Young-In Yoon, Sung-Gyu Lee ,?, Deok-Bog Moon , Shin Hwng , Ki-Hun Kim,Hui-Ju Kim, Ki-Hoon Choi

a Division of Hepatobiliary Surgery and Liver Transplantation, Department of Surgery, Asan Medical Center, University of Ulsan College of Medicine, 388-1 Poongnap-dong, Songpa-gu, Seoul 138-736, Korea

bDepartment of Surgery, Gold Coast University Hospital, Southport, QLD 4215, Australia

Keywords:Left hepatic trisectionectomy Major hepatectomy Liver neoplasms Survival rate Morbidity

ABSTRACT Background: Despite remarkable advances in surgical techniques and perioperative management, left hepatic trisectionectomy (LHT) remains a challenging procedure with a somewhat high postoperative morbidity rate compared with less-extensive resections.This study aimed to analyze the short- and long-term outcomes of LHT and identify factors associated with the postoperative morbidity of this technically demanding surgical procedure.

Introduction

Left hepatic trisectionectomy (LHT) was initially described as a left hepatic trisegmentectomy in 1982 by Starzl et al.[1]and was defined as the excision of Couinaud segments II, III, IV, V, and VIII with or without segment I by the Brisbane 20 0 0 Terminology Committee of the International Hepato-Pancreato-Biliary Association[2].Although this procedure allows for curative resection with negative margins of large left-sided or central advanced tumors that extend to the structures of the right anterior Glissonean pedicle, it remains the most challenging procedure of the major anatomical hepatectomies with a high postoperative morbidity rate [3,4].

As LHT removes approximately 65%–70% of the total liver volume, it is one of the most extensive types of hepatectomy and requires an accurate preoperative evaluation of liver function and future remnant liver volume (FRLV) to avoid post-hepatectomy liver failure.Preoperative biliary decompression and portal vein embolization (PVE) for compensatory hypertrophy of the posterior sector are mandatory to avoid postoperative liver failure when performing LHT [5,6].

Despite remarkable advances in surgical techniques and perioperative management, LHT is considered a complicated proce-dure because of the anatomical variations of the right portal triad structure [7,8].Although recent advances in preoperative diagnostic imaging have helped surgeons overcome this difficulty, the anatomical variations of the right posterior hepatic artery (HA) and bile duct (BD) relative to the right portal vein (PV) can result in misidentification, which is linked to postoperative morbidity[9].

Until recently, few studies have described the short- and longterm clinical outcomes of more than 50 patients undergoing LHT.Our study includes the data of 53 consecutive patients who underwent LHT for various indications over a 14-year period.Accordingly, this study aimed to analyze the short- and long-term outcomes and identify factors associated with postoperative morbidity.

Patients and methods

Patients

The medical records of 53 patients who underwent LHT between June 2005 and October 2019 at the Asan Medical Center,University of Ulsan College of Medicine (Seoul, Korea) were reviewed.This study was approved by the Institutional Review Board of the Asan Medical Center (2020-0358), and the requirement for informed consent was waived due to the retrospective nature of the study.

Preoperative evaluation

The preoperative radiological assessments performed in all patients included computed tomography (CT) and magnetic resonance imaging of the liver to define the anatomy and assess the extent of the tumor.Preservation of the preoperative liver function and safety of the hepatectomy were carefully determined using volumetric studies, biochemical assays (total bilirubin, coagulation profile, and indocyanine green clearance test), and the Child-Pugh classification.The presence of portal hypertension (esophageal varix, noticeable collaterals, and splenomegaly with thrombocytopenia) was also assessed, together with a fibroscan of the liver parenchyma.The cardiovascular and pulmonary risks of surgery were evaluated using electrocardiography, echocardiography, and pulmonary function testing.

Perioperative management

PVE was performed preoperatively to increase the FRLV based on the remnant liver volume observed on CT scan with a local picture archiving and communication system monitor and digital imaging and communications in medicine image-viewing software(PetaVision; Asan Medical Center, Seoul, Korea).PVE was indicated to patients identified as having an FRLV of less than 40%, and was performed on the left lobe in most patients.If the anterior sector volume was large, PVE was also performed on the right anterior portal branch of the PV.The FRLV was reevaluated via CT scans obtained 3–4 weeks after PVE.In patients with good liver function without cirrhosis and generally acceptable performance, LHT was performed without PVE, even if the FRLV remained at less than 40% in patients<50 years old.

Patients with obstructive jaundice with cholangitis underwent biliary decompression of the posterior sector via percutaneous transhepatic biliary drainage (PTBD) and/or endoscopic retrograde biliary drainage (ERBD).Surgery was scheduled when the serum bilirubin level was less than 2 mg/dL.

Surgical technique

Using an inverted T or reverse L incision, the xiphoid process was excised to facilitate the exposure of the entrance of the hepatic veins (HVs).The abdomen was thoroughly examined for peritoneal tumor seeding, and the intrahepatic tumor extent and liver anatomy were further examined using intraoperative ultrasonography.A no-touch approach was used in patients with hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (CCA) to minimize the intraoperative spread of the tumor.

Prior to the mobilization of the liver, the inflow to the liver was controlled using the Glissonean approach or individual hilar dissection, depending on the disease or anatomical findings of the vascular and biliary structures.The liver parenchyma was divided using the combined Kelly clamp crushing technique and a Cavitron Ultrasonic Surgical Aspirator (Excel; Integra LifeSciences Co., Plainsboro,NJ, USA) along the right HV, and the common trunk of the middle and left HVs was stapled and divided (Fig.1).During the parenchymal resection, the repeated Pringle maneuver was used for intervals of 15 min, with a resting period of 5 min until completion of the transection.When needed, caudate lobe resection, extrahepatic BD resection, and/or lymphadenectomy were performed during the surgery.Biliary reconstruction was performed using the Roux-en-Y method when a concomitant extrahepatic BD resection was performed.

Statistical analysis

Continuous variables are expressed as mean ±standard deviation (SD) or median (range) as appropriate, and categorical variables are presented as number (percentage).The risk of postoperative morbidity was analyzed using the logistic regression model to determine independent predictors of outcomes, and multivariate analyzes included prognostic factors withP<0.05 in the univariate analysis.Further, the Firth’s bias correction method was used.Patient survival was estimated using the Kaplan-Meier method and log-rank test.Significance was set atP<0.05.All statistical analyzes were performed using IBM SPSS Statistics for Windows, version 21.0 (IBM Corp., Armonk, NY, USA).

Results

Preoperative data and surgical procedures

A total of 53 patients who underwent LHT were included in this study.The most common indication for LHT was HCC (n= 21;39.6%), followed by hilar CCA (n= 14; 26.4%), intrahepatic CCA(n= 10; 18.9%), and other pathologies (n= 8; 15.1%), including colorectal liver metastasis, hepatolithiasis, gallbladder cancer,live donor, hemangioma, and a multilocular biliary cyst.Fortyseven patients (88.7%) had solitary tumors.The size of the tumor in 47 patients with solitary tumors was 7.98 ±14.60 cm.The size of the tumor according to each disease was as follows: HCC(10.33 ± 4.62 cm), hilar CCA (3.39 ± 4.68 cm), and intrahepatic CCA (7.10 ±4.90 cm).

Of the 53 patients, 35 were men and 18 were women.The median age was 53 years (25–75).Seven patients had an American Society of Anesthesiologists (ASA) fitness grade of I; 41, ASA fitness grade of II; and 5, ASA fitness grade of III.Preoperative PVE was conducted in 21 (39.6%) patients, including one (1.9%) diagnosed with intrahepatic CCA who underwent subsequent HV embolization followed by PVE to increase the FRLV.Preoperative anti-tumor treatment was performed in nine patients [transcatheter arterial chemoembolization (TACE),n= 9]with a diagnosis of HCC, one(systemic chemotherapy,n= 1) with a diagnosis of hilar CCA, and five (TACE,n= 4; systemic chemotherapy,n= 1) with a diagnosis of intrahepatic CCA.Preoperative biliary drainage was required in 19 patients (PTBD only,n= 6; ERBD only,n= 10; both PTBD and ERBD,n= 3).

A concomitant segment I resection, extrahepatic BD resection,or lymphadenectomy were performed in 29 (54.7%), 22 (41.5%),and 25 (47.2%) patients, respectively.All patients with hilar CCA underwent concomitant resections.The combined resection and reconstruction of the PV and HA were performed in one hilar CCA and two intrahepatic CCA patients.Extended LHT with resection of the involved right HV was performed in one patient with large infiltrating HCC and one patient with intrahepatic CCA because of the presence of a dominant inferior right HV.The pathology and surgical procedures performed combined with LHT are presented in Table 1.

Fig.1.Representative case.A: Computed tomography showed a 4.7-cm irregular tumor (arrowheads) involving the right anterior portal vein; B: the right anterior portal vein and left portal vein were embolized with an Amplatzer Vascular plug-II?; C: surgical field after left hepatic trisectionectomy.The right hepatic vein was clearly exposed; D:postoperative computed tomography of the patient 6 months after left hepatic trisectionectomy.RAPV: right anterior portal vein; RPPV: right posterior portal vein; LPV: left portal vein; RHV: right hepatic vein; IVC: inferior vena cava; RPHA: right posterior hepatic artery.

Perioperative data for patients with malignant disease

The duration of surgery was 359 ±498 min in patients with HCC, 502 ± 184 min in patients with hilar CCA, and 605 ± 232 min in patients with intrahepatic CCA.Twelve patients with HCC(57.1%) received red blood cell transfusion, with a median of 2 (0–44) U transfused.Three patients (21.4%) with hilar CCA and 5 with intrahepatic CCA (50.0%) received red blood cell transfusions, with a median of 0 (0–6) U and 0 (0–11) U transfused, respectively.The postoperative hospital stay was 16 ±9 days for patients with HCC,20 ± 7 days for patients with hilar CCA, and 22 ± 11 days for patients with intrahepatic CCA (Table 2).

Postoperative morbidity and predictors

Twenty-one patients (39.6%) had postoperative morbidities of Clavien-Dindo grade 3 or higher (Table 3).Right pleural effusion was the most common complication (n= 5, 23.8%), followed by operative site fluid collection (n= 4, 19.0%), and bile leakage(n= 4, 19.0%).Relaparotomy was performed in two patients due to PV stenosis and PV thrombosis, and the patients underwent thrombectomy and intraoperative PV metallic stenting.One HCC patient with macrovascular and BD invasion died due to postoperative bleeding; a PV injury occurred during the surgery, and PV plasty using a cadaveric iliac vein was attempted; however, the bleeding continued as the PV tore.After total hepatectomy for bleeding control, the patient was placed on a deceased donor waiting list for urgent life-saving liver transplantation; however, progression to disseminated intravascular coagulation occurred, and the patient died on postoperative day one.No other deaths occurred within 90 days after surgery.

Preoperative PVE (P= 0.337) was not associated with concomitant procedures, such as extrahepatic BD resection (P= 0.065),caudate lobectomy (P= 0.396), vascular reconstruction (P= 0.541),and lymphadenectomy (P= 0.065), whereas preoperative jaundice(P= 0.003) and operative time (P= 0.006) were significantly associated with postoperative morbidity on univariate analysis.Further,preoperative jaundice [hazard ratio (HR) = 6.15, 95% confidence interval (CI): 1.57–24.17,P= 0.009]and operative time>420 min(HR = 4.66, 95% CI: 1.27–17.17,P= 0.021) were positive independent predictors of postoperative morbidity on multivariate analysis(Table 4).

Table 1Pathology and surgical procedures in patients undergoing left hepatic trisectionectomy.

Table 2Perioperative data for patients with malignant disease.

Table 3Postoperative morbidity after left hepatic trisectionectomy ( n = 21).

Survival and recurrence

The 1-, 3-, and 5-year overall survival rates were 81.1%, 61.4%,and 44.6%, respectively.The 1-, 3-, and 5-year overall survival rates of patients with benign disease were 100.0%, 100.0%, and 100.0%,respectively, with a median follow-up of 46.6 months (Fig.2).On the contrary, the 1-, 3-, and 5-year overall survival rates of patients with malignant disease were 79.2%, 57.3%, and 39.7%, respectively, with a median follow-up of 21.8 months (Fig.2).The overall and disease-free survival rates were calculated only for the three most frequent types of tumors (HCC, intrahepatic CCA, and hilar CCA).The 1-, 3-, and 5-year overall survival rates for each disease were as follows: 71.4%, 55.7%, and 47.8% in patients with HCC; 80.0%, 60.0% and 16.0% in patients with intrahepatic CCA; and 92.9%, 64.3%, and 51.4% in patients with hilar CCA, respectively.In addition, the 1-, 3-, and 5-year disease-free survival rates for each disease were as follows: 32.7%, 32.7%, and 32.7% in patients with HCC; 33.3%, 11.1%, and 11.1% in patients with intrahepatic CCA; and 64.3%, 25.7%, and 25.7% in patients with hilar CCA, respectively(Fig.3).

HCC: hepatocellular carcinoma; CCA: cholangiocarcinoma; CRLM: colorectal liver metastases; GB: gallbladder; PV: portal vein; RHV: right hepatic vein; ERBD: endoscopic retrograde biliary drainage; PTBD: percutaneous transhepatic biliary drainage; HV: hepatic vein; BD: bile duct; HA: hepatic artery.

Discussion

In this study, LHT was performed with curative intent to achieve an R0 resection, and concomitant segment I resection,vascular and biliary reconstructions, and lymphadenectomies were performed in half of the patients.Despite the high risk of mortality for patients undergoing LHT [6,10–12], only one died in the hospital following LHT in our study.Although the incidence of ma-jor complications of Clavien-Dindo grade 3 or higher in this study was slightly higher than that after other types of hepatectomy, it was lower than what reported in other studies of LHT [4,10–15](Table 5).

Table 4Univariate and multivariate analyzes of predictors of postoperative morbidity.

Table 5Morbidity and mortality of left hepatic trisectionectomy for hepatobiliary malignancy in different studies.

Fig.2.Survival curve after left hepatic trisectionectomy.The 1-, 3-, and 5-year overall survival rates of patients with benign disease were 100.0%, 100.0%, and 100.0%, respectively, with a median follow-up of 46.6 months.The 1-, 3-, and 5-year overall survival rates of patients with malignant disease were 79.2%, 57.3%, and 39.7%, respectively, with a median follow-up period of 21.8 months.

Previous studies have demonstrated that preoperative obstructive jaundice is strongly associated with postoperative morbidity and in-hospital mortality [11,16–18], and our study similarly showed that preoperative obstructive jaundice is a positive independent predictor of postoperative morbidity.Although the optimal time interval between preoperative biliary drainage and hepatectomy is controversial, a total bilirubin level of 2 mg/dL or less is used to indicate the optimal time for surgery at our institution.In this study, 19 patients with jaundice required preoperative biliary drainage, including a significant number of patients with hilar CCA(13; 68.4%), and met the total bilirubin requirements prior to LHT.

Preoperative biliary drainage and PVE reduce the risk of postoperative liver failure and improve long-term survival in patients with obstructive jaundice who require major hepatectomy for hepatobiliary malignancies [19–22].The importance of these two procedures was described in a previous study of 302 patients undergoing surgery for hilar CCA in our institution[23].LHT is considered a high-risk procedure because of the relatively low FRLV following surgery.PVE, conducted to reduce this associated risk,increases FRLV and resectability in patients with inadequate FRLV and decreases the risk of post-hepatectomy liver failure.Therefore,PVE prior to LHT should be considered in patients with underlying liver disease or chemotherapy-induced liver injury with marginal FRLV.In principle, according to our indication, preoperative PVE was performed for patients identified as having an FRLV of lessthan 40% in this study.PVE was not performed in patients with hilar CCA if the CT volumetry demonstrated a dominant right HV and the posterior sector volume was expected to be more than 40%of the total liver volume.In contrast, in HCC measuring less than 5 cm, when the tumor was located adjacent to the right anterior and left Glissonean pedicle, PVE was performed.In this study, ipsilateral HV embolization was performed in one patient with intrahepatic CCA with insufficient liver regeneration after PVE[24], and the patient demonstrated a significant increase in FRLV and microscopically cancer-free surgical margins after LHT.

Fig.3.Survival curves according to tumor type.A: The 1-, 3-, and 5-year overall survival rates of patients with hepatocellular carcinoma (HCC) were 71.4%, 55.7%,and 47.8%, respectively.The 1-, 3-, and 5-year overall survival rates of patients with intrahepatic cholangiocarcinoma (CCA) were 80.0%, 60.0%, and 16.0%, respectively.The 1-, 3-, and 5-year overall survival rates of patients with hilar CCA were 92.9%,64.3%, and 51.4%.B: The 1-, 3-, and 5-year disease-free survival rates of patients with HCC were 32.7%, 32.7%, and 32.7%, respectively.The 1-, 3-, and 5-year diseasefree survival rates of patients with intrahepatic CCA were 33.3%, 11.1%, and 11.1%,respectively.The 1-, 3-, and 5-year disease-free survival rates of patients with hilar CCA were 64.3%, 25.7%, and 25.7%, respectively.

As LHT is performed for patients with large and advanced HCC and hepatobiliary malignancy invading the right anterior and left Glissonean pedicle, chemotherapy before surgery plays an important role in long-term survival by downstaging tumor aggressiveness, even when the surgery is curative.Moreover, in hepatobiliary cancer surgery, intraoperative manipulation should be limited to prevent tumor cell dissemination by the portal venous system.However, when operating large bulky tumors, it is difficult to limit manipulation and therefore, the roles of downstaging and minimal tumor manipulation of the tumor are considered more important.

Sequential TACE and PVE have been commonly performed prior to surgery for patients with HCC beyond Milan criteria associated with liver fibrosis and cirrhosis at our institution [25,26].This strategy allows patients with marginal FRLV to become candidates for major hepatectomy, such as LHT, by promoting compensatory FRLV hypertrophy without the deterioration of liver function or tumor progression.However, in this study, most of the patients had a large HCC; hence, the excised functional liver volume was small.Thus, only six of 21 HCC patients underwent PVE before surgery(28.6%), which is a relatively small proportion.Although preoperative TACE results in adhesions between the tumor and neighboring structures, including the diaphragm, which renders surgery diffi-cult[27], it may reduce the tumor bulk and tumor dissemination by intraoperative manipulation.When combined with an appropriate preoperative strategy, LHT can provide satisfactory and acceptable long-term survival for patients with large or advanced HCC tumors, as indicated by the 3- and 5-year survival rates of 55.7% and 47.8%, respectively.Recently, Zhang et al.showed that simultaneous TACE and PVE is safe and effective on increasing FRLV for patients with large HCC before major hepatectomy[28].Although it is an attractive strategy, simultaneous TACE and PVE requires further investigation in larger studies with longer follow-up periods because of concerns about risks of infarction or necrosis of the noncancerous liver parenchyma.

A negative resection margin is crucial to obtain a satisfactory oncological outcome, and one or more operative extensions with LHT are sometimes necessary to achieve negative margins.Most patients with hilar CCA, which extends longitudinally along the BD,require BD resection, segment I resection, and lymphadenectomy.In this study, operative extensions were performed as needed in patients with other malignant and benign diseases.The effects of operative extensions on the operative risk of LHT are unclear [4,6,11].However, concomitant extrahepatic BD resection and lymphadenectomy were not associated with postoperative morbidities in this study.Although these operative extensions may elongate the operative time, which may increase postoperative morbidity, based on the survival outcomes in this study, biliary or vascular reconstruction in patients undergoing LHT should not be considered contraindications for surgery with curative intent.

Despite the remarkable advances in surgical techniques and preoperative diagnostic imaging, LHT remains a challenging procedure even for experienced hepatobiliary surgeons.Additionally,identifying and transecting the liver at an appropriate intersectional plane is challenging.For other anatomical hepatectomy procedures, the intersectional plane can be indented using the major HV as a landmark.However, as the right HV does not always run along the right intersectional plane, identifying the right intersectional plane is sometimes difficult [29,30].Patients with a large inferior HV tend to have a relatively small right HV draining mainly from segment VII.In these patients, full exposure of the right HV may lead to the unnecessary resection of some territories of segment VI.The resulting low FRLV may be a risk for the development of postoperative liver failure.

The anatomical variation of the right portal triad structure is another factor contributing to the difficulty of LHT.Type III PV anatomic features are advantageous for performing LHT; however,this variation is found in only 2.2%–22.6% of patients[9].In this study, seven patients had type III PV.A clear understanding of the course of the right posterior HA and BD relative to the right PV is more important than the identification of the type of anatomical features for each portal triad prior to surgery.The precise identification of the three-dimensional (3D) positional relationships of the right posterior HA and right posterior sectional BD to the right PV, such as supraportal right posterior HA or infraportal right posterior sectional BD, via preoperative evaluation,is essential to prevent the occurrence of postoperative morbidities [9,29,30].Although more research is still needed, the latest technologies, such as indocyanine green-based fluorescence imag-ing, 3D visualization, and 3D printing, are expected to optimize surgical planning and improve the controllability and safety of intraoperative surgery by improving anatomical understanding before and during surgery [31,32].

Fig.4.Left hepatic trisectionectomy for liver donation.A: The right hepatic artery arose from the superior mesenteric artery; the middle hepatic artery arose from the gastroduodenal artery, and A2+ 3 emerged from the LGA.B: The right posterior portal vein emerged early from the main portal vein.C: The right anterior bile duct,accessory right anterior bile duct (B5), right posterior bile duct, and left bile duct separately drain into the common hepatic duct.Dotted lines indicate the division points.D:After dissection of the hilum of recipient, each structure was encircled using a vessel loop (yellow: hepatic artery; blue: portal vein; orange: bile duct).E: The right anterior hepatic artery, middle hepatic artery, and left hepatic artery of the donor were anastomosed to the left, right anterior, and right posterior hepatic arteries of the recipient,respectively.F: Postoperative computed tomography of the recipient 1 week after transplantation.LGA: left gastric artery; RAHA: right anterior hepatic artery; RPHA: right posterior hepatic artery; MHA: middle hepatic artery; LHA: left hepatic artery; RAPV: right anterior portal vein; RPPV: right posterior portal vein; LPV: left portal vein;aRABD: accessory right anterior bile duct; RABD: right anterior bile duct; RPBD: right posterior bile duct; LBD: left bile duct.

The role of LHT, which is a technically demanding and extensive resection in contemporary hepatobiliary practice, is changing.Recently, rare cases of successful living donor liver transplantation(LDLT) using a left trisection graft from a donor with favorable anatomic features for LHT have been reported [33–35].The donor included in this series was our first LHT for LDLT in January 2018(Fig.4).All reported cases of LDLT using a left trisection graft, including this case, have been successful; however, the justification of the complexity and risks of the LDLT procedure for the donor remains controversial[31].As shown in our case, a left trisection graft can be an option in extremely specific situations in donors with type III (early branching of the posterior PV from the main trunk) or IV (late branching of the anterior PV from the umbilical portion of the left PV) PV associated with the acceptable posterior sector liver volume for donor safety; however, in our experience of performing over 60 0 0 adult living donor hepatectomies since February 1997, few donors are eligible for the indication.In addition, this procedure should be performed by surgeons with extensive experience in donor hepatectomy for LDLT and aggressive hepatectomy for locally advanced hepatobiliary disease.

The first limitation of this study is the retrospective nature.The second is the small number included which did not provide suffi-cient power for a meaningful statistical analysis of the prognostic factors.The third is more than 14 years of the study period which may have resulted in historical bias, as significant advances in the perioperative assessment of liver function and diagnostic imaging,surgical techniques, and perioperative management occurred during the study period.Despite these limitations, this study has considerable value because it involves a relatively large cohort of patients who underwent LHT compared with previous studies.

In conclusion, despite the remarkable advances in surgical techniques and perioperative management, LHT remains a challenging procedure with a somewhat higher postoperative morbidity than less-extensive resections.However, LHT has curative potential and an acceptable long-term survival rate in patients with large leftsided and central tumors that extend to the structures of the right anterior Glissonean pedicle.A successful LHT surgery requires a reliable preoperative evaluation of liver function and anatomy of the hilar structures and major HVs and their tributaries, active and appropriate preoperative management for obstructive cholangitis, and an experienced surgeon.

Acknowledgments

None.

CRediT authorship contribution statement

Young-In Yoon:Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Writing –original draft, Writing –review & editing.Sung-Gyu Lee:Conceptualization, Supervision, Writing –review & editing.Deok-Bog Moon:Formal analysis, Methodology, Supervision.Shin Hwang:Data curation, Investigation, Project administration.Ki-Hun Kim:Data curation, Investigation, Project administration.Hui-Ju Kim:Data curation, Investigation, Project administration.Ki-Hoon Choi:Data curation, Investigation, Project administration.

Funding

None.

Ethical approval

This study was approved by the Institutional Review Board of the Asan Medical Center (2020-0358), and the requirement for informed consent was waived due to the retrospective nature of the study.

Competing interest

No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the sub ject of this article.