Branch pulmonary artery (PA) stenosis often accompanies congenital heart diseases, and left pulmonary artery (LPA) stenosis is a common cause of reoperation after surgical repair of tetralogy of Fallot (TOF) [1-3]. LPA stenosis is also frequently detected in patients with TOF who undergo reoperation for other reasons, such as pulmonary valve (PV) implantation for pulmonary regurgitation (PR) or pulmonary angioplasty for pulmonary stenosis [1,4]. However, whether the anatomical improvement of LPA stenosis by surgical angioplasty leads to an actual improvement in perfusion to the ipsilateral lung is unclear, especially in adolescent or adult patients whose somatic growth has almost ended. We evaluated the effect of branch PA angioplasty on perfusion to the ipsilateral lung and its predictive factors in adolescents and adults.

The study protocol was approved by the institutional review board of Seoul National University Hospital (approval no., H-2111-009-1257), and all procedures were performed in accordance with our institutional guidelines to protect patient confidentiality. The requirement for patient consent was waived because of the retrospective nature of the study.

Patients and data

This was a retrospective review of patients who underwent surgical angioplasty for LPA stenosis between January 2004 and June 2019. Preoperative and postoperative perfusion rates in the left lung (p-left) were obtained using lung perfusion scanning (LPS) or cardiac magnetic resonance imaging (cMRI). Only patients aged ≥13 years at the time of pulmonary angioplasty with both preoperative and postoperative information on p-left were included. Patients with a preoperative p-left of ≥40% were excluded, even if they underwent LPA angioplasty for anatomical stenosis, because this study investigated the effect of LPA angioplasty on the improvement of a significantly decreased preoperative perfusion ratio to the ipsilateral lung. Anatomical information regarding PA size was mainly obtained from computed tomography (CT) or cMRI in patients without available CT data.

Surgical techniques

Various surgical techniques, including patch angioplasty, cutback angioplasty, intraluminal tissue excision, and angle correction techniques, were used alone or in combination with LPA angioplasty. The surgical techniques were selected according to the anatomical characteristics of the LPA causing the stenosis and the preference of each surgeon.

Statistical analysis

All continuous variables are expressed as means±standard deviations (SDs) or medians with interquartile range (IQR) values, as appropriate. Categorical variables are expressed as frequencies and percentages. Continuous variables were compared using the Student t-test, Mann-Whitney U test, paired t-test, or Wilcoxon signed-rank test. Statistical significance was set at p<0.05. All analyses were performed using IBM SPSS ver. 23.0 (IBM Corp., Armonk, NY, USA).

Patients and overall findings

During the study period, 634 LPA angioplasties were concomitantly performed with other cardiovascular procedures. Among them, 34 met the inclusion criteria (age range and both preoperative and postoperative imaging data). One patient who demonstrated progressive right PA stenosis requiring additional postoperative intervention was excluded because contralateral stenosis could ultimately affect ipsilateral perfusion. Thus, 33 patients were included in this study. Among the 33 patients, 17 were adolescents (age range, 13–18 years) [5,6] and 16 were adults at the time of pulmonary angioplasty. The median age at the time of operation was 17 years (IQR, 14.5–25.5 years), and the median follow-up duration after LPA angioplasty was 72 months (IQR, 14.5–143 months). The original diagnoses were TOF (n=29, 87.9%), pulmonary atresia with ventricular septal defect (n=1, 3.0%), transposition of the great arteries (n=1, 3.0%), congenital aortic stenosis (n=1, 3.0%), and coarctation of aorta with ventricular septal defect (n=1, 3.0%). There was no patient who showed pulmonary hypertension in the preoperative evaluation. In all patients, LPA angioplasty was concomitantly performed with procedures on the right ventricular outflow tract (RVOT) including PV repair or replacement, or Rastelli conduit change 16.3 years (median, IQR 13.1–24.5 years) after total correction of congenital heart disease. All patients had a left aortic arch. Three of the TOF patients had patent ductus arteriosus (PDA). Two of them underwent PDA double ligation, while the other underwent division. Seven patients underwent a Blalock-Taussig shunt prior to total repair, whereas the remaining 22 TOF patients underwent primary total repair. Excluding 5 patients, the remaining 24 patients underwent LPA angioplasty during total repair. Among the remaining 5 patients, 4 received LPA angioplasty once before total repair, and 1 patient underwent LPA angioplasty twice. During the follow-up period after surgery, 1 adolescent and 1 adult patient underwent balloon angioplasty for stenotic LPA. No cases of early or late mortality occurred. The patients’ clinical characteristics are presented in Table 1.

Table 1 . Clinical characteristics of the patients.

Characteristic	Overall (n=33)	Adolescent (n=17)	Adults (n=16)
Sex
Male	21 (63.6)
Female	12 (36.4)
Age at total correction (yr)	0.8 (0.4–2.1)	0.45 (0.40–0.48)	0.8 (0.62–0.87)
Age at LPA angioplasty (yr)	17.0 (14.5–25.5)	14.0 (12.7–15.3)	25.1 (19.5–27.4)
Surgical interval (yr)	16.3 (13.1–24.5)	13.6 (12.2–14.8)	24.3 (18.9–26.5)
Follow-up duration (mo)	72.0 (14.5–143.0)	112.0 (58.0–156.0)	19.0 (11.0–97.0)
Primary diagnosis
Tetralogy of Fallot	29 (87.9)
Pulmonary atresia with VSD	1 (3.0)
Transposition of great arteries	1 (3.0)
Congenital AS	1 (3.0)
CoA with VSD	1 (3.0)

Values are presented as number (%) or median (interquartile range)..

LPA, left pulmonary artery; VSD, ventricular septal defect; AS, aortic stenosis; CoA, coarctation of aorta..

Perfusion ratio to the left lung: overall

In the entire cohort, p-left significantly increased postoperatively (24.3%±10.0% in preoperative evaluations versus 29.9%±9.9% in postoperative evaluations, p=0.005). The preoperative p-left values demonstrated no significant difference between the 2 age groups (24.6%±8.8% in adolescents and 24. 0%±11.4% in adults, p=0.868), whereas the postoperative p-left values were higher in the adolescent group (33.5%±6.6%) than in the adult group (26.2%±11.6%, p=0.038). The p-left increased in both age groups after LPA angioplasty; however, the change was statistically significant only in the adolescent group (p=0.003 in adolescents; p=0.433 in adults) (Table 2).

Table 2 . Preoperative and postoperative left lung perfusion (p-left).

Variable	Overall (n=33)	Adolescents (n=17)	Adults (n=16)	p-value
Preoperative (%)	24.3±10.0	24.6±8.8	24.0±11.4	0.868
Postoperative (%)	29.9±9.9	33.5±6.6	26.2±11.6	0.038
p-value	0.005	0.003	0.433

Values are presented as mean±standard deviation unless otherwise stated..

Perfusion ratio to the left lung according to the anatomical characteristics of the LPA

CT and cMRI data were available for 30 patients (15 adolescents and 15 adults). We divided the patients into 4 groups according to the preoperative anatomical characteristics of the LPA (Table 3), particularly according to hilar stenosis. Patients with small LPA diameter at the hilum (z-score <−2.0) were classified into group I (n=7) (Fig. 1A); patients with normal LPA diameter at the hilum (−2.0< z-score <2.0) and no significant focal stenosis (z-score <−2.0) but diffuse narrowing from the LPA before the hilum were classified into group IIa (n=11) (Fig. 1B); patients with normal LPA diameter at the hilum (−2.0< z-score <2.0) and presence of significant focal stenosis from LPA proximal to the hilum (z-score <−2.0) were classified into group IIb (n=7) (Fig. 1C); and patients with enlarged LPA (z-score >2.0) at the hilum were classified into group III (n=5) (Fig. 1D) regardless of the characteristics of proximal stenosis. The changes in the z-score of the LPA diameter at the hilum and narrowest site before the LPA, and the ratio of perfusion to the left lung in each group are presented in Table 4. The diameters at the narrowest site of LPA before the hilum increased significantly in group IIa (p=0.017) and IIb (p=0.028); however, the diameters at the hilum did not show significant changes in any group. The improvement in p-left was statistically significant only in group IIb (p=0.018).

Figure 1. Preoperative computed tomography images representing the anatomical characteristics of left pulmonary artery (LPA) in 4 groups. (A) Diffuse small LPA with a small hilar diameter (z-score <-2.0). (B) Relatively small LPA, but no significant focal stenosis, with a z-score less than -2.0. (C) Significant stenosis (z-score <-2.0) at the proximal LPA, but normal hilar diameter. (D) Proximal LPA stenosis with hilar dilatation (z-score >2.0).

Table 3 . Four groups according to the preoperative anatomical characteristics of LPA.

	Diameter of LPA at the hilum
	z-score <-2.0	-2.0< z-score <2.0		z-score >2.0
		Absence	Presence
Significant focal stenosis at the LPA proximal to hilum (z-score <-2.0)	Group I (n=7)	Group IIa (n=11)	Group IIb (n=7)	Group III (n=5)

LPA, left pulmonary artery..

Table 4 . Preoperative and postoperative z-scores of the LPA diameter and perfusion ratio to left lung in 4 groups.

LPA diameter z-score	Group I (n=7)	Group IIa (n=11)	Group IIb (n=7)	Group III (n=5)
Narrowest site proximal to the hilum
Preoperative	-0.37 (-2.75 to 1.49)	-1.13 (-1.67 to 0.79)	-3.12 (-4.59 to -2.53)	0.20 (-1.44 to 0.92)
Postoperative	0.30 (-1.16 to 1.60)	0.83 (-0.57 to 2.32)	-0.72 (-1.87 to 0.03)	0.43 (-0.81 to 1.41)
p-value	0.310	0.017	0.028	0.50
Hilum
Preoperative	-2.31 (-6.13 to -2.13)	0.38 (-1.04 to -0.70)	-0.87 (-1.66 to 0.58)	2.77 (2.23 to 3.43)
Postoperative	-2.07 (-2.96 to 0.87)	0.49 (-0.95 to 1.33)	0.49 (-0.99 to 0.69)	1.80 (1.34 to 2.49)
p-value	0.128	0.499	0.116	0.14
P-left
Preoperative	19.1±7.6	28.4±8.9	18.3±6.2	34.6±5.2
Postoperative	18.9±7.6	31.1±7.5	35.5±3.8	35.0±11.9
p-value	0.799	0.248	0.018	0.686

Values are presented as median (interquartile rage) or mean±standard deviation for continuous variables. The Mann-Whitney test and Wilcoxon signed rank test were used to compare continuous variables. p<0.05 was considered statistically significant..

LPA, left pulmonary artery; P-left, perfusion ratio to left lung..

Patients with remarkable improvement

When we defined a postoperative p-left >40% in absolute terms or a ≥9% increase in p-left as a “remarkable improvement,” 9 of 30 patients with preoperative CT or cMRI data demonstrated remarkable improvement. None of the 7 patients in group I (0%), 1 of the 11 patients in group IIa (0.1%), 6 of the 7 patients in group IIb (85.7%), and 2 of the 5 patients in group III (40.0%) demonstrated remarkable improvement. Remarkable improvement was observed in both adolescent patients in group III (2/2, 100.0%), whereas none of the adult patients in group III demonstrated remarkable improvement (0/3, 0. 0%) (Fig. 2). Patients in group IIb and adolescent patients in group III were more likely to exhibit remarkable improvement, with high positive and negative predictive values (88.9% and 95.2%, respectively).

Figure 2. The numbers of adolescent and adult patients with remarkable improvement in 4 groups.

Difference in the perfusion ratio to the left lung between LPS and cMRI

Overall, 30 patients had both LPS and cMRI data during the same period (patients with any preoperative or postoperative examination). In patients with a PR grade ≥ moderate, the mean difference of p-left values between LPS and cMRI was 9.7%±5.5%. In patients with a PR grade ≤mild, the mean difference of p-left values between LPS and cMRI was 4.6%±3.1%. The difference was significantly greater in patients with PR grade ≥moderate than in patients with PR grade ≤mild (p=0.005).

Subgroup analysis: the independent effect of PA angioplasty based on cMRI

Regarding the data on PA flow, cMRI can provide the amount of forward flow, regurgitant flow, and the difference between them—in other words, net forward flow. As mentioned earlier, when patients had no significant PR, no significant difference was observed between forward flow and net forward flow, and LPS and cMRI had similar ratios between the right and left lung perfusion or flow. To evaluate the effect of pulmonary angioplasty on lung perfusion in adolescents and adults, we analyzed the cMRI data of 20 patients who underwent both preoperative and postoperative cMRI. To exclude the effect of preoperative PR on the LPS, the fraction of forward flow combining net forward flow and regurgitant flow of the left pulmonary perfusion, rather than net forward flow, was used as an index corresponding to p-left on preoperative MRI. Of 20 patients with preoperative cMRI, 18 had a PR grade ≥moderate preoperatively, and the PR was corrected by PV repair or PV replacement postoperatively. In the subgroup analysis, the preoperative and postoperative p-left values were 26.8%±11.0% and 32.9%±10.8%, respectively (p=0.031). Hence, PA angioplasty alone was able to increase left lung perfusion. When we divided the 20 patients into adult (n=11) and adolescent (n=9) groups, this change was statistically significant only in the adolescent group (p=0.021 in adolescents versus p=0.328 in adults; Wilcoxon signed-rank test).

Branch PA stenosis often accompanies congenital heart diseases, such as TOF. Although previous reports have demonstrated excellent outcomes of LPA angioplasty at the time of total repair of TOF [7,8], many patients still have LPA stenosis after TOF repair. In cases where percutaneous interventions, such as ballooning or stenting, do not work well for these patients, surgical treatment is required independently or concomitantly with other procedures, such as PV implantation [4,6,9-11]. Most patients have reached adolescence or adulthood by the time when they require RVOT reoperations; however, whether anatomical improvement of LPA stenosis by concomitant surgical angioplasty leads to actual improvement of perfusion to the ipsilateral lung in this age group remains unclear.

Characteristics of perfusion improvement according to LPA anatomy

Our study demonstrated that LPA angioplasty significantly improved perfusion of the ipsilateral lung in adolescents. Although the improvement in perfusion in the adult group was not significant overall, a non-negligible improvement in perfusion was still observed in some subgroups of adults. We identified the characteristics of these adult and adolescent patients who exhibited improved perfusion. Since the average increase of p-left in the adolescent patients was approximately 9% in our study, we defined an increase of p-left of >9% as a remarkable improvement. Moreover, considering a postoperative p-left of ≥40% as indicating remarkable improvement is reasonable because a right-to-left lung perfusion ratio of 55/45 is considered normal, and a 60/40 split is often considered to be within the normal range in the clinical setting [12]. Hence, we identified common anatomical features in patients demonstrating remarkable improvement, focusing on the LPA, and divided the cohort into 4 groups (Table 3). Three adult patients demonstrated remarkable improvement; 2 belonged to group IIb, and the other belonged to group. Among the 9 patients showed remarkable improvement, improvement was observed in their preoperative and postoperative LPS studies in 6 patients and in their cMRI studies in 2 patients. We compared preoperative LPS and postoperative cMRI in 1 patient because of the lack of the corresponding study. Among these patients, 5 underwent both postoperative LPS and cMRI, and the p-left values for LPS and cMRI in these 5 patients showed no significant difference.

Group I (small hilar diameter)

Patients with a small LPA diameter (z-score <−2.0) at the hilum were classified into group I. A small LPA diameter at the hilum may reflect poor peripheral, intraparenchymal pulmonary vasculature distal to the hilum, which is unlikely to be surgically corrected by LPA angioplasty. The improvement in p-left was not significant in group I, and none of the 7 patients in this group demonstrated remarkable improvement. The apparent absence of remarkable results in both adolescents and adults within group I may be attributed to the diffuse diminution in size observed to the hilar level. Specifically, the reduced size of pulmonary arteries at the pulmonary parenchymal level, distal to the hilar level, likely leads to smaller z-scores than those observed at the level proximal to the hilum. Indeed, upon calculating the Nakata index for the patients in group I, the results showed an average of 93.85±20.32, suggesting that, according to Nakata index reference values, achieving reversibility might be challenging in both adolescent and adult populations.

Group II (normal hilar diameter)

Patients with a normal LPA diameter (−2.0< z-score <2.0) at the hilum and without significant focal stenosis (z-score <−2.0) from the LPA proximal to the hilum, but demonstrating diffuse narrowing, were classified into group IIa. The perfusion decrease in the left lung despite the anatomically mild stenosis of the LPA in this group may be possibly caused by multiple factors, including poor peripheral pulmonary vasculature distal to the hilum, as well as a small LPA size. Although the anatomical size of the LPA was significantly increased by surgical angioplasty, the improvement in p-left was not significant in group IIa, and none of the 6 adolescent patients and only 1 of the 5 adult patients in this group demonstrated remarkable improvement.

Patients with a normal LPA diameter (z-score between −2.0 and 2.0) at the hilum and focal significant stenosis (z-score <−2.0) in the middle of the LPA (from proximal to the hilum) were classified into group IIb. In this group, the perfusion decrease in the left lung would be expected to improve if the focal stenosis of the LPA is resolved, which is the main cause of the perfusion decrease in the ipsilateral lung. Focal stenosis of the LPA was significantly resolved anatomically, and p-left also increased significantly in group IIb. Four of the 5 adolescent patients and both adult patients in this group demonstrated remarkable improvement. Belonging to group IIb was considered a positive predictive factor for remarkable improvement. A previous study on TOF reported a better left lung perfusion rate after LPA angioplasty in patients with focal LPA stenosis than in those with diffuse LPA stenosis [7]. This finding has similar implications to our study, despite the differences in patient populations between the 2 studies.

Group III (enlarged hilar diameter)

Patients with an enlarged LPA (z-score >2.0) in the hilum were classified into group III. These patients underwent LPA angioplasty because of the presence of stenosis at the orifice or middle of the LPA, although the degrees varied. Dilatation of the distal LPA at the hilum can be interpreted as post-stenotic dilatation resulting from prolonged exposure to stenosis of the proximal LPA. The improvement in p-left was not significant in group III, and none of the 3 adult patients in this group demonstrated remarkable improvement. However, both the adolescents in this group demonstrated remarkable improvement. Although all adolescents in group III exhibited remarkable improvement, the small number of patients in this age group (n=2) might have weakened the statistical power, and the improvement in the p-left in these 2 patients was not statistically significant (p=0.180). The decrease in perfusion in the left lung with proximal LPA stenosis, which is severe and sufficiently prolonged to dilate the distal LPA, may cause underdevelopment of peripheral pulmonary vascularity. Adult patients with longer periods of exposure to such situations may already have poor peripheral pulmonary vascularity and may not exhibit remarkable improvement, whereas adolescent patients with relatively shorter periods of exposure may have preserved peripheral vascularity and demonstrate similar features to those in group IIb. Belonging to group III might be considered a positive predictive factor for remarkable improvement in adolescent patients.

Using the normal LPA diameter (z-score between −2.0 and 2.0) at the hilum with focal significant stenosis (z-score <-2.0) of the LPA before the hilum in all ages and enlarged LPA (z-score >2.0) at the hilum in adolescent patients to predict remarkable improvement, the positive and negative predictive values were 88.9% and 95.2%, respectively, which are considered highly acceptable. However, this does not necessarily mean that LPA angioplasty is not useful for improving perfusion in patients who do not meet these criteria, because the criteria in this study are only used to predict remarkable improvement, which was defined as postoperative p-left >40% or a 9% or more increase in p-left. Determining what constitutes a clinically meaningful increase in p-left is challenging. Even an increase of less than 9% in p-left might be considered clinically significant in patients with a severe preoperative decrease in perfusion to the left lung. Considering that the previously reported SD of the p-left in the normal population was 2.1% [12], only a 4.2% (twice the SD) or greater increase in the p-left might be accepted as a non-negligible improvement.

Evaluation of the perfusion ratio to the left lung

LPS is considered the gold standard for quantitative evaluation of pulmonary perfusion in most patients with congenital heart disease [13,14]. Recently, cMRI has also been used for quantitative measurements of pulmonary blood flow distribution in patients with congenital heart disease [14,15]. Although many previous studies have reported excellent correlations between LPS and cMRI data measuring pulmonary blood flow distribution [14,16,17], the accuracy of cMRI has been reported to be affected by factors associated with PV function or PA status. Roman et al. [14] noted that in the presence of branch PA stenosis, turbulent flow at the stenosis site can produce errors in the cMRI data. Wu et al. [18] reported a discrepancy between the LPS and cMRI data on p-left in the presence of PR. Our study revealed a similar result; the discrepancy between LPS and cMRI data for p-left was significantly greater in patients with PR grade ≥moderate than in those with PR grade ≤mild. When a patient has PR, a difference between LPS and cMRI data can be expected because lung perfusion measured by cMRI is obtained based on the net forward flow, whereas pulmonary perfusion in LPS is obtained by Tc 99m macroaggregated albumin (Tc-MAA). The mechanism of Tc-MAA localization in pulmonary capillaries is capillary blockade, which typically results in the microembolization of hundreds of thousands of capillaries in the lungs of adults. When we consider that the lung perfusion ratio is obtained during 3–5 respiratory cycles, the LPS appears to reflect forward flow summation from a couple of cardiac cycles, rather than net forward flow in a cardiac cycle (difference between initial forward flow and regurgitant flow), as in cMRI. We speculate that the patient’s PR status influences the observed difference in perfusion ratios between LPS and cMRI.

Therefore, when patients have more than moderate PR, we suggest using LPS to evaluate and compare the ultimate preoperative and postoperative perfusion status of the lungs.

Pulmonary vascular bed maturation versus anatomy

According to the results of a previous study by Ochs et al. [19], the mean alveolar size remains unchanged after the age of 18. Alveolar duct arteries mature around 18 months postnatally, but microvascular maturation at the alveoli level and capillaries continues until the age of 21. As a result, in adults with completed pulmonary bed maturation, it is believed that there will be little variability in vascular bed resistance, and correcting distal anatomical stenosis may not have a significant impact on perfusion changes. Our question was whether surgical pulmonary angioplasty could be effective even in adult patients with pulmonary arterial stenosis whose pulmonary vascular bed and alveolar maturation seemed to have already been completed in terms of perfusion. After performing an analysis of diameter changes at the proximal LPA ostium and hilar level following LPA angioplasty, we observed a tendency for a larger diameter at the hilar level in the adolescent group (Table 5). This is likely attributed to the less mature development of the pulmonary vascular bed in adolescents.

Table 5 . LPA diameter changes.

	Overall (n=18)	Adolescents (n=6)	Adults (n=12)
LPA os (delta) (mm)	2.52±2.74	2.53±3.46	2.52±2.48
LPA hilum (delta) (mm)	0.61±2.31	1.39±2.61	0.22±2.15

Values are presented as mean±standard deviation for continuous variables..

LPA, left pulmonary artery; os, proximal ostium..

In our investigation, we conducted an analysis based on the morphological classification of anatomical stenosis. Even in adults, our findings suggest that if focal stenosis is confined solely to the proximal LPA, performing PA angioplasty can lead to an increase in ipsilateral lung perfusion.

Limitations

First, this was a retrospective study with a relatively small cohort. The small number of patients made it difficult to perform a statistical analysis and weakened the statistical power, especially in the subgroup analysis.

Information on the preoperative and postoperative perfusion ratio to the left lung from the same modality (LPS or cMRI) was used in this study if available, and data from LPS were used if data from both LPS and cMRI were available preoperatively and postoperatively. However, only preoperative cMRI and postoperative LPS were available for 7 patients. The comparison of perfusion rates obtained from different modalities may confuse the results of this study, considering the limitations of cMRI in the evaluation of perfusion rates to the lungs, particularly in cases of preoperative uncorrected PV function and branch PA stenosis. We performed a subgroup analysis of patients with both preoperative and postoperative cMRI data to overcome this limitation, and the results were as expected.

This study did not include data on patients’ functional status, such as the New York Heart Association (NYHA) functional class or exercise pulmonary function testing (ePFT). All patients in this study belonged to NYHA class I or II preoperatively and there was little room for improvement after surgical treatment. Very few patients underwent ePFT before and after surgery. Additionally, since LPA angioplasty was performed concomitantly with other procedures, such as PV repair or replacement and Rastelli conduit change, it is difficult to determine whether the improvement in functional status is due to the improvement in perfusion decrease in the left lung. However, for the improvement of perfusion decrease to the left lung to be clinically meaningful, further studies on whether it actually helps improve patients’ functional status are necessary.

Conclusion

In conclusion, PA angioplasty leads to a significant increase in perfusion in the ipsilateral lung in adolescent patients, but not in adult patients. Focal stenosis confined to the proximal LPA is a positive predictive factor for increased perfusion after PA angioplasty, regardless of age. We assume that post-stenotic dilatation of the distal LPA could also be a positive predictive factor for a better perfusion increase after PA angioplasty in adolescent patients, and we suggest that PA angioplasty should not be delayed in patients with this anatomical feature.

Author contributions

Conceptualization: JGK, WHK, SC. Data curation: JM, DHS, JGK. Formal analysis: DHS, JM, JGK. Methodology: JGK, JM. Project administration: JGK. Visualization: JM, DHS. Writing–original draft: JM, DHS. Writing–review & editing: JM, DHS. Final approval of the manuscript: all authors.

Conflict of interest

No potential conflict of interest relevant to this article was reported.

Funding

This study was supported by research grant from Seoul National University College of Medicine (grant No. 800-20230598).