Trichoscopic evaluation of dental pulp stem cell conditioned media for androgenic alopecia.

Background: Conditioned media (CM) derived from mesenchymal stem cells (MSC) is known to induce hair regrowth in androgenic alopecia. Objectives: The objectives of the study were to assess the efficacy and safety of one type of MSC‐CM, the CM derived from dental pulp stem cells obtained from human exfoliated deciduous teeth (SHED‐CM) and to compare the efficacy of SHED‐CM with and without dihydrotestosterone synthesis inhibitor (DHT‐inhibitor). Methods: Eighty‐eight male androgenic alopecia subjects with Hamilton‐Norwood Classification (H‐N C) I–VII were evaluated by trichoscopy to explore which trichoscopic factors statistically correlated with H‐N C. After being screened, 33 subjects received six SHED‐CM treatments at 1‐month intervals. Clinical severity was assessed through global and trichoscopic images from baseline to 9th month. Results: SHED‐CM was effective for 75% of subjects regardless of disease severity, concomitant DHT‐inhibitor use, and age. Adverse effects including pain and small hemorrhages were transient and mild. We also found that clinical hair status evaluated by absolute values of three quantitative trichoscopic factors (maximum hair diameter, vellus hair rate, and multi‐hair follicular unit rate) showed a good correlation with H‐N C stages, and what is more—a scoring system of these three factors can be a possible predictor of SHED‐CM efficacy. Conclusions: We have shown that SHED‐CM provides global and trichoscopic image improvement for androgenic alopecia, regardless of concomitant DHT‐inhibitor use.

Keywords: androgenic alopecia; conditioned media; dental pulp stem cell; hair loss; trichoscopy

Androgenic alopecia (AGA) is a polygenic condition with progressive hair loss. Age, genetic preposition, and androgen are the main known driving factors in its progression.[1]

As patients with AGA experience loss of self‐esteem and socioemotional deprivation, we need visually effective treatment strategies that can more accurately be seen in global images. However, as Rossi, et al described, classification methods currently used for AGA only show the clinical pattern and the extent of the disease, but not an objective classification.[2] Furthermore, these methods do not describe the real trend of hair loss because of the wide gaps between stages. Accordingly, trichoscopy which leads to more detailed descriptions is now considered to be a fundamental tool for the management of hair loss diseases. Ummiti, et al also indicated that trichoscopy can reveal early variation in hair diameter long before hair loss becomes clinically visible.[3] For this reason, we aimed to explore whether there exist any ideal quantitative trichoscopic factors which have a strong correlation to clinical stages.

Commonly used trichoscopic parameters are hair thickness (hair diameter) and hair density (total hair counts in a certain area). Hence, we first conducted regression analysis to explore whether these parameters have a coefficient correlation to clinical stages. Along with them, maximum hair diameter (Max D) which could indicate the ability to create thick hair, vellus hair rate (VH%) which reflects follicular miniaturization,[[2], [4]] multi‐hair follicular unit rate (MFU%) which could be the expression of shortening of anagen with prolongation of telogen[[2], [5]] and nine other numerical variables were analyzed to find any other useful trichoscopic factors in addition to the three mentioned above.

Hair thickness is a major parameter for describing hair follicle miniaturization. Progressive hair follicle miniaturization in AGA is an expression of the terminal hair (TH: ≥60 μm) transformation into vellus hair (VH: <30 μm). Rossi, et al showed VH% was 24% at AGA stage I, while 73% at AGA stage VI, indicating that the more severe AGA is, the more its hair follicle miniaturization progresses. De Lacharriēre reported that there is a good correlation between hair density and AGA severity.[4] Accordingly, we performed a statistic analysis to explore whether these two trichoscopic values have a coefficient correlation to clinical stages or not.

Mesenchymal stem cells (MSC), mesoderm‐derived immature precursors, have self‐renewal potential and multilineage differentiation capacity. Accumulated studies suggest that the main driving force behind the therapeutic activity observed in MSC are paracrine factors secreted in conditioned media (CM). Administration of CM is a cell‐free therapeutic strategy that activates signaling pathways based on the transfer of extracellular vesicles (EVs) and soluble factors. EVs and soluble factors include bioactive materials such as growth factors, cytokines, mRNAs, and miRNA, which work on the affected tissues.[6] Recent retrospective human studies have shown that CM regulate the hair cycle and hair follicle regeneration.[[7], [9]]

It was reported that CM derived from dental pulp stem cells (dental‐MSC) activates hair follicles, neural cells, adipocytes, and dentine‐producing odontoblast[10] and demonstrates superior nerve regeneration, differentiation, and maturation potentials.[11] In the same way, CM from dental‐MSC has been reported to be effectively used for neural diseases besides odontogenic applications.[12]

Exfoliated teeth are one resource of MSC. Stem cells from human exfoliated deciduous teeth (SHED) were isolated and identified to be a highly proliferative population with multipotential differentiation abilities.[13] Using CM derived from SHED, Gunawardena, et al demonstrated that animal hairs were stimulated by increasing the number of anagen hairs.[14] This study also reported that SHED‐CM contained higher levels of growth factors such as vascular endothelial growth factor and hepatocyte growth factor that induce quicker transition of hair follicles from resting phase (telogen) to growing phase (anagen).

Follicular unit (FU) is known to be composed of terminal hair (TH)s, vellus hair (VH)s, sebaceous glands, arrector pili muscle (APM), and sympathetic nerve. A single APM is shared by all the hair follicles contained within the follicular unit.[15] In the progression process of AGA, APM attachment to VHs is lost, although attachment to THs remains preserved.[[16]]

The APM provides stable anchors that maintain sympathetic innervation to hair follicle stem cells (HFSC). Thus APM, and sympathetic neurons form a dual‐component niche that regulates HFSCs. By the orchestration of these three (nerve, APM, and HFSC), this tri‐lineage unit regulates stem cell activation via synapse‐like structure.[18]

Collectively, our clinical trial indicated that SHED‐CM, which has shown high potential for nerve regeneration, will be effective for AGA through this tri‐lineage.

It is thought that dihydrotestosterone (DHT), derived from testosterone, is responsible for early hair regression.[19] It was found to induce gradual miniaturization of genetically susceptible hair follicles resulting in a reduction of the cellular hair matrix volume and a decrease in the duration of the anagen growth phase.[20] DHT synthesis inhibitor (DHT‐inhibitor) is known as a major metabolite in hair development, although it cannot fully cure the disease. But still, long‐term AGA treatment with DHT‐inhibitor has been documented in well‐controlled clinical trials to show high efficacy and safety.[21] Yanagisawa et al reported that DHT‐inhibitor was quite beneficial, especially when started at early stages (H‐N C I–III). Hence, we divided 33 subjects into three sets of two groups and aimed to compare these two groups: (1) non‐severe (H‐N C II–III) or severe (H‐N C IV–VI); (2) SHED‐CM with or without DHT‐inhibitor; (3) younger or older than 50‐year‐old, so as to determine efficacy differences between the groups.

Based on the capability of SHED‐CM to regenerate hair follicles through their neural regenerative potentials, we investigated the hair growth effects of SHED‐CM on AGA, by conducting clinical and trichoscopical evaluations.

We focused on vertex‐shedding male AGA and designed an open‐label, prospective, pilot study to evaluate the efficacy and safety of SHED‐CM.

We first screened 88 healthy male AGA patients with vertex shedding (mean age: 52.8, SD ± 10.7) who visited the dermatology department of Tokyo Midtown Skin/Aesthetic Clinic from March 2021 to December 2022. All subjects had a physical examination and blood test to exclude severe systemic diseases, or hair loss diseases other than AGA. Those receiving DHT‐inhibitors (oral finasteride or dutasteride) for less than 6 months or other AGA treatment were excluded. Among them, 33 subjects aged 27–69 (mean age 50.7, SD ± 10.5) were enrolled. Table 1 details the baseline characteristics of the 33 subjects. Subjects who had received DHT‐inhibitors longer than 6 months continued with it during and after the treatment until 9th month.

1 TABLE Characteristics of 33 subjects at baseline.

1 Abbreviations: 3QTF‐ts, three quantitative trichoscopic factor‐total score; Max D, maximum hair diameter; MFU%, multi‐hair follicular unit rate; N, number of subjects, Age (standard deviation), variables below (standard error); THC, total hair count/25 square millimeters of hair‐whorl area; VH%, vellus hair rate.

All subjects received a consultation each time before and after being treated in order to check for any adverse events.

We used a SHED‐CM commercial product (SGF; Ginza Solaria Clinic, Tokyo, Japan).[22] Manual microinjections of SHED‐CM in affected areas were administered in deep dermis with 34G quadruple‐tip needles (Quantron; ASTI Corporation), injecting 0.1 mL into each injection site, 1 cm apart. A total volume of 4.5 mL was injected each session.

Before the treatment and at the 3rd, 6th, and 9th month (T3, T6, T9 month) after the initial treatment, global and close contact trichoscopic photographs were taken to evaluate hair status. All subjects were classified by H‐N C at baseline (T0) and subsequent improvements were noted. Trichoscopic images were taken with a dermocamera (Casio DZ‐D100, Japan) around the hair whorls. Since vertex shedding AGA usually starts from hair‐whorl areas, and the individual whorls never change,[23] the benchmark of the measurement area was set within a 5 × 5 mm whorl centering grid. Hair characteristics were measured using image management software (D'z image Viewer, Casio, Japan).

The measurement data included the following:

• hair shaft diameters (HD)
• hair count of each follicular unit
• total number of follicular unit (TFU)
• terminal hair count (TH; diameter > = 60 μm, 2 mm < length)
• vellus hair count (VH; diameter < 30 μm)
• indeterminate hair defined as hair between 30 and 60 μm

The calculated data included the following:

• maximum hair diameters and minimum hair diameters (Max D , Min D) which were the mean diameter of the three thickest or thinnest hairs
• the difference between Max D and Min D (M‐MD)
• total hair count (THC),
• follicular unit rate with single hair (1FU%), double hairs (2FU%), triple hairs (3FU%) double and triple hairs (FFU%), and more than single hair (MFU%)
• terminal hair rate (TH%), vellus hair rate (VH%), and indeterminate hair rate (IH%)

Using these 12 trichoscopic factors (Max D, M‐MD, THC, TFU, VH%, TH%, IH%, 1FU%, 2FU%, 3FU%, FFU%, and MFU%), multiple regression analysis was conducted. A final model was revealed and used for evaluating treatment efficacy. In addition to global and trichoscopic images, H‐N C stages, trichoscopic factors, and radar charts plotting from the trichoscopic factors were also given for each subject.

Data were summarized using standard error of the mean for quantitative variables, relative frequencies, and mean differences (percentages) for categorized variables. Spearman's correlation coefficient was used for analyzing the correlation between H‐N C and each trichoscopic factor. The difference of Max D, VH%, MFU%, and THC from T0 to T3, T6, T9 in each subject were compared using Wilcoxon signed‐rank test. The mean difference between baseline and T3, T6, and T9 were compared using Mann–Whitney U test. Within these models, statistical tests were performed that adjusted for the baseline response levels for each of the post‐baseline time points. Fisher's Exact test was used to determine whether or not there was a significant association between two categorical variables. p‐values less than 0.05 were considered statistically significant.

Based on Spearman's coefficient correlation between male AGA subjects classified by H‐N C and absolute values of trichoscopic factors, six (Max D, M‐MD, VH%, TH%, 1FU%, and MFU%) trichoscopic factors were chosen (Table 2A). Using the correlation matrix depicting the correlation between all the possible pairs of values (Table 2B), a multiple regression was calculated to correlate with baseline H‐N C based on combinations of the independent variables above (data not shown). The final multiple regression model revealed that a combination of the absolute values of Max D, VH%, and MFU% showed the highest coefficient of determination (adjusted R‐square = 0.59), measuring how well a statistical model predicts disease severity. The same analysis was conducted for 33 subjects enrolled in this study, and Max D, VH%, and MFU% showed similar results (Table 4A). On the other hand, hair density (THC) showed a weaker correlation to H‐N C than the three factors above in our study (Table 2A).

2 TABLE (A) Spearman's coefficient correlation between H‐N C and 3QTF. (B) Correlation matrix.

2 Abbreviations: 1FU%, follicular unit rate with single hair; 2FU%, follicular unit rate with double hairs; 3FU%, follicular unit rate with triple hairs; FFU%, follicular unit rate with double and triple hairs; IH%, indeterminate hair rate; Max D, maximum hair diameter; MFU%, multi‐hair follicular unit rate; M‐MD, the difference between Max D and Min D; TFU, total follicular unit; TH%, terminal hair rate; THC, total hair count; VH%, vellus hair rate.

Box and whisker plots illustrate the distribution of Max D, VH%, and MFU% at T0, T3, T6, and T9 (Table 3). Exploiting the interquartile range of Max D, VH%, and MFU%, we set up a score table of these three factors (Table 3). Each absolute value was converted into three quantitative trichoscopic factor individual scores (3QTF‐is) using the score table, and we compared the results of multiple regression analysis conducted by 3QTF absolute values (3QTF‐av), 3QTF‐is, and 3QTF total scores (3QTF‐ts). From this calculation, 3QTF‐is and 3QTF‐ts were identified to have equally high correlational values of H‐N C at T0 (Table 4A).

3 TABLE Box and whisker plots with score table.

• 3 Abbreviations: 0, 3, 6, 9, baseline, 3rd, 6th, 9th month after initial treatment; 3QTF, three quantitative trichoscpic factor (Max D, VH%, MFU%); Max D, maximum hair diameter; MFU%, multi‐hair follicular unit rate; VH%, vellus hair rate.
• 4 TABLE (a) Multiple regression models for predicting baseline H‐N C. (b) Multiple regression models for predicting 9th month 3QTF.

4 Abbreviations: 3QTF, three quantitative trichoscopic factors (Max D, VH%, MFU%); (0): baseline, (3), (6), (9), 3rd, 6th, 9th month after initial treatment.

Absolute values (Figure 1A) and the mean difference (Figure 1B) of Max D, VH%, and MFU% at T3, T6, T9 were significantly improved from T0. Detailed values and 3QTF‐ts at T0 and T9 are listed in Table 5. According to these results, 3QTF‐ts at T9 all increased from the baseline. The highest scores at T9 were non‐severe group and DHT‐inhibitor combined group, while the lowest score was SHED‐CM alone group with only one score improving. Subgroup analysis showed that there were significant differences of Max D and MFU% between non‐severe and severe groups (Figure 2, red asterisk). There were also significant differences of Max D and VH% between with and without DHT‐inhibitor. However, no statistical significance was noticed in Max D, VH%, and MFU% between two age groups. Consequently, disease severity and with or without DHT‐inhibitor should be mainly considered for efficacy evaluation of SHED‐CM treatment.

5 TABLE Absolute values and 3QTF total scores at baseline and 9th month.

5 Abbreviations: 3QTF, three quantitative trichoscopic factors (Max D, VH%, MFU%); 3QTF‐ts, 3QTF total score; Max D, maximum hair diameter; MFU%, multi‐hair follicular unit rate; NS, non‐severe stage (H‐N C II–III); S, severe stage (H‐N C IV–VI); t (−), treated with SHED‐CM alone; t (+), treated with dihydrotestosterone synthesis inhibitor combined with SHED‐CM; VH%, vellus hair rate.

Collectively, SHED‐CM improved 3QTF‐ts in all three sets of two subgroups, because of its ability to increase Max D and MFU%, and decrease VH% when combined with DHT‐inhibitor.

H‐N C and 3QTF‐ts improved 33.3% and 75.8%, respectively (Figure 3A). No subjects' H‐N C deteriorated, while 9.1% subjects' 3QTF‐ts decreased. We suppose that 3QTF‐ts is a more sensitive parameter than H‐N C for evaluating SHED‐CM efficacy.

Apart from age subgroups, 100% of subjects who belonged to non‐severe stage with SHED‐CM alone group had no change of H‐N C stage, yet 80% of them improved 3QTF‐ts. In contrast, 75% of subjects belonging to severe stage with DHT‐inhibitor improved H‐N C stage, and 100% of them also improved 3QTF‐ts (Figure 3B). Fisher's Exact test was used to determine whether there was a significant association between severity subgroups, with or without DHT‐inhibitor subgroups, and younger and older than age 50 subgroups and their combinations. There was not a statistically significant association between all three sets of two groups (p > 0.05). That is, all categories were independent.

Therefore, regardless of severity, concomitant DHT‐inhibitor use, and age difference, SHED‐CM improved their hair status as evaluated by 3QTF‐ts.

To predict 3QTF‐ts at T9, a multiple regression analysis was calculated based on 3QTF‐av, 3QTF‐is, and 3QTF‐ts at T0, T3, T6 (Table 4B). Comparing the R‐squared between nine models (Model 1–9), the greatest explanatory power was carried out in Model 8 and 9. However, we think Model 2 (3QTF‐is at T0) is the most preferable one since we would like to know the final predicted results as early as possible.

Consequently, there is a 48% probability that we can predict T9 3QTF‐ts from T0 3QTF‐is (T9 3QTF‐ts = 8.87 + 0.41*[Max D] + 1.27*[VH%] + 0.03*[MFU%]). The current results highlight the importance and ability of both 3QTF‐is and 3QTF‐ts to predict SHED‐CM efficacy.

Mild adverse events, such as small hemorrhages and needle pain at the injection sites, were observed in all subjects on treatment day. Neither existed at the next day of treatment.

Case 1: VH% did not improve (Figure 4A). However, both global and trichoscopic images showed improvement. A 55‐year‐old, severe stage, and treated with SHED‐CM alone improved his H‐N C stage and 3QTF‐ts. Not only did the score change, but also the radar chart displayed a balanced triangle after the treatment. As in this case, there were a few other subjects who attained good results even if VH% did not improve.

Case 2: No change in his global image and H‐N C stage (Figure 4B). However, trichoscopic images and 3QTF‐ts presented marked improvement.

A 54‐year‐old, non‐severe, and treated with SHED‐CM combined with DHT‐inhibitor did not change his global image and H‐N C stage. However, trichoscopic images and 3QTF‐ts improved, together with a large balanced triangle radar chart at T9.

For a consultation in a dermatology clinic, easy‐to‐see visual tools are needed to display the patients' current hair status. Ideally, treatment efficacy can be shown with image photographs and charts from measured values. However, the most used classifications of AGA, for example, H‐N C, only describe clinical patterns and the extent of the disease.[2] In addition, frequently they are not able to describe the real trend of hair loss because of a wide gap between the different stages. Because of these problems, several scales using trichoscopy have been proposed for giving a clearer, more objective evaluation of a patient's hair status and the ability to better manage it.[[2], [5], [24]]

However, despite these improvements, no one has yet studied the coefficient correlation between global stages and trichoscopical image grading, comparing pre‐ and post‐treatment. In this study, therefore, we explored how to describe the relationship between H‐N C and 3QTF before and after SHED‐CM treatment.

Hair follicle diameter heterogeneity (VH and IH increasing with TH decreasing) is observed as the most common features in AGA. As a previous report documented, trichoscopy can reveal early variation in hair follicle diameter far before hair loss becomes clinically visible.[3] Hence Max D will decrease and VH% will increase when hair follicle heterogeneity progresses. Moreover, since MFU is related to low severity of AGA,[5] MFU% will decrease in severe AGA. It therefore seems likely that 3QTF (Max D, VH%, and MFU%) are logical parameters. That is, 3QTF‐av, 3QTF‐is and 3QTF‐ts show promise for diagnosing AGA treatment; the calculation is simple, and radar charts illustrated by 3QTF‐is have been quite helpful to explain the patient's current hair status in our clinic (Figure 4A,B).

Focusing on the 24.3% (8 subjects) whose 3QTF‐ts did not improve, as seen in the breakdown of three sets of two groups in Figure 3B, in the severity subgroups four were in non‐severe stage, four were in severe stage; in the concomitant treatment subgroups five were in SHED‐CM alone, three were in combination with DHT‐inhibitor (red circles). It is interesting that there were a few ineffective subjects in two different severity groups and in two different treatment groups. So, next we compared two sets of combined groups.

Regarding the group of non‐severe stage with concomitant DHT‐inhibitor treatment, nine out of 12 subjects improved 3QTF‐ts, while three did not. In the group of severe stage with concomitant DHT‐inhibitor treatment, all of them improved 3QTF‐ts (Figure 3B). This means that DHT‐inhibitor does not always lead to good results. The reason is estimated to be related to the polygenic condition of AGA as documented in other reports.[[1], [25]]

As described above, 100% of the subjects improved their 3QTF‐ts in the severe stage with DHT‐inhibitor group. This result suggests that even for those subjects in the severe stage there was high success in improving hair status if treated with SHED‐CM and DHT‐inhibitor. The group of non‐severe stage with SHED‐CM alone showed a high improvement rate; four out of five (80%) subjects improved, while one (20%) stayed the same. Thus, non‐severe stage subjects have a chance to improve even if treated without DHT‐inhibitor.

Next, we would like to discuss 3QTF‐is in more detail. Since progressive hair follicle miniaturization in AGA is an expression of TH transformation into VH, hair thickness is a major parameter for describing hair follicle miniaturization. For the assessment of hair thickness, we measured hair diameters, as the average of the three thickest hairs in a specified area (Max D) were analyzed. Regarding Max D, both absolute values and mean difference improved significantly. Max D in three sets of two groups ended up with very high scores (6, 7). These results suggest that SHED‐CM has high ability to increase Max D (Figure 1A,B, Figure 2).

The improvement of VH% was not as good as Max D; the final scores were 2–6 at the T9, especially low in severe stage and SHED‐CM alone groups (Table 5). Comparing the two groups of severity and concomitant DHT‐inhibitor treatment, a significant difference was present only in the latter groups (Figure 2). These results lead us to believe that DHT‐inhibitor mostly effects VH%.

The absolute values and mean differences in MFU% also showed improvement as well as Max D (Figure 1A,B), and final 3QTF‐is of MFU% were relatively high (5, 6) except in the severe stage group (4). Significant differences between three sets of two different groups were only present in the two different severity groups (Figure 2). In short, SHED‐CM increased MFU%, yet the final score in the severe group was relatively lower than in the non‐severe group.

Hair density, measured by the hair count of a specified area, THC in our study (hairs/5 × 5 mm2), showed a weaker correlation to H‐N C than other parameters of 3QTF as shown in Table 2A. For this reason, we did not use THC to evaluate the treatment efficacy here. However, as hair density has been measured to assess hair regrowth such as by Gentile, et al.,[[26]] it also needs to be considered. The mean difference in THC significantly improved from T0 to T9 (10.7–11.1%) as shown in Figure 1B. The increased rate of THC was 24.3% in total, and the rate stayed at the same levels even 3 months after the last treatment (36 weeks: T9).

In summary, SHED‐CM mainly increases Max D and MFU%. In contrast, the combination of SHED‐CM with DHT‐inhibitor affects VH% alternation. Therefore, we hypothesized that a patient with a higher 3QTF‐is for VH% should start SHED‐CM alone, but with a lower score it is preferable to combine SHED‐CM with DHT‐inhibitor. Patients with lower 3QTF‐is for Max D and MFU% have a good chance of improving with SHED‐CM treatment.

This was an open‐label, prospective, pilot study to evaluate the efficacy and safety of SHED‐CM in subjects of vertex shedding male AGA. Our results demonstrated that SHED‐CM was effective for 75% of subjects regardless of disease severity, concomitant DHT‐inhibitor use, and age. Pain and other adverse events caused by SHED‐CM injection were mild and patients could tolerate it.

We also reported that hair status evaluated by 3QTF‐av (Max D, VH% and MFU%) showed a good correlation with H‐N C stages, and what is more—that a 3QTF scoring system can be a possible predictor of the treatment effects of SHED‐CM. The hair density parameter THC showed significant increase after the treatment, although it showed a lower coefficient correlation with H‐N C than 3QTF.

Since this is a preliminary study, future research is needed to determine the efficacy of SHED‐CM on a larger scale, double‐blind examination. At the same time, we are eager to find additional treatment options such as combination therapy with autologous micrografts[26] to induce better results for the 24.3% of the subjects who did not show improvement in this study.

T.K. and T.O. conceived of the presented idea. T.K. developed the theory and performed the computations. T.K., C.H. and K.Z.Y. verified the analytical methods. S. K. provided study material. J.T. encouraged T.K. to investigate and supervised the findings of this work.

The authors thank Sakiko Kawano, Sachiyo Kiyoku for data acquisition and management, Lawrence T. Knipfing for critical feedback and preparation on this manuscript.

This research did not receive any specific grant from funding agencies in the public, commercial, or not‐for‐profit sectors. The authors declare no potential conflicts of interest.

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

This study protocol was approved by the Tokyo Midtown Medical Center Ethics Review Board (倫‐2021‐03/ February 10, 2021) and registered with the University Hospital Medical Information Network Clinical Trial Registry (UMIN000045897/ October 28, 2021). All subjects agreed to participate in this study, use for publication of images, and provided written informed consent. The study followed the principles outlined in the Declaration of Helsinki. We state that the manuscript has not been published in any form and that it is not considered for publication elsewhere.