovering multifocal ileal strictures during a follow-up CTE, subsequently maintaining remission. No patients in the VCE group exhibited locational progression during the follow-up period.
Fig. 3 shows the 10-year cumulative incidence of complicated behavior, use of biologics, CD-related hospitalization, and surgery. Overall, the results suggested that the VCE group patients had a lower risk of developing each outcome than the control group during a follow-up period of up to 10 years (log-rank test, P<0.05 in all outcomes).
(A) Cumulative incidence of complicated behavior in video capsule endoscopy (VCE) and control groups. (B) Cumulative incidence of biologics use in VCE and control groups. (C) Cumulative incidence of Crohn’s disease (CD)-related hospitalization in VCE and control groups. (D) Cumulative incidence of CD-related surgery in VCE and control groups.
4. Follow-up Study after Diagnosis in VCE Group
Out of 24 patients in the VCE group, 5 underwent follow-up VCE. Among them, 4 patients underwent the examination between 6 months and 1 year to assess the treatment response. Of these 4 patients, 3 showed a response to 5-ASA and continued the treatment, whereas 1 patient did not respond to 5-ASA (showing multiple ulcers and erosions at follow-up VCE), resulting in the addition of an immunomodulator. Another patient exhibited disease aggravation 7 years after the diagnosis of CD, which was detected by follow-up VCE and also by ileocolonoscopy and CTE, leading to the initiation of infliximab therapy.
All the patients underwent ileocolonoscopy and cross-sectional imaging every 2–3 years to monitor the progression of the disease. In 4 of 24 patients, new endoscopic findings indicating CD were detected in the terminal ileum during follow-up ileocolonoscopy at a median of 58 months (IQR, 26.3–87.2 months) post-diagnosis. Follow-up enterography revealed new active inflammatory findings consistent with CD in 6 out of 24 patients, with a median follow-up of 37 months (IQR, 26.5–59.0 months) post-diagnosis. In 17 patients, neither ileocolonoscopy nor cross-sectional imaging showed specific findings during the follow-up period.
5. Comparison of Perianal Fistula Characteristics between Groups
Table 4 summarizes the fistula in ano (FIA) characteristics of patients diagnosed with FIA at the time of diagnosis of CD or during follow-up. Between the 2 groups, no significant difference was observed in the rate of FIA diagnosis (29.2% vs. 40.3%, P=0.330). The mean follow-up period was 89.1 months (SD, 57.7) and 83.6 months (SD, 50.0), respectively, with no significant differences between the groups. The rate of FIA surgery, immunomodulator use, and biologics use in each group did not display significant differences. Moreover, there were no significant differences in the recurrence or aggravation of FIA during the follow-up period (1 case [14.3%] vs. 4 cases [13.8%], P>0.999).
Comparison of Perianal Fistula Characteristics between VCE Group and Control Group
DISCUSSION
Our study demonstrated the characteristics and long-term prognosis of patients with CD diagnosed based on VCE, despite negative findings in ileocolonoscopy and cross-sectional imaging. These cases were remarkably rare, and we could collect data on 24 patients from 3,752 patients with CD over approximately 16 years from 3 tertiary hospitals. We established a matching control group of conventionally diagnosed CD patients. The VCE group showed differences in disease location and behavior compared to the control group, and exhibited better long-term disease course during the follow-up period in terms of complicated behavior, use of immunomodulators and biologics, and CD-related hospitalization and surgery.
VCE is a highly sensitive test for detecting small bowel inflammation [13,15]. It is known to be superior in detecting superficial and proximal small bowel lesions compared to cross-sectional imaging [5,6,8,9]. Additionally, VCE allows for direct visualization of small bowel segments inaccessible by ileocolonoscopy. Leighton et al. [16] reported that VCE detected 16% more lesions in the terminal ileum than ileocolonoscopy, indicating that lesions identified by VCE were unreachable by ileocolonoscopy. Therefore, several guidelines have recommended conducting VCE when patients present with clinical features consistent with CD but demonstrate negative findings at ileocolonoscopy or cross-sectional imaging [17-19].
In our study, the majority (91.7%) of the patients in the VCE group had non-stricturing and non-penetrating type of behavior, and 2 patients (8.3%) had a stricturing type and none with a penetrating type at the time of diagnosis. Previous studies have indicated that CTE and MRE have high sensitivity and specificity (over 80%) in detecting CD complications such as stenosis and fistulas [20]. However, 2 patients (8.3%) in this study had stricturing type of lesions despite negative enterography results. Occasionally, limitations in cross-sectional imaging may arise from insufficient small bowel distention by inadequate ingestion of oral contrast agents [21]. In such cases, VCE could be beneficial.
VCE is also advantageous as it is less invasive and has a high safety profile [6,15]. The most common adverse event in VCE is capsule retention. In this study, capsule retention occurred in 2 patients (8.3%). They were not excreted during the observation time of VCE but were naturally excreted after subsequent medical treatment. Results from a recent study in Japan demonstrated capsule retention in 7.4% of patients with documented CD and 6.3% with suspected CD [22]. These findings were similar to that of our present study. Some reports suggest that patency capsules can predict the risk of retention, but this study did not employ it [23]. A careful clinical and radiographical screening of patients can reduce the risk of small bowel retention of the capsule [24]. VCE is limited by the inability to perform biopsies. In the VCE group, 3 patients underwent device-assisted enteroscopy; however, in 2 cases, deep insertion of the enteroscope was difficult due to angulation and the biopsy was non-diagnostic. In the remaining patient, biopsy of the lesion assisted diagnosis. Inflammatory and ulcerative lesions confined to the small intestine require differentiation from other rare conditions like chronic enteropathy associated with SLCO2A1, cryptogenic multifocal ulcerous stenosing enteritis, and small intestine-invasive infections such as tuberculosis. Our diagnosis of CD was based on VCE findings consistent with criteria from previous studies [13,14,25], along with clinical features. However, the limited ability to perform biopsies and the absence of established diagnostic criteria for VCE findings in CD are limitations of our study.
Chateau et al. [11] reported on 13 patients with CD only visible on small bowel capsule endoscopy. The study showed that during a mean follow-up of 27.5 months, none of the patients developed stricturing or penetrating disease, or perianal lesions. None of them required CD-related hospitalization or surgery for CD complications [11]. These findings are similar to the present study wherein the VCE group showed favorable disease prognosis compared to the control group. However, among the 24 patients, 10 exhibited complicated behavior from diagnosis throughout the follow-up period. This is relatively high compared to that reported in previous studies, which possibly could be due to a longer follow-up duration (mean, 66.3 months). Furthermore, 7 of these patients had perianal lesions, aligning with previous reports indicating a higher proportion of perianal disease in Asian IBD patients compared to Western populations [26,27]. Additionally, a retrospective single-center study conducted by McCurdy et al. [25] reported that VCE showed higher detection of luminal inflammation in patients with isolated perianal fistula (IPF) compared to the control group who underwent VCE for alternative indications (26% vs. 3%, P<0.01). These findings suggest that VCE is useful in selected patients with IPF and negative conventional investigations for luminal disease. In our study, 6 patients in the VCE group underwent VCE for perianal disease, which is consistent with the results from that previous study.
A better prognosis in patients diagnosed with VCE might be due to an early diagnosis. Patients in the VCE group notably exhibited lower CDAI, CRP, and fecal calprotectin levels at diagnosis compared to the control group. As CRP or fecal calprotectin has been reported to better reflect colonic inflammation, the results of these biomarkers might be lower than the actual inflammation in small bowel CD [28,29]. Nevertheless, these biomarkers are well-established indicators of mucosal inflammation [30,31]. This implies that diagnosis of CD using VCE is possible even at an early stage when disease activity is not severe. Determining the extent and severity of CD during diagnosis is an important part of the clinical assessment to determine optimal treatment and monitoring strategies [32]. CD can progress with gut inflammation even without symptoms, leading to complicated behavior [33,34]. Therefore, the importance of accurate diagnosis and early intervention is recognized [32,35]. Diagnosing early lesions using VCE might lead to a better prognosis with proper treatment and follow-up.
Patients with proximal small bowel CD involvement have been reported to have a poor prognosis, with more frequent relapse and surgery [36-38]. On the other hand, CD diagnosed using VCE only, primarily involves the proximal small bowel but demonstrates a favorable prognosis, suggesting it as a distinct disease entity from typical CD [11]. However, one needs to be cautious in interpreting these results because of the potential for lead time bias. Early detection of CD and initiating appropriate treatment with follow-up examinations likely influence the prognosis. Nonetheless, in our study, the age at diagnosis of the VCE group was consistent with that of general CD [1], and the time from symptom onset to diagnosis did not differ significantly between the case and control groups. Additionally, both groups comprised of patients who received suitable medical care and treatment from expert physicians at tertiary hospitals post-diagnosis. These factors could diminish lead time bias and suggest that patients diagnosed based on VCE might have a better prognosis.
In the VCE group, follow-up VCE was conducted only in 5 patients. Among them, 3 patients showed a treatment response and maintained their treatment, while 2 patients required treatment escalation due to disease aggravation. All patients underwent regular follow-up examinations with ileocolonoscopy and cross-sectional imaging. Among these follow-up examinations, active inflammatory findings consistent with CD were observed in 6 out of 24 patients on enterography and 4 out of 24 patients on ileocolonoscopy. The results of this study demonstrated the progressive nature of CD [33]. Additionally, given that the lesions were detected exclusively by VCE at diagnosis, follow-up using only enterography or ileocolonoscopy may be inadequate for accurately monitoring the course of the lesions. It is difficult to determine conclusively that the disease has not progressed even if no active inflammation is observed on CTE/MRE and ileocolonoscopy. Appropriate follow-up examinations are necessary even in CD patients with lesions initially detected only by VCE.
Our study has several limitations. First, it has a small sample size and is of a retrospective study design. Although the patients solely diagnosed by VCE were rare [11,39], we have confirmed the long-term disease course of patients collected from 3 hospitals. Second, different endoscopists at each hospital analyzed VCE images, which could have introduced some interpretational differences. Nevertheless, the participating physicians were experts in their field and adhered to the Korean guidelines of VCE for examination and interpretation [40]. Third, this study was conducted at tertiary hospitals, possibly introducing selection bias in patient recruitment. Additionally, the control group comprised patients matched from the registry of a tertiary center, possibly representing individuals with severe disease severity compared to the overall CD patients. Nevertheless, our study benefitted from a direct comparison between a matched control group used for analyzing the long-term prognosis of rare cases.
In conclusion, patients diagnosed with CD exclusively detectable on VCE were rare, comprising less than 1% of the patient population. These patients presented different disease characteristics at diagnosis and had a favorable long-term disease course compared to conventionally diagnosed matched CD patients. However, some patients experienced disease progression, emphasizing the need for regular follow-up examinations. The results of this study suggested that CD diagnosed solely by VCE represents a distinctive disease entity. These results might be beneficial in diagnosing and monitoring such rare cases in clinical practice.
Notes
Funding Source
This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (No. 2023R1A2C2005817 and No. 2021R1A5A2021614) and a grant (No. 2023IT0006) from the Asan Institute for Life Sciences, Asan Medical Center, Seoul, Korea.
Conflict of Interest
Myung SJ is an editorial board member of the journal but was not involved in the peer reviewer selection, evaluation, or decision process of this article. No other potential conflicts of interest relevant to this article were reported.
Data Availability Statement
All data, analysis methods, and study materials relevant to the study are included in the article or are available upon request from the corresponding authors, Park SH (16th) and Kim ES.
Author Contributions
Conceptualization: Bae JH, Park SH (2nd), Kim ES, Park SH (16th). Data curation: Bae JH, Park SH (2nd), Park JB, Baek JE, Hong SW, Hwang SW, Yang DH, Kim KO, Lee MR. Formal analysis: Bae JH, Park SH (2nd), Park JB, Baek JE, Hong SW, Hwang SW, Yang DH, Kim KO, Lee MR. Funding acquisition: Kim ES, Park SH (16th). Investigation: Bae JH, Park SH (2nd), Kim ES, Park SH (16th). Methodology: Bae JH, Park SH (2nd), Kim ES, Park SH (16th). Supervision: Ye BD, Byeon JS, Myung SJ, Yang SK, Jang BI. Writing – original draft: Bae JH, Park SH (2nd). Writing – review & editing: Kim ES, Park SH (16th). Approval of final manuscript: all authors.