Functional outcomes in early-stage tongue cancer patients following partial glossectomy: a pilot study comparing primary closure and secondary healing with biomaterials

Abstract

Background

This retrospective study explores primary closure (PC) and secondary healing (SH) using biomaterials (a double layer of Matriderm^® and Neoveil^®) in tongue cancer surgery, focusing on postoperative outcomes and indicators of functional recovery, including speech evaluation, movement, and discomfort.

Methods

A retrospective analysis of 20 patients who underwent partial glossectomy without free flap reconstruction at Seoul National University Dental Hospital between September 2020 and December 2023 was conducted. Patients were divided into PC (n = 10) and SH with biomaterials (n = 10) groups. Data included demographic details, clinical characteristics, and speech evaluation results from the urimal test of articulation and phonology (U-TAP), multi-dimensional voice program (MDVP), speech mechanism screening test (SMST), intelligibility, movement, and discomfort.

Results

The SH group showed a significantly larger mean surgical defect dimension (P < 0.05). No significant differences were found in sex, depth of invasion, tumor–node–metastasis (TNM) stage, and tumor location. Functional outcomes showed no significant differences in U-TAP scores, MDVP parameters, SMST scores, intelligibility, movement, or discomfort.

Conclusions

In this retrospective pilot study, biomaterial-assisted SH was feasible in selected patients and showed broadly similar postoperative speech-related outcomes to PC; however, comparisons should be interpreted cautiously due to baseline imbalances and confounding by indication.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12903-026-08491-4.

Background

The tongue plays a critical role in articulation, swallowing, taste, and overall oral health. The occurrence of tongue cancer necessitates surgical interventions, which present significant challenges not only in tumor removal but also in preserving the tongue’s functionality and the patient’s quality of life [1, 2]. Partial glossectomy inevitably creates a mucosal and soft-tissue defect, and the method of defect management can substantially influence recovery of speech and swallowing as well as long-term tongue contour. In clinical practice, the reconstructive plan is primarily guided by defect size and geometry, with the goal of achieving adequate coverage while minimizing distortion and contracture [3]. Although primary closure (PC) is often favored for smaller defects, even tension-free closure can alter tongue shape and mobility depending on defect configuration, and these changes may translate into bleeding, delayed healing, or functional compromise requiring prolonged adaptation of speech and swallowing [4–6].

Accordingly, reconstructive strategies after partial glossectomy span a spectrum from PC and skin grafting to local tongue flaps, pedicled flaps (e.g., submental or platysma-based flaps), and microvascular free flaps [7–9]. Microvascular free flaps can restore bulk and mobility for larger composite defects—particularly when a substantial portion of the mobile tongue is resected—but they require microsurgical expertise, longer operative time, and entail donor-site morbidity [8, 10]. Pedicled flaps provide reliable vascularized tissue with shorter operative time than free flaps; however, they may introduce external scarring, bulk mismatch, and potential limitations in tongue mobility depending on flap design [9, 11, 12]. For selected small defects, skin grafting or local tongue flaps may provide coverage while avoiding microsurgery, but these approaches can still be associated with contraction, contour distortion, or tethering that affects function [7, 10, 13].

In many patients with small-to-moderate mucosal defects, free-flap reconstruction may be unnecessary, yet tension-free primary closure is not always feasible without distorting tongue contour or creating undue tension that can contribute to discomfort and scar contracture [7, 10, 13]. This intermediate clinical scenario highlights the need for alternative defect-management strategies that preserve native tongue shape and mobility while providing stable coverage during early wound healing [7, 10]. Biomaterial-assisted secondary healing (SH) is one such approach: by offering temporary protection and a scaffold for re-epithelialization, it may facilitate healing with less tethering and without additional donor-site morbidity [14].

Building on this concept, SH approaches using innovative biomaterials such as Matriderm^® (MedSkin Solution Dr. Suwelack AG, Billerbeck, Germany) and Neoveil^® (Gunze Ltd., Ayabe, Japan) offer promising avenues. Matriderm^®, an acellular dermal matrix, and Neoveil^®, a biodegradable mesh sheet, have shown considerable potential in reconstructing oral mucosa defects [15–17], suggesting their capability to maintain structural integrity, reduce discomfort, and expedite functional recovery.

This study aims to address the comparative evidence between PC and novel SH techniques employing biomaterials in tongue cancer surgery. It focuses on small tongue cancer cases and evaluates postoperative outcomes, including wound healing, functional recovery assessed through speech evaluation, and patient-reported discomfort. By analyzing the impact of each surgical method on the recovery process, this study strives to provide evidence-based insights to optimize surgical strategies and improve patient quality of life.

Materials and methods

Patient selection and data collection

A retrospective analysis approved by the Institutional Review Board of Seoul National University Dental Hospital (IRB No. ERI23033) was conducted. Consecutive patients treated between September 2020 and December 2023 were screened for eligibility (Fig. 1). Inclusion criteria were: (1) surgical management with partial glossectomy; (2) histopathologically confirmed squamous cell carcinoma of the tongue; (3) no microvascular free-flap reconstruction (i.e., primary closure or secondary intention healing); and (4) completion of a postoperative speech evaluation as part of routine follow-up care. Patients not meeting any of these criteria were excluded. No additional sampling or discretionary selection was applied.

Fig. 1.

Flow diagram of patient selection

Patients were classified into two groups according to the wound management strategy documented in the operative and follow-up records: primary closure (PC) versus secondary healing (SH) using a double-layer biomaterial (Matriderm^® and Neoveil^®). The choice between PC and SH was primarily based on defect size/geometry and the feasibility of achieving tension-free closure without distorting tongue contour or mobility. PC was preferred when tension-free closure was feasible, whereas SH with biomaterials was selected when primary closure was expected to create excessive tension or distortion.

Comprehensive patient data were collected from electronic medical records, which included demographic details (age and sex), oral cancer staging according to the American Joint Committee on Cancer tumor–node–metastasis (TNM) classification (8th edition), and specific measurements of the surgical defect such as depth of invasion (DOI) and maximum dimensions. Additionally, data on the anatomical location of the cancer, patient-reported symptoms, and quantitative results from speech evaluation were meticulously compiled.

Surgical procedures

For all participants, operative records were reviewed to extract procedural details (e.g., surgical instruments/energy devices) and to classify wound management. Patients were classified into two groups according to the documented closure strategy: PC with sutures (Fig. 2) versus SH with a Matriderm^®/Neoveil^® double-layer (Fig. 3). Partial glossectomy was performed using a combination of monopolar cautery and an ultrasonic surgical device (Harmonic Scalpel^®, Ethicon Endo-Surgery Inc., Cincinnati, OH, USA), according to surgeon preference and intraoperative requirements. PC was achieved with non-absorbable sutures when tension-free approximation was feasible. In the SH group, Matriderm^® was trimmed to match the defect and placed in direct contact with the wound bed as a collagen–elastin scaffold intended to support granulation and epithelial migration. Neoveil^® was then applied as an outer protective layer to provide mechanical stability, reduce friction/irritation in the mobile tongue environment, and facilitate secure coverage during early healing. The double layer was fixed according to the operative note using non-absorbable sutures and fibrin glue. This double-layer concept was selected to combine scaffold integration (Matriderm^®) with early protective stabilization (Neoveil^®), which may be less consistently achieved with a single-layer dressing in the oral cavity.

Fig. 2.

Surgical procedures of PC (A) Preoperative appearance showing squamous cell carcinoma on the right lateral tongue. B Defect of the tongue following glossectomy. C Immediate postoperative view following PC. PC; primary closure

Fig. 3.

Surgical procedures of SH (A) Intraoperative view showing squamous cell carcinoma on the right lateral tongue. B Defect of the tongue following glossectomy. Defect coverage with a double layer of (C) Matriderm^® and (D) Neoveil^®. SH; secondary healing

Speech evaluation

Speech evaluations were conducted in the maxillofacial clinic by a single speech-language pathologist. Standardized measures targeting various aspects of speech and speech function were administered to assess patients’ linguistic capabilities approximately 6–12 months following partial glossectomy. In addition to objective speech measures, patient-reported oral discomfort was routinely queried at the time of speech assessment and documented in the standard speech-evaluation record. For this retrospective study, discomfort data were extracted from these records and analyzed as present/absent (symptom type categorized when described).

Articulation proficiency examination

Articulation proficiency was systematically examined using the urimal test of articulation and phonology (U-TAP), developed by Kim and Shin [18]. The U-TAP is an assessment tool designed to calculate consonant and vowel accuracy. It involves naming 30 individual word pictures and evaluates 43 consonants and 10 monophthongs within the initial, medial, and final positions of syllables in the words. For the present analysis, U-TAP results are reported as a percentage score reflecting overall articulation accuracy, with higher values indicating better consonant/vowel production accuracy. The percentage score was calculated as the proportion of correctly produced target phonemes out of the total targets elicited in the test items, following the standard scoring approach.

Speech mechanism screening test

The speech mechanism screening test (SMST) evaluates abnormalities in the structure or function of the speech mechanism [19]. It assesses the coordination process of the speech organs and measures the maximum movement speed, regularity, and accuracy of the muscles that comprise the speech organs. It measures the maximum phonation time (MPT) during vowel prolongation to indirectly evaluate lung capacity, intra-lung pressure, vocal fold contact, and airflow through the vocal folds. The SMST also assesses speech rate, articulation errors, alternating motion rate (AMR), and sequential motion rate (SMR) to screen for abnormalities in speech production or neuromuscular coordination. Each test item is scored and compared against age-specific norms for mean, standard deviation, and cutoff scores to benchmark the subject’s performance. In our study, we assigned numerical scores to each assessment category as follows: normal (0), mild (1), mild to moderate (2), moderate (3), moderate to severe (4), and severe (5). The MPT was calculated separately, while AMR and SMR scores were combined to calculate a single score. These scores were then summed to calculate the overall total score for each subject. In this scoring framework, higher total scores indicate greater abnormality/severity of impairment in the speech mechanism.

Speech intelligibility assessment

Speech intelligibility, a critical component of speech function, was evaluated using a listener-based word and sentence intelligibility task adapted for Korean routine clinical assessment, following commonly used intelligibility evaluation frameworks and Korean-language test methodologies [20–22]. Patients were prompted to articulate a predetermined set of Korean words and sentences representing varying linguistic complexity. A listener rated overall comprehensibility for each item and classified it as normal or abnormal. Intelligibility was recorded as a percentage, calculated as the proportion of items judged as having normal intelligibility out of the total items administered. This assessment was intended to capture the clarity and comprehensibility of speech after different closure methods. Ratings were performed by a single listener during routine clinical speech evaluation.

Voice analysis

The multi-dimensional voice program (MDVP) was used to delve into the acoustic aspects of patients’ voices [23]. Patients sustained the vowel /a/ for approximately 5 s, allowing for a detailed examination of voice quality and stability. The MDVP analyzed parameters such as frequency in Hertz (Hz), jitter (related to frequency variation) in percentage, shimmer (related to amplitude variation) in percentage, and the noise-to-harmonics ratio (NHR), which assesses the relationship between noise and harmonic components. Although MDVP primarily captures phonatory stability, these acoustic parameters were included as supplementary descriptive measures to characterize overall speech-related voice quality, which can be secondarily influenced by articulatory adaptation and compensatory speech patterns after tongue resection.

Statistical analysis

The data collected from the patient cohort was subjected to rigorous statistical analysis to evaluate the impact of different closure techniques on speech and healing outcomes following partial glossectomy. Statistical analysis was performed using SPSS software (version 26; IBM Corporation, Armonk, NY, USA). Key variables included demographic information (age, sex), clinical features (cancer stage, defect dimensions including DOI and maximum dimensions), and the method of surgical closure (PC vs. SH). Normality of continuous variables was assessed using the Shapiro–Wilk test. Because of the small sample size and non-normal distribution of several variables, between-group comparisons for continuous outcomes were performed using the Mann–Whitney U test, and categorical variables were compared using Fisher’s exact test. In addition, the Mann–Whitney U test was employed to compare the U-TAP scores, voice parameters (frequency, jitter, shimmer, and NHR), SMST total scores (including MPT, AMR, and SMR), and speech intelligibility scores between the primary and secondary closure groups. All tests were two-sided, and P values < 0.05 were considered statistically significant. This pilot study was designed as a retrospective exploratory analysis; therefore, no a priori sample size calculation was performed.

Results

Demographic and clinical characteristics

The following results should be interpreted as preliminary findings from this pilot study, given the limited sample size. The SH group was significantly older than the PC group (P < 0.05), and defect dimensions were also significantly larger in the SH group. The maximum dimension of the measured parameters was significantly different between the groups, with the PC group having an average dimension of 2.7 ± 0.9 cm compared to 3.8 ± 0.9 cm in the SH group (P < 0.05). No significant difference in the DOI was observed, with the PC group having a DOI of 0.24 ± 0.14 cm and the SH group having a DOI of 0.18 ± 0.09 cm. Regarding the stage distribution, 90% of patients in the PC group were at stage I, while 100% of patients in the SH group were at stage I. One patient (10%) in the PC group was ultimately classified as stage II (pT2) on final histopathology, although the lesion had been managed surgically under the preoperative clinical impression of a T1 tumor (cT1). This patient was clinically node-negative (cN0) and therefore did not undergo neck dissection. None of the patients in either group received postoperative radiotherapy. The location of the condition was anterior in three patients (30%) in the PC group and seven patients (70%) in the SH group versus posterior in seven patients (70%) in the PC group and three patients (30%) in the SH group. This difference was also not statistically significant (Table 1). Detailed clinicopathologic variables, including TNM categories (T/N/M), DOI, and additional demographic characteristics, are provided in Supplementary Table S1.

Table 1.

Demographic data of included patients

Group	PC (n = 10)	SH (n = 10)	p value
Sex (M: F)	5:5	5:5	1
Age, mean ± SD (yr)	53.3 ± 11.9	63.4 ± 9.4	0.049*
Dimension, mean ± SD (cm)	2.7 ± 0.9	3.8 ± 0.9	0.013*
DOI, mean ± SD (cm)	0.24 ± 0.14	0.18 ± 0.09	0.436
TNM Stage, n (%)			1
I	9 (90%)	10 (100%)
II	1 (10%)	0 (0%)
Location, n (%)			0.18
Anterior	3 (30%)	7 (70%)
Posterior	7 (70%)	3 (30%)

* Statistically significant difference at p value < 0.05

Abbreviations: PC primary closure, SH secondary healing, DOI depth of invasion, TNM tumor-node-metastasis

In the PC group, complete wound healing (as defined by full epithelialization without exudate/bleeding requiring further dressing) was typically achieved within 2–3 weeks. In contrast, patients in the SH group showed partial second-intention healing with partial absorption of the biomaterials at the 3-week follow-up. By 2 months postoperatively, complete healing was observed (Fig. 4).

Fig. 4.

Postoperative clinical characteristics of PC and SH. In the PC group, complete wound healing was achieved at 3 weeks (A) and maintained at 2 months postoperatively (B). In the SH group, partial healing with biomaterial absorption was noted at 3 weeks (C), whereas complete healing was achieved by 2 months (D). PC; primary closure, SH; secondary healing

Functional outcomes

For the U-TAP scores, the PC and SH groups had similar mean values across all categories with no significant differences between the groups. SMST scores, including MPT and AMR/SMR, also did not show significant differences between the groups. In terms of intelligibility, 90% of patients in the PC group were classified as having normal intelligibility compared to 80% in the SH group, but this difference was not statistically significant. Regarding movement, 60% of patients in the PC group had normal movement compared to 50% in the SH group, with no significant difference. In terms of discomfort, one patient (10%) in the PC group and six patients (60%) in the SH group reported discomfort, although this difference was not statistically significant (Table 2). The MDVP analysis showed no significant differences between the groups in terms of frequency, jitter, shimmer, and NHR. Following analysis, the mean fundamental frequency was revealed to be 177.45 ± 52.64 Hz in the PC group versus 168.07 ± 32.72 Hz in the SH group. Separately, the mean jitter was 2.26% ± 2.06% in the PC group and 2.45% ± 2.59% in the SH group, and the mean shimmer was 6.94% ± 4.95% and 8.53 ± 4.45%, respectively. Finally, the NHR was 0.26 ± 0.15 in the PC group and 0.25 ± 0.12 in the SH group (Fig. 5).

Table 2.

Comparison of functional outcomes between PC group and SH group

Group	PC (n = 10)	SH (n = 10)	p value
U-TAP, mean ± SD (points)
C_W	98.14 ± 3.77	97.21 ± 5.24	0.657
C_S	97.44 ± 5.2	98.60 ± 2.94	0.816
V_W	98.89 ± 3.33	99.54 ± 1.47	0.878
V_S	100	100	1
SMST, mean ± SD (points)			0.651
MPT	0.6 ± 1.26	0.9 ± 1.29
AMR/SMR	6.11 ± 5.69	7.5 ± 6.35
Intelligibility, n (%)			0.542
Normal	9 (90%)	8 (80%)
Abnormal	1 (10%)	2 (20%)
Movement, n (%)			0.661
Normal	6 (60%)	5 (50%)
Abnormal	4 (40%)	5 (50%)
Discomfort, n (%)			0.061
Yes	1 (10%)	6 (60%)
No	9 (90%)	4 (40%)

* Statistically significant difference at p value < 0.05

Abbreviations: PC primary closure, SH secondary healing, U-TAP urimal test of articulation and phonology, C_W consonants_word, C_S consonants_sentence, V_W vowels _word, V_S vowels _sentence, SMST speech mechanism screening test, MPT maximum phonation time, AMR alternating motion rate, SMR sequential motion rate

Fig. 5.

Postoperative MDVP parameters, including (A) fundamental frequency, (B) jitter, (C) shimmer, and (D) NHR, in the PC and SH groups. Each dot represents an individual patient: circles correspond to the SH group, and triangles correspond to the PC group. The red “X” indicates the mean value of the PC group, while the yellow “X” indicates the mean value of the SH group. No statistically significant differences were observed between the two groups for any MDVP parameter. MDVP; multi-dimensional voice program, PC; primary closure, SH; secondary healing, NHR; noise-to-harmonics ratio

Discussion

This retrospective pilot study explored whether biomaterial-assisted secondary healing yields postoperative speech-related outcomes comparable to primary closure in selected patients; the findings are preliminary and should be interpreted cautiously. By analyzing the demographic and clinical characteristics of the patient groups, alongside detailed speech evaluations, we aimed to determine the impact of these surgical techniques on postoperative recovery and functional outcomes.

The functional outcomes are particularly noteworthy. No significant differences were found in U-TAP scores, MDVP parameters, SMST scores, intelligibility, and movement between the groups. This finding is crucial as it suggests that, within the limitations of this small retrospective cohort, SH with biomaterials was not associated with an obvious deterioration in postoperative speech-related outcomes. Given the critical role of the tongue in articulation, preserving speech functionality is a primary concern in tongue cancer surgery. However, it is important to note that patients in the SH group reported higher discomfort levels. This increased discomfort could be attributed to the exposure of the wound, which may lead to pain and an elevated risk of bleeding.

Udayakumara et al. [24] used polyglycolic acid (PGA) sheets for oral mucosal defects, noting effective prevention of severe scar contracture or secondary deformity post-resection of lesions. These findings suggest biomaterials can be alternatives to traditional methods like skin grafts or free flap reconstructions, particularly in moderate-sized defects, without directly addressing functional advantages. Compared with PGA-sheet–based coverage, the present double-layer strategy conceptually separates wound-bed scaffolding from surface protection. PGA sheets are typically used as a temporary protective dressing (often combined with fibrin glue) to cover oral mucosal defects during the early healing phase, with gradual degradation over subsequent weeks. In contrast, Matriderm^® is a collagen–elastin matrix intended to interface directly with the wound bed and may support granulation and epithelial migration, whereas an overlying PGA mesh (Neoveil^®) can provide mechanical stability and reduce frictional irritation in the highly mobile tongue environment.

With respect to degradation and tissue interaction, PGA functions primarily as a resorbable barrier, while a collagen–elastin matrix is designed to act as an extracellular-matrix–like scaffold that may facilitate tissue integration and remodeling. These differences may have implications for mucosal pliability and scar contracture; however, we acknowledge that direct comparative data between PGA-only strategies and Matriderm^®/Neoveil^® double-layer regarding degradation kinetics, long-term scarring, and functional outcomes after partial glossectomy are limited.

Bhattacharya et al. [6] categorized tongue defects and assessed their impact on speech, noting significant impairments with defects involving more than one-third of the tongue. The nature of our cohort, containing predominantly smaller defects (less than one-third of the tongue), likely explains the absence of the significant differences in speech outcomes observed in our study. In other words, our results suggest that the extent of tongue resection plays a crucial role in postoperative speech and functional recovery, supporting our findings that, for smaller defects, the choice of closure technique may not substantially impact speech outcomes.

Extending this discussion, it is notable that most existing research tends to focus on functional recovery in cases of more extensive surgeries, such as hemi-glossectomy or larger resection [1, 3, 25–27]. These studies emphasize the need to assess and optimize functional outcomes after surgery, as larger resections pose a greater risk of impairing speech, swallowing, and other critical functions. However, there is a significant gap in the literature regarding the functional outcomes of SH using biomaterials for smaller lesions.

Even with smaller lesions, patients can experience functional discomfort that significantly impacts their quality of life [27, 28]. Therefore, it is essential to consider functional outcomes for both larger and smaller defects to ensure comprehensive patient care and improve overall well-being. This highlights the importance of our findings in providing alternative healing techniques that do not compromise on patient comfort and functionality, regardless of defect size.

Previous studies have contributed significantly to our understanding of the use of biomaterials for structural recovery after surgery [6, 24]. These studies primarily discuss how materials like PGA sheets, decalcified collagen scaffolds, dermal substitutes like Integra (Integra Lifesciences Corp, Plainsboro, NJ), and fibrin glue spray used in reconstructing tongue defects show effectiveness in promoting wound healing [9, 29–33]. However, these studies often overlooked critical aspects such as speech, swallowing, and overall patient comfort and satisfaction post-reconstruction. Our study addresses these gaps by examining both structural and functional outcomes, demonstrating that SH with biomaterials (a double layer of Matriderm^® and Neoveil^®) can achieve comparable results to PC, even in smaller defects.

This pilot study has several limitations. Its retrospective, single-center, non-randomized design and small sample size limit statistical power and external validity and may introduce selection bias. Treatment allocation was determined by defect characteristics and closure feasibility, making confounding by indication likely, while propensity-based adjustment was not feasible due to the limited sample size. In addition, baseline speech function was not routinely documented, preventing within-patient change analyses, and postoperative intelligibility was assessed by a single non-blinded listener. Patient-reported symptoms were extracted from routine records with limited detail, and swallowing outcomes and validated quality-of-life measures were not systematically collected. Therefore, these findings should be interpreted as preliminary and hypothesis-generating.

Future studies should predefine clinically meaningful functional endpoints and include an a priori sample-size calculation to ensure adequate power to detect differences between closure strategies. Prospective designs with standardized defect stratification and predefined criteria for selecting primary closure versus biomaterial-assisted secondary healing may reduce confounding by indication. Baseline assessments of speech and swallowing should be incorporated to allow within-patient change analyses, and patient-centered outcomes should be evaluated using validated measures of pain and quality of life. Blinded speech intelligibility ratings by multiple listeners with reporting of inter-rater reliability would further strengthen measurement validity. Longer-term follow-up focusing on tongue pliability and functional adaptation may help clarify the potential benefits and trade-offs of biomaterial-assisted healing in small-to-moderate defects.

Conclusions

In this retrospective pilot study, biomaterial-assisted SH was feasible in selected early-stage tongue cancer patients undergoing partial glossectomy, including those with relatively larger defects, and postoperative speech-related outcomes were not markedly different from those observed after primary closure. However, given baseline imbalances and confounding by indication in a small non-randomized cohort, these results should be viewed as preliminary and require confirmation in adequately powered prospective studies with standardized assessments.

Abbreviations

TNM

Tumor–node–metastasis

DOI

Depth of invasion

PC

Primary closure

SH

Secondary healing

U-TAP

Urimal test of articulation and phonology

MDVP

Multi-dimensional voice program

Hz

Hertz

NHR

Noise-to-harmonics ratio

SMST

Speech mechanism screening test

MPT

Maximum phonation time

AMR

Alternating motion rate

SMR

Sequential motion rate

C_W

Consonants (word)

C_S

Consonants (sentence)

V_W

Vowels (word)

V_S

Vowels (sentence)

PGA

Polyglycolic acid

Authors' contributions

Conceptualization, I.J.K.; methodology, J.D. and I.J.K.; investigation, J.D. and J.J.H.; speech-language assessment, K.H.Y.; writing—original draft preparation, J.D.; writing—review and editing, I.J.K and H.M. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the New Faculty Startup Fund from Seoul National University (No. 860-20250053).

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the Declaration of Helsinki. The Institutional Review Board of Seoul National University Dental Hospital reviewed the protocol and granted an exemption from formal IRB review and waived the requirement for informed consent on the grounds of minimal risk and use of de-identified data (IRB No. ERI23033).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.