Trabecular Bone Remodeling as a New Indicator of Osteointegration After Posterior Lumbar Interbody Fusion

Abstract

Study Design

Retrospective cohort study.

Objectives

We newly found that trabecular bone remodeling (TBR) often appeared in the fixed adjacent vertebrae during bony fusion. Thus, TBR might indicate osteointegration. Hence, we aimed to investigate whether TBR in the early postoperative period could predict future bony fusion after posterior lumbar interbody fusion (PLIF).

Methods

We retrospectively analyzed 78 patients who underwent one-level PLIF. Demographic data were reviewed. Using computed tomography (CT) images taken at 3 months and 1 year postoperatively, we investigated the vertebral endplate cyst (VEC) formation, TBR in the vertebral body, cage subsidence, and clear zone around pedicle screw (CZPS).

Results

TBR had high interobserver reliability regardless of cage materials. VECs, TBR, and both were found in 30, 53, and 16 patients at 3 months postoperatively and in 30, 65, and 22 patients at 1 year postoperatively, respectively. The incidence of VEC, which indicates poor fixation, was lower in early (3 months postoperatively) TBR-positive patients, with a significant difference at 1 year postoperatively (3 months, P = .074; 1 year, P = .003). Furthermore, 3 (5.7%) of the 53 early TBR-positive patients had CZPS without instability at 1 year postoperatively. In 25 TBR-negative patients, 1 (4.0%) had pedicle screw cutout requiring reoperation, 1 (4.0%) had pseudarthrosis, and 4 (16%) had CZPS.

Conclusions

Patients with early TBR (3 months) did not experience pedicle screw cutout nor pseudarthrosis and had significantly fewer VECs than those without early TBR. Thus, TBR may be a new radiological marker of initial fixation after PLIF.

Introduction

Posterior lumbar interbody fusion (PLIF) is widely performed for lumbar canal stenosis with instability and/or malalignment because it promotes reliable nerve decompression and alignment restoration. ¹ Achieving bony fusion is important for PLIF because postoperative pseudarthrosis is associated with symptom recurrence, malalignment, and quality-of-life worsening.^2,3

Fujibayashi et al ⁴ observed the vertebral endplate by computed tomography (CT) and showed that the expansion and contraction of vertebral endplate cyst (VEC) formation potentially predict subsequent nonunion. Obtaining CT images repeatedly is necessary to evaluate the expansion or contraction of VEC associated with pseudarthrosis. Furthermore, VEC is a potential risk indicator of pseudarthrosis; thus, it does not directly provide the evidence of osteointegration to the operated disk level after PLIF. Recently, titanium-coated polyetheretherketone (PEEK) cage (TiP cage) or 3D porous tantalum cage (Tn cage) has been found to promote bioactivity that induces bone ingrowth or osteointegration on cage surfaces, as demonstrated in animal models.^5-10 Detectable bone ingrowth or osteointegration on the cages might be a useful radiographic indicator of bony fusion. Makino et al ¹¹ evaluated bone ingrowth on the surface of TiP cages by using CT color mapping based on Hounsfield unit (HU) value. Kashii et al ¹² also reported that the vertebral cancellous condensation value calculated from HU indicates bone growth on the cage–vertebral endplate surface. However, considering the technical complexity in such evaluation methods, an easier radiographic indicator of osteointegration is desired.

In our postoperative CT review, we noticed that the trabeculae of adjacent vertebrae showed characteristic changes after PLIF, suggesting osteointegration to the grafted cage (Figure 1). Furthermore, this trabecular bone remodeling (TBR) can be observed even in the early postoperative period (approximately 3 months); hence, this finding may be clinically related to future bony fusion. In the current study, we aimed to examine radiological differences in the early stages up to 1 year postoperatively, focusing on VEC and TBR in the vertebral body, and to evaluate whether the early appearance of TBR indicates future bony fusion.

Figure 1.

Trabecular bone remodeling. Characteristic changes in the bone trabeculae of the adjacent vertebral bodies after posterior lumbar interbody fusion.

Methods

Patient Population

We retrospectively reviewed 105 consecutive patients who underwent one-level PLIF for lumbar degenerative disease between April 2017 and March 2020. This study was approved by the Institutional Review Board of Anjo Kosei Hospital, Anjo, Japan (No. R21-046). All clinical and radiological interventions were routine assessments, and all participants provided written informed consent. Patients who underwent concomitant laminectomy at other levels were included. The indications for PLIF were as follows: 1) spondylolisthesis with more than 3 mm of slippage and/or more than 5° of posterior opening on dynamic lateral X-ray images and 2) stenosis requiring total facetectomy for decompression. We excluded 17 patients because of previous spinal fusion, hemodialysis for renal failure, and severe scoliosis (Cobb angle >30°). For postoperative evaluation, CT was performed at 3 months and at 1 year postoperatively. To reduce radiation exposure, we did not perform CT after 1 year, except in patients with complications. Thus, we excluded 11 patients who did not undergo CT scans at 3 months and/or 1 year. Ultimately, 78 patients (67.8 ± 12.3 years, 34 males and 44 females) were analyzed (Figure 2).

Figure 2.

Flowchart for patient selection.

Surgical Procedure

One surgeon (RS) performed the PLIF in a standardized manner, using a bilateral pedicle screw (PS)-rod system and 2 interbody cages in all cases. Bilateral facet joints were resected as required, and grade 1 osteotomy ¹³ was performed in most cases. Seven patients who had other levels of concomitant stenosis underwent partial laminectomy concurrently. PSs were inserted, temporary fixation was provided, and the disc was removed to avoid damaging the vertebral endplate. The local bone was morselized to graft, and the graft bone was inserted into the intervertebral space before inserting 2 cages of the same type. 3D porous Tn cages (TM Ardis; Zimmer Biomet, Warsaw, IN) and titanium-coated PEEK cages (MectaLIF-TiPEEK; Medacta, Castel San Pietro, Switzerland; Concorde Pro Ti; DePuy Synthes, Raynham, MA; Capstone PTC; Medtronic Sofamor Danek, Memphis, TN) were used. Tn cages had no space for the bone grafting, but the interior of the TiP cages was filled with grafted bone. For 3 months postoperatively, the patients wore a hard corset.

Patients’ Demographic and Operative Data

Patients’ age, sex, body mass index (BMI), smoking history, preoperative bone mineral density (T-score of the proximal femur measured using dual-energy X-ray absorptiometry [Prodigy; GE healthcare, Chicago, IL]), and PLIF level were recorded. The Japanese Orthopedic Association (JOA) ¹⁴ scores (range: 0-29 points) were recorded before and 1 year after surgery. The JOA scoring system consists of 3 subjective symptoms (0-9 points); 3 clinical symptoms, such as neurological deficit (0-6 points); 7 activities of daily living (0-14 points); and bladder function (−6 to 0 points); the higher the scores, the better the patient condition.

Radiological Assessments

One year postoperatively, the range of motion (ROM) of the local angle (measured using the rostral endplate of the upper instrumented vertebra and the caudal endplate of the lower instrumented vertebra) was estimated using standing lateral forward and backward bending X-ray images. In addition, CT images were taken using standard methods at our institution, and multi-planar reconstruction (MPR) images were produced.

Radiological images were examined by 2 examiners (NS and YK) blinded to the clinical outcome but not blinded to the type of cage, given that both X-ray and CT clearly show the cage type. Continuous variables were measured using the mean value of each examiner’s measurements. The presence or absence of a finding was considered to be present if both examiners judged it to be present.

Trabecular Bone Remodeling

TBR was defined in the coronal images of CT-MPR (Figure 1). TBRs were imaging findings in the adjacent vertebrae and were distinguished by morphology and density compared with normal vertebral trabeculae. TBRs were new bone structures, which appeared “postoperatively,” and were not observed at pre- and immediate postoperative CT. TBRs could be observed both vertically and obliquely from the contact portion of the cage and vertebral endplate toward the PSs. Furthermore, TBRs were distinct from the background trabeculae owing to their high density (high HU value), which is similar to the cortical bone (Figure 3A and B).

Figure 3.

Computed tomography coronal images of representative cases. (A) Early TBR-positive (arrow) case using Tn cages. (B) Early TBR-positive (arrow) case using TiP cages. (C) A case that changed from VEC-positive/TBR-positive to VEC-negative/TBR-positive status. Three months postoperatively, the central cyst was enlarged (arrowhead) but was also TBR-positive (arrow). One year postoperatively, the cyst had shrunk, and bridging bone (arrowhead) had developed, confirming bony fusion. (D) VEC-positive/TBR-negative case with CZPS. Three months postoperatively, cysts had developed (arrow), and TBR was not present. One year postoperatively, the cysts were enlarged, and cage subsidence (arrow) and CZPS (arrowhead) were observed. (E) A case of bony fusion with cage subsidence using Tn cages. Three months postoperatively, cage subsidence occurred (arrowhead), but the cage that had not subsided was TBR-positive (arrow). One year postoperatively, a bridging bone (arrowhead) had developed, confirming bony fusion. Abbreviation: TBM, trabecular bone remodeling; VEC, vertebral endplate cyst; CZPS, clear zone around pedicle screw.

The number of occurrences was counted for each cage, and the total number was recorded as 1 score, which ranged from 0 to 2 (0: no TBR, 1: 1 cage had TBR, and 2: both cages had TBR).

Vertebral Endplate Cyst

VEC formations^4,15 were evaluated by the sagittal and coronal views of CT-MPR (Figure 3C and D). VEC formation was also evaluated in the same way as TBR and scored from 0 to 2 (0: no cysts occurred at all, 1: cysts formed in only 1 cage, and 2: cysts formed in both cages).

Cage Subsidence

We defined cage subsidence as cage entry into the vertebral endplate of 2 mm or more (Figure 3D and E). Subsidence was assessed on CT-MPR images and scored on a scale of 0 to 2, similar to TBR or VECs. The current series showed no cage deviation from the intervertebral space.

Clear Zone Around Pedicle Screw

CZPS was defined as a translucent zone measuring more than 2 mm around a PS. CZPS was assessed using CT axial images. Given that 4 PSs were inserted in all cases, CZPS was scored from 0 to 4.

Pseudarthrosis

Pseudarthrosis was determined if 1 year after surgery, the X-ray showed a ROM of more than 5° in the local angle ¹⁶ and/or the CT images showed loosening of all 4 PSs.

Statistical Analysis

Data are presented as mean and standard deviation for the continuous variables and as number and percentage for the categorical variables. Statistical data were analyzed using R version 4.1.0 (http://www.R-project.org) on Wilcoxon rank sum test, Fisher’s exact test, Pearson’s chi-square test, and (weighted) kappa coefficient. A P value of less than .05 was considered significantly different.

Results

For all 78 patients, the PLIF level was L4-5 in 57 (73%) patients, L5–S in 16 (21%) patients, and L3-4 in 5 (6.4%) patients. The JOA score was 13.7 ± 3.3 preoperatively, and it significantly improved to 27.0 ± 3.0 at 1 year postoperatively (P < .001). Tn cages were used in 43 (55%) patients, whereas TiP cages were used in 35 (45%) patients. Age, sex, BMI, smoking history, preoperative bone mineral density, PLIF level, and pre- and postoperative JOA score were not significantly different between the cage types.

The interobserver reliability of CT image evaluation depended on the cage materials (Table 1). With TiP cages, the kappa coefficient for VEC evaluation exceeded .8, as in previous reports, ⁴ whereas with Tn cages, the kappa coefficient was .593-.607, considering that the radiation halo of tantalum affected the area around the Tn cage. Similarly, the kappa coefficients of cage subsidence were .779-.785 for the TiP cage and .628-.650 for the Tn cage. Regarding TBR, the kappa coefficients were constant, ranging from .686 to .739, regardless of the cage type.

Table 1.

Interobserver Reliability for Computed Tomography Image Evaluations.

	κ
	Tn cage	TiP cage
VEC, 3 months a	.593	.845
VEC, 1 year a	.607	.804
Subsidence, 3 months a	.650	.779
Subsidence, 1 year a	.628	.785
TBR, 3 months a	.739	.711
TBR, 1 year a	.703	.686
CZPS, 3 months a	1.000	1.000
CZPS, 1 year a	1.000	.936

Abbreviations: κ, kappa coefficient; Tn cage, 3D porous tantalum cage; TiP cage, titanium-coated polyetheretherketone cage; VEC, vertebral endplate cyst; TBR, trabecular bone remodeling; CZPS, clear zone around pedicle screw.

^aWeighted kappa coefficient.

The flowchart in Figure 4 shows the distribution of VEC and TBR at 3 months and 1 year postoperatively. At 3 months, VEC, TBR, and both were found in 30, 53, and 16 patients, respectively. Two of the 11 patients who were VEC-negative/TBR-negative at 3 months changed to VEC-positive/TBR-negative at 1 year, and 1 of them had CZPS. One of the 14 patients who were VEC-positive/TBR-negative at 3 months underwent revision surgery because of PS cutout secondary to poor stability; of the remaining 13 patients, 6 remained VEC-positive/TBR-negative at 1 year, of which 4 had CZPS and 1 had pseudarthrosis because all 4 PSs were loosened on CT images and showed a local ROM of more than 5° on dynamic X-ray image. ¹⁶ The remaining 7 changed to VEC-positive/TBR-positive status and had no CZPS. Of the 37 patients who were VEC-negative/TBR-positive at 3 months, 3 changed to VEC-positive/TBR-positive status after 1 year, and 1 of them had cage subsidence and CZPS despite showing stability on lateral X-ray image. Of the 16 patients who were VEC-positive/TBR-positive at 3 months, 12 remained unchanged after 1 year, and 2 of them had CZPS; meanwhile, VECs disappeared in the 4 remaining patients.

Figure 4.

Computed tomography evaluation at 3 months and 1 year postoperatively.

We also compared the patients according to TBR presence in the early stage (after 3 months) (Table 2). The early TBR-negative group had a larger BMI and lower bone mineral density than the early TBR-positive group, but the difference was not statistically significant. No significant differences were observed in other patient backgrounds or clinical outcomes between the 2 groups. The early TBR-positive group had significantly more cases of Tn cage (P < .001), and it had a lower incidence of VEC, with a significant difference at 1 year postoperatively (3 months, P = .074; 1 year, P = .003). No significant difference was found in cage subsidence and CZPS, although they occurred less in the early TBR-positive group. Early TBR-positive patients did not manifest clinically problematic implant-related complications such as PS cutout or unstable pseudarthrosis. However, early TBR positivity alone did not completely predict some imaging problems, such as CZPS, 1 year after surgery. For example, 1 early TBR-positive patient who changed from VEC-negative to VEC-positive between 3 months and 1 year postoperatively exhibited CZPS, although it was not associated with clinical symptoms and revision surgery.

Table 2.

Comparison Between Positive and Negative Early Trabecular Bone Remodeling.

	Early TBR-positive	Early TBR-negative	P-value
	N = 53	N = 25
Age	67.7 (10.7)	68.0 (15.4)	.40
Sex, Male	25 (47.2%)	9 (36.0%)	.35
BMI	22.5 (3.2)	23.8 (2.9)	.081
Smoking	6 (11%)	1 (4.0%)	.66
T-score (Hip)	−1.7 (.8)	−1.9 (.8)	.38
Osteoporosis treatment			.88
No	45 (85%)	20 (80%)
PTH	3 (5.7%)	2 (8.0%)
BP	2 (3.8%)	2 (8.0%)
Denosumab	2 (3.8%)	1 (4.0%)
SERM	1 (1.9%)	0 (0%)
JOA Score, pre-op.	13.5 (3.4)	14.2 (3.1)	.40
JOA Score, 1year a	27.1 (3.1)	26.8 (2.6)	.31
PLIF level			.70
L3-4	3 (5.7%)	2 (8.0%)
L4-5	40 (75%)	17 (68%)
L5–S	10 (19%)	6 (24%)
Cage			<.001
Tn cage	37 (70%)	6 (24%)
TiP cage	16 (30%)	19 (76%)
Pseudoarthrosis			.10
Follow-up	0 (0%)	1 (4.0%) b
Reoperation	0 (0%)	1 (4.0%) c
VEC, 3 months			.074
0	37 (70%)	11 (44%)
1	12 (23%)	9 (36%)
2	4 (7.5%)	5 (20%)
VEC, 1 year a			.003
0	38 (72%)	9 (38%)
1	9 (17%)	4 (17%)
2	6 (11%)	11 (46%)
Subsidence, 3 months			.34
0	41 (77%)	18 (72%)
1	9 (17%)	3 (12%)
2	3 (5.7%)	4 (16%)
Subsidence, 1 year a			.20
0	36 (68%)	15 (62%)
1	13 (25%)	4 (17%)
2	4 (7.5%)	5 (21%)
TBR, 1 year a			<.001
2	31 (58%)	4 (17%)
1	22 (42%)	8 (33%)
0	0 (0%)	12 (50%)
CZPS, 3 months			.063
0	50 (94%)	21 (84%)
1	2 (3.8%)	0 (0%)
2	1 (1.9%)	4 (16%)
CZPS, 1 year a			.079
0	50 (94%)	19 (79%)
2	3 (5.7%)	4 (17%)
4	0 (0%)	1 (4.2%) b

Abbreviations: BMI, body mass index; PLIF, posterior lumbar interbody fusion; T-score, bone mineral density assessed by T-score measured by dual-energy X-ray absorptiometry at the proximal femur; PTH, teriparatide; BP, bisphosphonate; SERM, selective estrogen receptor modulator; JOA score, Japanese Orthopedic Association scoring system (0-29 points); Tn cage, 3D porous tantalum cage; TiP cage, titanium-coated polyetheretherketone cage; VEC, vertebral endplate cyst; TBR, trabecular bone remodeling; CZPS, clear zone around pedicle screw.

^aOne patient without early TBR positivity who underwent reoperation for PS cutout was excluded.

^bPseudarthrosis was determined, and follow-up was performed.

^cReoperation was performed for PS cutout.

Discussion

The current study investigated the radiological changes after PLIF with a focus on TBR to evaluate whether TBR appearance at 3 months postoperatively indicates future bony fusion. The early TBR-positive cases not only had no clinically problematic pseudarthrosis or implant-related problems but also had significantly fewer VECs 1 year postoperatively. Regardless of cage material, TBR had stable interobserver reliability. Thus, TBR could be a reliable indicator of osteointegration after PLIF.

The present study focused on a new finding regarding TBR; as bony fusion progresses, the remodeled trabecular bone was placed obliquely toward the cages in the vertebral body (Figure 3A and B). This trabecular bone placement probably resulted from remodeling^17,18 because the trabeculae ran obliquely similar to the load directed from the PS to the interbody cages. With cages made of radiolucent materials such as PEEK, TBR also appeared along with the trabeculae that passed through the cage, making the intervertebral space continuous. Therefore, TBR can be 1 of the indicators of osteointegration. Considering that the strong radiological halo of tantalum makes the area around Tn cages difficult to observe, ¹⁹ the Tn cage has less stable interobserver reliability than the TiP cage in assessing cage subsidence and VEC. However, the interobserver reliability of TBR was comparable irrespective of the cage materials. In fact, we noticed the TBR on the images of patients using the Tn cage. Initially, we suspected that TBR is a phenomenon unique to the Tn cage. However, given that we were also able to observe TBR in the TiP cage, we believe that TBR is widely expressed after PLIF.

The evaluation of post-PLIF osteointegration by CT imaging has been widely investigated. Makino et al ¹¹ focused on the HU values at the contact surface of the cage and vertebral endplate in CT images and then evaluated bone ingrowth by using color mapping. Kashii et al ¹² also focused on the HU values at the contact surface and reported that higher HU values of the vertebral endplate indicate better fixation between the cage and the bone. Their findings are probably another observation of TBR described in our study. However, TBR does not require complex operations on the picture archiving and communication system software or the assistance of an external software that calculates HU values; it can evaluate cage–bone fixation by visually capturing the changes in the vertebrae body. Moreover, VEC is a good predictor of nonunion or delayed fusion after PLIF, and it appears when micromotion occurs in the fixed segment where stress-induced microfracture or bone resorption can take place.^4,15 Therefore, VEC formation is a high-risk factor for spinal implant failures. In the present series, only those who were VEC-positive at 3 months postoperatively experienced implant-related problems (Figure 4). Thus, VEC formation is a good predictor of fixation failure in PLIF. However, being a risk factor for fixation failure, VEC cannot directly predict fusion; repeated CT evaluation is necessary to observe cyst expansion or contraction.

Although early VEC positivity suggests intervertebral instability, patients who were also early TBR-positive did not experience clinically problematic mechanical failures. However, early VEC-positive/TBR-negative patients suffered a high rate of clinically problematic mechanical failure or CZPS. Thus, TBR is an imaging finding complementary to VEC and can provide another perspective on the stabilization process that cannot be determined by VEC alone. Although VEC is a good marker of instability, its interobserver reliability is relatively low with Tn cages because of the radiation halos of tantalum. Given that VEC is a sufficiently reliable index,^4,15 the combination of TBR and VEC can further improve post-PLIF evaluation. Furthermore, TBR can be observed in the adjacent vertebral body, rather than the contact surface between the cage and the vertebral endplate. Therefore, TBR can be used to evaluate osteointegration/bone ingrowth even with radiation halos of tantalum.

TBR assessment, although easy, may be arbitrary because it is not objective. In addition, TBR may not be detected in some patients despite the progress of bony fusion. In this case, vertebral endplates are highly sclerotic because of degeneration, or the trabeculae in the vertebrae appear rough because of osteoporosis. Thus, the presence or absence of TBR should be examined in comparison with preoperative CT images. This problem may be 1 of the reasons why the interobserver reliability of TBR is lower than that of VEC in TiP cage cases. Therefore, TBR is not an absolute index for determining osteointegration after PLIF. However, in our findings, the weighted kappa coefficient for TBR assessment was .686-.739, which is high enough for clinical evaluation combined with VEC.

Cage material influenced early TBR positivity, with Tn cages being significantly more common than TiP cages. The relationship between cage material and bone ingrowth has been investigated in animal studies. McGilvray et al ²⁰ found that bone ingrowth into the porous architecture of the TiP cage was 46.35% ± 4.70% and 41.40% ± 2.60% at 12 and 18 weeks in an ovine model. Conversely, Bobyn et al ⁷ reported that the average bone ingrowth into the porous architecture of the Tn cage was 63.1%-69.2% at 16 weeks in a canine model. Thus, Tn cages are better for early bone ingrowth, and its intervertebral stabilization is achieved earlier than that of TiP cages, resulting in an earlier TBR positivity. In other words, TBR positivity suggests intervertebral stability and is associated with less subsequent mechanical failure and vigorous bone ingrowth.

CZPS was adopted as 1 of the negative outcome measures, but CZPS did not indicate PLIF failure, unlike PS cutout. Seven cases of CZPS without obvious pseudarthrosis were identified in the current study, whereas 2 cases of pseudarthrosis (including PS cutout) were identified. Bone fusion is a race between biology and biomechanics, and if the race is close, CZPS will occur. However, even if CZPS occurs, it is quite possible that interbody bone fusion will be achieved later. In that case, only the CZPS would remain; CZPS reflects relative lower “initial” stability of the PLIF. This was an important factor in the decision to adopt early postoperative (3 months) imaging as the primary assessment. However, there were only 2 cases of clinically problematic instrumentation-related problems. Therefore, it is acknowledged that there are insufficient negative outcomes in the current study to assess positive or negative predictive values. The aim of this study was to determine imaging results that are predictive of the occurrence of instrument-related problems after PLIF. Furthermore, the strength of the study lies in the fact that TBR is not an indicator of high-risk condition, but rather an indicator that can predict a favorable course.

This study has several limitations. First, it is not a randomized controlled study. To standardize the clinical and mechanical conditions, we only included patients who underwent one-level PLIF operated by 1 sufficiently skilled surgeon. Second, the TiP cage group was not a unified group because of the varied cage geometries. Third, the sample size was relatively small. Despite these limitations, statistically significant results were obtained, and this study reveals that TBR is a finding complementary to VEC and is a sufficient finding to directly suggest post-PLIF osteointegration. However, it is acknowledged that caution should be used when interpreting interobserver reliability because the agreement rate is lower for numerical endpoints and higher for two-way decisions. Furthermore, obesity, age, osteoporosis, and smoking are also risk factors for pseudarthrosis in lumbar fusion according to previous studies.^21-23 Owing to the relatively small sample size, these risk factors may not have been identified as such in the current study; therefore, risk factors should be validated in future large-scale studies.

Conclusion

The new imaging finding “TBR” demonstrated stable interobserver reliability regardless of the cage material used. Patients who were TBR-positive at 3 months postoperatively not only had no implant-related problems requiring reoperation but also had significantly fewer VECs at 1 year postoperatively. Therefore, TBR may be a new indicator of initial fixation after PLIF, that is, osteointegration/bone ingrowth, which can be evaluated independent of the cage material.

ORCID iDs

Naoki Segi https://orcid.org/0000-0001-9681-2422

Hiroaki Nakashima https://orcid.org/0000-0002-0039-9678

Kei Ando https://orcid.org/0000-0002-1088-2903

Shiro Imagama https://orcid.org/0000-0003-1721-9626