The efficacy of transbronchial indocyanine green instillation for fluorescent-guided wedge resection

Abstract

 

OBJECTIVES

The purpose of this study was to investigate the feasibility of lung wedge resection by combining 3-dimensional (3D) image analysis with transbronchial indocyanine green (ICG) instillation, in order to delineate the intended area for resection.

METHODS

From December 2017 to July 2020, 28 patients undergoing wedge resection (17 primary lung cancers, 11 metastatic lung tumours) were enrolled, and fluorescence-guided wedge resection was attempted. Virtual sublobar resections were created preoperatively for each patient using a 3D Image Analyzer. Surgical margins were measured in each sublobar resection simulation in order to select the most optimal surgical resection area. After transbronchial instillation of ICG, near-infrared thoracoscopic visualization allowed matching of the intended area for resection to the virtual sublobar resection area. To investigate the effectiveness of ICG instillation, the clarity of the ICG-florescent border was evaluated, and the distance from the true tumour to the surgical margins was compared to that of simulation.

RESULTS

Mean tumour diameter was 12.4 ± 4.3 mm. The entire targeted tumour was included in resected specimens of all patients (100% success rate). The shortest distances to the surgical margin via 3D simulation and by actual measurement of the specimen were11.4 ± 5.4 and 12.2 ± 4.1 mm, respectively (P = 0.285) and were well correlated (R² = 0.437). While all specimens had negative malignant cells at the surgical margins, one loco-regional recurrence was observed secondary to the dissemination of neuroendocrine carcinoma.

CONCLUSIONS

ICG-guided lung wedge resection after transbronchial ICG instillation and preoperative 3D image analysis allow for adequate negative surgical margins, providing decreased risk of local recurrence.

INTRODUCTION

Recent advances in computed tomography (CT) have allowed for the detection of small lung nodules that cannot be palpated during surgery. To resolve the problem of locating the nodule for resection, various methods for identifying and delineating the target nodule have been developed in order to confirm the location of the nodule seen on imaging [1–3]. Although wedge resection carries an inferior survival rate compared to segmentectomy in stage I non-small-cell lung cancer [4], wedge resection is frequently and easily performed in compromised patients [5]. Should on-site pathology reveal lung cancer, an additional operation, such as lobectomy, is performed for which surgical margins are not an issue. Lung wedge resection, on the other hand, references intentional limited resection and is intended for patients with a peripheral lung cancer smaller than 2 cm with ground-glass opacities (GGOs) on CT [6]. Passive limited resection can be utilized in severely compromised patients with cardiopulmonary disease or for patients with a metastatic lung tumour, situations where the surgical margin of the tumour is critical. Prior methods used to find and mark the locations of small lung tumours have been limited to pinpoint marks, which are associated with increased risks of positive surgical margins.

We recently developed a novel method for tumour location during segmentectomy. Our method utilizes transbronchial instillation of indocyanine green (ICG) followed by near-infrared thoracoscopy [7, 8]. Preoperatively, a simulation of anatomical sublobar resection is created from chest CT imaging. Next, ICG is instilled into the associated bronchi, which indicates the resection area by simulation immediately prior to operation under general anaesthesia. A notable characteristic of this method is its ability to demonstrate the precise and exact area associated with ICG instilled bronchus matched with the simulation of sublobar resection [8]. We now use this method for lung wedge resections, as it allows us to delineate the area for resection with increased chances of adequate negative tumour margins. The purpose of our study was to investigate the feasibility of this method for margin-free wedge resections in an effort to avoid local recurrence in lung cancer.

PATIENTS AND METHODS

Ethical statement

This study was registered in the Japan University Hospital Medical Information Network Clinical Trial Registry (UMIN-CTR) on 11 August 2014 with the Registration No. R000017236 (IRB: 140902, approved on 26 September 2014).

Patient enrolment

Our study protocol was approved by the Research Ethics Board of Tokyo Women’s Medical University, Japan. This was a prospective single-centre, phase II, feasibility study. Written informed consent was provided by all study participants. The inclusion criteria were as follows: (i) early-stage lung cancer located in the parenchyma of the outer two-thirds of the lobe from the hilum, defined as pure or mixed GGOs with a nodule of 2 cm in diameter or less with a tumour disappearance ratio of 0.5 or greater, or solid tumour of 1 cm in diameter or less without evidence of nodal and distant metastasis (c-stage 0 or IA1) (intentional limited resection); (ii) metastatic lung tumour amenable to wedge resection; and (iii) a patient with lung cancer at high risk due to poor cardiopulmonary function or poor general condition who was not a candidate for lobectomy or segmentectomy (c-stage IA1 to IA2) (compromised limited resection). The exclusion criteria were as follows: benign diseases including pneumothorax with bullae and benign tumours, and history of iodine allergy. Patients with multiple lung tumours necessitating lobectomy or segmentectomy combined with wedge resection were also excluded.

From December 2017 to July 2020, 48 wedge resections for 43 patients with malignant tumours were performed at our department. When the tumour was expected to be easily detected under a thoracoscope due to pleural indentation or protrusion of the nodule from the lung surface under a condition of a collapsed lung, we felt ICG fluorescence image would not be necessary, resulting in the exclusion of 15 patients from the study. Ultimately, 28 patients undergoing 28 wedge resections were enrolled in the study. The sample size of 28 was based on a previous clinical study demonstrating the safety of fluorescence image-guided pulmonary resection after endobronchial ICG instillation [7]. The sample size was justified by rationale on feasibility, precision reflected by mean values and variance, and regulatory considerations [9].

Surgical margin measurements

Postoperative histopathological examinations were performed to confirm whether the target lesions were in the resected specimen and that the surgical margins were free of malignant tumour cells. We measured the shortest distance from the surgical margin to the tumour in the deflated condition and compared it to the measurement of the simulation in an inflated condition.

Postoperative follow-up

To evaluate the safety of our procedure, all study patients were followed for 1 month after surgery to determine whether any complications arose. Prolonged air leakage was defined as air leakage still present by postoperative day 5. Recurrent air leakage was defined as lung collapse on the operative side after removal of the chest tube. Other evaluated complications included postoperative bleeding, pneumonia, arrhythmia, empyema and acute lung injury.

Patients underwent follow-up at 3-month intervals after surgery, which included blood sampling and radiological examination. CT was performed at 3 months after surgery and then at 6-month intervals to assess for tumour recurrence and/or new lesions by CT. Patients were followed to investigate intermediate surgical outcomes, particularly local recurrence at the site of resection by 31 October 2020. The average duration of follow-up was 15.6 ± 9.1 months (median 14.0 months).

Statistical analysis

Data were analysed using SPSS Statistics version 20 (IBM Japan, Tokyo, Japan). Results were expressed as means ± standard deviation. The Mann–Whitney U-test was used to compare the difference between margin distances determined by the simulation and the actual measurement of surgical specimens. All reported P-values were 2-sided, and P-values <0.05 were considered statistically significant.

SURGICAL TECHNIQUE

Creation of a virtual sublobar resection

High-resolution CT, 3-dimensional pulmonary angiography and virtual bronchoscopy were performed prior to surgery to confirm tumour and associated vessels and bronchi. Several simulated sublobar resections were performed to determine appropriate tumour resection margins. For preoperative planning of a sublobar resection, each participant underwent multislice CT by a 320-slice scanner (Aquilion ONE/ViSION Edition; Toshiba Medical Systems, Tokyo, Japan) to create a virtual bronchoscopic procedure for sublobar resection based on bronchial structure and to measure lung volume utilizing the 3-dimensional Volume Analyzer ‘Synapse VINCENT’ (Fujifilm Co., Tokyo, Japan) (Fig. 1A–C). With the selection of 4th–6th bronchial branches, the tumour was included in the simulated area. The shortest distance from the tumour to the resection margin was measured by the VINCENT analyser. The most appropriate region to undergo sublobar resection was selected, based on the following criteria: resection margin ∼2 cm or greater from the tumour or greater than the diameter of the tumour. The nomenclature on lung segments was based on the report by Jafek and Carter [10] and Nomina Anatomica [11].

Figure 1:

A representative case of small-sized peripheral lung cancer. The 13-mm-diameter nodule located at S²aii. Depth from lung surface was 10 mm (A). Several potential associated bronchi (#1: B²aiiα, #2: B²aiα, #3: B²aiiβ) were identified on a virtual bronchial tree (B). Simulation of right S²aiiα (green) + S²aiiβ (red) removal revealed that the tumour is located in the centre of this area with an 8.2-mm surgical margin (C).

Transbronchial indocyanine green instillation

After induction of general anaesthesia, a single-lumen endotracheal tube or a laryngeal mask was introduced for transbronchial instillation of ICG. The ICG (25 mg/10 ml of diluent) was further diluted in 70 ml of saline and 20 ml of autologous blood for a 10-fold dilution of ICG, as adsorption of ICG to human serum albumin can increase its intensity of fluorescence [12].

With the patient in a supine position, a thin bronchoscope (BF-P260F; Olympus Medical Co., Tokyo, Japan) was inserted into the targeted bronchus. A bronchial catheter with balloon (Olympus Disposable Balloon Catheter B5-2C/2LA; Olympus Medical Co.) was inserted, and the balloon was inflated at the bronchial orifice. Ten millilitres of the ICG solution was instilled into each targeted 4th–6th bronchial branch that was a subsegmental or more peripheral bronchus, and 200–400 ml of air was then directed into the bronchus to distribute ICG to the peripheral areas (Fig. 2). During this manoeuvre, the bronchoscope was fitted over the balloon to prevent leakage of ICG and to visualize the tip of the catheter over the balloon (Video 1). Overall, instillation required ∼5–15 min. After instillation of ICG, a double-lumen Broncho-Cath tube was introduced, and 5 cm H₂O of positive end-expiratory pressure ventilation was maintained until the start of the operation. The duration from the end of ICG instillation to the start of operation was ∼15–30 min.

Figure 2:

Representative anatomical bronchial findings by virtual bronchoscopy and actual bronchoscopic images of the patient in Fig. 1 from the orifice of the right upper lobe to B²aiiα (A–D). The bronchoscope was inserted into B²aiiα under the guidance of virtual bronchoscopy.

Fluorescence-guided lung wedge resection

At the beginning of the surgery, a near-infrared thoracoscope (PINPOINT; Stryker, Kalamazoo, MI, USA) was used to visualize the clear border of the planned resection area. The visceral pleura was marked by electric cautery along the border of the ICG fluorescence (Fig. 3A). The lung was divided by electric cautery and/or endostaplers along the marked line to complete the sublobar resection (Fig. 3B, Video 2). Resection margins were measured to ensure adequate distance from tumour to specimen edge. The visceral pleura was then closed if needed to stop air leakage.

Figure 3:

Intraoperative indocyanine green-fluorescent image of the patient in Fig. 1 was clearly visualized during the surgery (A) and a resected lung specimen (B). The tumour was identified in the resected specimen with a sufficient margin distance of 18 mm.

RESULTS

Patient characteristics are shown in Table 1. Seventeen patients had primary lung cancer and 11 patients had a metastatic lung tumour. The mean size of the tumour including GGO was 12.4 ± 4.3 mm (range 6–24 mm) with the consolidation of 7.5 ± 5.1 mm. The consolidation/tumour ratio was 62.3 ± 43.0% (0–100). The mean depth from the surface of the lung by CT was 7.2 ± 6.5 mm (range 0–20 mm). The mean number of created virtual sublobar resections was 3.2 ± 1.0 (range 2–5). Maximum standardized uptake value by positron emission tomography was 3.1 ± 2.8 and serum carcinoembryonic antigen level was 3.3 ± 2.1 ng/ml. No patients required additional postoperative treatment.

Table 1:

Patient characteristics and preoperative image analysis

Age (years)	69.4 ± 9.5 (41–83)
Gender (male/female)	17/11
Consolidation diameter (mm)	7.5 ± 5.1 (0–16)
Tumour diameter (mm)	12.4 ± 4.3 (6–24)
Consolidation/tumour ratio (%)	62.3 ± 43.0 (0–100)
Depth from the lung surface (mm)	7.2 ± 6.5 (0–20)
Average number of simulations	3.2 ± 1.0 (2–5)
Positron emission tomography (SUVmax)	3.1 ± 2.8 (0.8–9.8)
Serum CEA (ng/ml)	3.3 ± 2.1 (1.0–9.5)
Preoperative pulmonary function
FVC (l)	3.15 ± 0.8
%FVC (%)	105.8 ± 17.0
FEV1.0 (l)	5.47 ± 16.5
%FEV1.0 (%)	104.3 ± 22.2
FEV1.0/FVC (%)	76.0 ± 12.6

The data are expressed as mean ± standard deviation (range).

CEA: carcinoembryonic antigen; FEV1: forced expiratory volume in 1 s; FVC: forced vital capacity; SUVmax: maximum standardized uptake value.

Surgical procedures and outcomes are presented in Table 2. Intentional limited resection was performed in 12 patients and compromised limited resection was performed for 4 patients with primary lung cancer. Indications for compromised limited resection included post-left pneumonectomy with age over 80 years old in 1 case, post-lobectomy at a contralateral side with age over 80 years old in 2 cases and severe chronic obstructive pulmonary disease in 1 case. Preoperative pulmonary function tests in these patients were notable for forced vital capacity 2.5 ± 0.7 l (97% predicted), forced expiratory volume in 1 s 1.5 ± 0.6 l (84.4% predicted) and forced expiratory volume in 1 s/forced vital capacity 61.8 ± 24.8%. Planned wedge resections were done in 16 right lungs and 12 left lungs. In terms of pathology for lung cancer, invasive carcinoma with papillary component was seen in 4 intentional limited resections and 3 compromised limited resections. Minimally invasive carcinoma and adenocarcinoma in situ were diagnosed in 9 intentional limited resections. Based on the simulations, ICG was instilled into the 4th branch (sub-subsegmental bronchus level) to the 6th branch (sub-sub-sub-sub-bronchus level), and wedge resection was completed with 100% of resected specimens containing the intended tumour. The border of ICG fluorescence was visualized clearly in all patients from the beginning of the surgery and could be marked circumferentially by electric cautery. ICG fluorescence was clearly observed during the entire procedure so that division of the lung by staplers or electric cautery could be easily performed. Almost immediately after removal of the intended lung tissue, the majority of ICG disappeared from the residual lung. The mean number of ICG-injected bronchi was 2.4 ± 1.0 (range 1–5). The difference between the mean shortest width of the surgical margin measured on the simulations (11.4 ± 5.4 mm) and the width measured on the specimen (12.2 ± 4.1 mm) was not significant (P = 0.285). The distance on the simulation was well correlated to that on the actual measurement (R² = 0.438, P = 0.009, Fig. 4A). We obtained a surgical margin distance of more than 5 mm in a collapsed condition in all cases. In terms of the relation between the actual margin distance and the depth from the tumour, there was no correlation (R² = 0.0001, P = 0.972) (Fig. 4B). Surgical margins were negative for malignancy in all cases (100% success rate). We also evaluated the success of surgical margins by wedge resection based on the recommendations of Sawabata et al. [13]. They insisted that successful margins were those 2 cm or more from the tumour, or greater than the diameter of the tumour. Based on these criteria, 17 cases were considered to have successful surgical margins leading to a 60.7% (17/28) success rate (Fig. 5A). However, when we compared consolidation maximum diameter and the actual surgical margin distance, actual distance was longer than the consolidation diameter in 24 cases (success rate 85.7%, Fig. 5B).

Table 2:

Surgical procedures and outcomes

Reason for wedge resection
Intentional limited resection	12
Compromised limited resection	4
Metastatic lung tumour	11
Pathological diagnosis
Primary lung cancer	17
Invasive adenocarcinoma	7
Intentional	4
Compromised	3
Minimally invasive adenocarcinoma (all intentional)	5
Adenocarcinoma in situ (all intentional)	4
Combined large cell neuroendocrine carcinoma (compromised)	1
Metastatic lung tumour	11
Colon cancer	8
Rectal cancer	1
Lung cancer	1
Urothelial cancer	1
Number of ICG-injected bronchi, mean ± SD (range)	2.4 ± 1.0 (1–5)
Operating time (min), mean ± SD (range)	69.9 ± 19.7 (33–116)
Blood loss (ml), mean ± SD (range)	24.3 ± 43.6 (0–68)
Surgical margin (n = 28)	P = 0.285
Virtual (mm), mean ± SD (range)	11.4 ± 5.4 (4–23.1)
Actual (mm), mean ± SD (range)	12.2 ± 4.1 (6–22)
Hospitalization (days), mean ± SD (range)	5.3 ± 2.3 (3–14)
Postoperative complications	2
Prolonged air leakage > 5 days	2
Pneumonia	0
Acute lung injury	0
Arrhythmia	0
Follow-up duration (months), mean ± SD (range)	15.6 ± 9.1 (3–35)
Local recurrence (resection margin site)	1
Distant metastasis	0

ICG: indocyanine green; SD: standard deviation.

Figure 4:

The relationship between the distance of virtual and actual margins (A). There was a good correlation, with R² = 0.438 (P = 0.009). In terms of the relation between the actual margin distance and the depth from the tumour, there was no correlation (R² = 0.0001, P = 0.972, B).

Figure 5:

Successful surgical margins were defined as margins 2 cm or more from the tumour, or greater than the diameter of the tumour reported, per Sawabata et al. [13]; 60.7% (17/28) of cases were successful (A). When the consolidation diameter was compared with actual margin distance, 85.7% (24/28) were successful (B).

There were no ICG-related adverse events noted, including allergic reactions and inflammatory changes of the lung, in all enrolled cases. The postoperative complications consisted of prolonged air leakage lasting more than 5 days requiring pleurodesis in 2 patients, and complications including pneumonia, acute lung injury and arrhythmia were not observed in any patients. One patient had multiple distant metastases with intrapulmonary dissemination 1 year after surgery. His pathology was consistent with large cell neuroendocrine carcinoma, 16 mm in diameter. He underwent a compromised limited resection because he had a left pneumonectomy for squamous cell carcinoma 15 years ago. This patient had a surgical margin distance of 17 mm but died of lung cancer 21 months after surgery. No other patients developed local recurrence at the resection site and all remain alive.

DISCUSSION

Minimally invasive surgery for lung cancers, such as lung wedge resection, is lauded for its low morbidity and mortality rates compared to lobectomy, but compromised patients undergoing minimally invasive surgery have a high recurrence rate [14]. Several investigators have reported that tumour size and the margin are crucial factors for local recurrence [13, 15, 16]. Sawabata et al. [13] reported that a ratio of margin width to maximum tumour diameter >1 or a margin width >2 cm was a critical factor in local control. Inadequate resection margins of the central side are the most frequent cause of local recurrence. Several methods have been explored for securing the central surgical margin [17–19]. However, these methods are complicated, expensive, and not well established. Sato et al. [17, 18] reported on the utility of virtual-assisted lung mapping, a preoperative bronchoscopic multispot dye-marking technique. This enables effective wedge resection with sufficient surgical margins. However, this technique requires an additional CT examination, and the distances from the tumour to the marking points are diverse and irregular. Furthermore, the security of the central margin is unresolved. Therefore, they planned virtual-assisted lung mapping with a microcoil implementation for securing the central margin [20].

A unique feature of our procedure is that it not only marks where the tumour exists but also allows for guidance as to how much lung to remove to ensure safe and successful surgical margins. Although the requirement of sufficient surgical margins per Sawabata et al. [13] was fulfilled in only 60.7% of our cases, all specimens contained the intended tumour and the pathology of the surgical margin was negative for malignancy in all cases (success rate 100%). As lung adenocarcinoma with a GGO component is associated with excellent survival compared to solid lesions because of a no or minimally invasive feature [21], negative surgical margins with longer margin distance compared to the maximum solid diameter may be meaningful. From this point of view, 24 resections in 28 patients (85.7%) fulfilled this condition. One compromised patient with prior left pneumonectomy experienced local recurrence with multiple distant metastases due to an underlying high-grade malignancy. This case was thought to be a contra-indication of limited resection. Neither local nor distant metastases were observed in any other patients, suggesting the effectiveness of this method.

Even though underlying pleural change can facilitate greater ease in identifying tumour location, the security of the central margin is challenging. From the viewpoint of the anatomical structure of the lung, when the lung is divided deeply, the lung is severed into a conical shape. Therefore, the deeper the lung is severed, the wider the range of required cutting surface is necessary. In our method, as the resected area was displayed by fluorescence, an accurate and deep wedge resection is possible. Furthermore, our method allows for identification of sub-subsegmental (or more peripheral) borders via ICG fluorescence. When the lung is anatomically divided by an electric cautery along the fluorescence border, the lung can then be divided with less bleeding and less air leakage as there is less transection of vessels and bronchi.

Fluorescence-guided surgery based on the florescent contrast agent ICG has been developed for use in various medical fields, and intraoperative fluorescence imaging has already been used to guide lung segmentectomies [7, 22, 23]. There are 2 types of ICG injection: intravenous and transbronchial. The benefit of intravenous injection is that it is easy to use and repeatable. However, the disadvantages are very short illumination time (∼30–180 s) and unclear demarcation in cases of emphysema and fibrosis. On the other hand, the benefit of transbronchial instillation is the capability of direct instillation into lung tissues and the persistence of visible fluorescence during surgery. We previously reported that transbronchial instillation of ICG could target a small bronchus and that an appropriate and sufficient surgical margin could be obtained by additional instillation into the 3rd branch (subsegmental) to 6th branch (sub-sub-sub-subsegmental) [8]. Therefore, a segmentectomy that was a complete match to the virtual segmentectomy could be performed. We applied this method to wedge resection. For tumours located deep in the lung, the volume of resected lung is equal to or greater than a subsegment. This suggests that a very wide large resection is necessary. When securing an adequate surgical margin is impossible, a segmentectomy or lobectomy must be performed. Our method can remove a large lung volume that approaches the volume of a subsegment or segment without the need for ligating individual blood vessels.

We acknowledge the challenges associated with our method. It requires advanced technical skills for successful instillation of ICG into the lung and carries the possible risk of allergic or inflammatory reaction to the lung. Additionally, VINCENT is the only available software for simulating segmental or subsegmental lung resection and this software is only recently expanding its commercial availability in the world. The mastery of understanding the bronchial structure and bronchoscopic technique is mandatory. By understanding the structure of bronchial trees and following the virtual bronchoscopy, the bronchoscope can be inserted correctly. In terms of ICG, most of the transbronchial injected ICG is removed during surgery and the remaining ICG is drained by the lymphatic stream, and ultimately eliminated by the liver. Furthermore, as ICG was diluted by saline and human blood, there were no ICG-related adverse events noted.

Surgical margin distances during simulation were measured in the inflated state. However, actual measurements were made in the deflated state, yielding a different position for measurement compared to the simulation. Furthermore, this is not an anatomical removal of the lung despite the simulation of anatomical sublobar resection. We used ICG for identifying the area of lung removal, which confirmed sufficient margin distance. Even though the measurement position is different between the simulation and resected specimen, security of margin distance is most important. We measured the shortest distance in the deflated state of the tissue and expected the actual distance to be longer in the inflated state. Given we had successful negative margins in the deflated state, we do not feel inflation or deflation yields a challenge. Additionally, there is currently no validation method to determine whether or not ICG was instilled correctly. If we can take a cone-beam CT into the operating theatre, we may be able to confirm whether ICG was distributed in the objected area, as the instilled dilute liquid ICG should be visualized as ground-glass shadows.

There are various indications for large wedge resections, such as pure GGO without pleural indentation, a non-palpable tumour in the peripheral or intermediate portion, and desires to avoid severe adherent vascular transections during completion pneumonectomy for a secondary tumour after upper lobectomy (Fig. 6).

Figure 6:

Indications for transbronchial indocyanine green instillation technique for this method. From left image to right image: pure GGO without pleural change; non-palpable tumour localized deep from the surface; second primary cancer after contralateral pneumonectomy requiring minimally invasive surgery with short operating time; and avoidance of completion pneumonectomy after left upper lobectomy. GGO: ground-glass opacity.

Limitations

We acknowledge the limitations of this study. First, it was a small single-centre non-comparative study. Secondly, the strict indications for enrolment led to exclusion of many patients with malignant tumours. Lastly, our study did not have sufficient long-term outcomes reported. To address these limitations, and further validate our technique, a multicentre randomized clinical trial of standard wedge resections for further evaluation and long-term outcomes is needed.

CONCLUSION

In conclusion, virtual sublobar resection combined with transbronchial-injected ICG-fluorescence-guided wedge resection is applicable to the establishment of margin-free wedge resections for patients with lung tumours. This technique advances currently available methods due to its instillation technique and intraoperative confirmation of the ICG instillation area.

Conflict of interest: none declared.

ABBREVIATIONS

3D

3-dimensional

CT

Computed tomography

GGO

Ground-glass opacity

ICG

Indocyanine green

Author contributions

Yasuo Sekine: Formal analysis; Investigation; Methodology; Project administration; Writing—original draft; Writing—review & editing. Eitetsu Koh: Data curation; Investigation. Hidehisa Hoshino: Data curation; Investigation; Writing—review & editing.

Reviewer information

Reviewer information Interactive CardioVascular and Thoracic Surgery thanks the anonymous reviewer(s) for their contribution to the peer review process of this article.