Age Thresholds and Gender Differences in Midpalatal Suture Maturation: A Cross-sectional Observational Study with CBCT for Maxillary Expansion Guidance

Abstract

Objective

This study analyzed 3162 CBCT images to explore age and gender features in midpalatal suture (MPS) maturation staging, and identify the statistically significant age threshold for MPS ossification.

Methods

Axial CBCT sections of 3162 participants (1690 females, 1472 males, aged 3–90 years) were classified into six MPS maturation stages (A, B, C, D1, D2, E) based on Angelieri’s criteria. The Youden index defined the diagnostic age threshold for MPS maturation; chi-squared tests assessed correlations between age, gender and stages. Kolmogorov-Smirnov, t-test and Wilcoxon test analyzed subgroup data distribution similarity, with statistical significance set at p < 0.05.

Results

The optimal age threshold for distinguishing MPS stages A, B, C and D was 11.2 years (males:11.2 years, females:10.75 years), with high diagnostic sensitivity and specificity. Age correlated significantly with MPS maturation (p < 2.2e-16). D2 stage, morphologically intermediate between D1 and E yet distinct, had a significantly different age distribution from both (p < 0.05), warranting independent staging. Adolescents (0–20 years) showed significant gender differences (p < 0.05), with females maturing earlier.

Conclusion

The critical MPS ossification age threshold is 11.2 years, supporting reduced unnecessary radiological examinations in younger patients. MPS maturation advances with age and is earlier in females, especially in 0–20 years.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s00784-026-07093-2.

Clinical Significance: Defining the threshold age of MPS ossification provides theoretical guidance for clinical expansion strategies and reduces unnecessary radiographic examinations in adolescents.

Introduction

The MPS is a fibrous joint located along the midline, connecting the two maxillary halves. It extends anteriorly from the incisive foramen to the posterior nasal spine This structure serves as a critical growth site for transverse maxillary development, facilitating coordinated dental arch expansion and occlusal alignment through osteogenic activity at the suture margins [1]. During childhood and adolescence, active bone apposition along the MPS enables gradual maxillary widening. MPS ossification is a gradual, age-dependent process progressing from childhood into late adolescence. Fusion typically begins in the palatine bone, with maturation proceeding from posterior to anterior. Complete fusion typically occurs by late adolescence or early adulthood, marking the transition from dynamic growth to structural consolidation [2].

Current research continues to explore methods for predictive and interventional strategies in maxillary development through monitoring the ossification of the MPS [3]. Cone-beam computed tomography (CBCT), a low-radiation imaging modality, enables comprehensive three-dimensional reconstruction of the morphology and bone density distribution of MPS [4]. Angelieri et al. proposed a CBCT-based classification system, dividing suture closure into five stages (A to E), stratifying suture closure progression and informing clinical maxillary expansion protocols [5]. However, the representativeness and clinical utility of this classification system remain contested. Furthermore, under clinical low-dose protocol, no single imaging workflow can simultaneously achieve low radiation exposure, high spatial resolution and effective artifact suppression, which limits diagnostic performance of CBCT for subtle bony lesions and quantitative bone analysis [6].

Insufficient maxillary width development in adolescents is a significant clinical concern. Tooth-supported maxillary arch expansion performed before complete palatal suture ossification reduces risks such as periodontal damage, root resorption, alveolar bone fenestration and gingival recession while enhancing biomechanical efficiency [7]. However, the exact correlation between MPS staging and age, as well as the reliability of this staging in guiding clinical treatment strategies remains unrevealed [8]. Additionally, the rate and pattern of facial development in children, particularly in East Asia and southern China, may be influenced by regional factors such as climate and diet.

This retrospective cohort study analyzed 3162 CBCT images of individuals aged 3 to 90 years, examining MPS stage distribution characteristics across gender and age subgroups, with a specific focus on the age range associated with the rapid fusion and ossification of the palatal suture. Based on the common imaging features of the palatal suture, the staging criteria were refined, offering a more streamlined and clear approach. The primary objective of this study is to establish a data-driven framework grounded in these findings to minimize unnecessary CBCT reassessment in adolescents, optimize intervention timing, reduce iatrogenic risks, and improve orthodontic treatment outcomes.

Materials and methods

The CBCT data source were obtained from our hospital and the ethical approval has been granted by the local ethics committee.

The inclusion criteria: (1) The MPS is fully encompassed within the field of view; (2) Authentic and comprehensive patient demographic data.

The exclusion criteria: (1) Presence of artifacts that impair the evaluation of the palatine suture; (2) Incomplete or inaccurate demographic information; (3) Images that do not fully include the maxilla; (4) Presence of cleft lip and palate, supernumerary teeth, cysts, or other conditions affecting the clarity of the palatal cleft; (5) Severe asymmetry in maxillary development.

Images were acquired using a CBCT scanner (OP 3D VISION, Dental Imaging Technologies Corporation). Scanning parameters were: tube voltage 120 kVp, tube current 5 mA, total exposure time 7.4 s. A cylindrical field of view of 16 *10 cm was selected, with isotropic voxel size of 0.25 mm and an average effective dose of 985.6 mGy-cm2. All participants were required to sit upright with the Frankfort horizontal plane aligned parallel to the ground during the scanning process. All images are exported in DICOM format and imported into Invivo7 (Osteoid Inc., USA) for analysis. Age and gender data were extracted algorithmically and compiled into a dataset. All images were independently and blindly evaluated by two experienced dental clinicians—one general dentist (Y. S.) and one orthodontic specialist (X. Y.). Both dentists possessed over three years of independent clinical practice experience involving CBCT-based radiographic interpretation and diagnosis. Prior to participating in this study, the two clinicians received standardized training delivered by an oral radiologist on the imaging characteristics and evaluation criteria of the midpalatal suture, and both passed the corresponding assessment. Interpreter consistency was assessed, and intra-interpreter error was evaluated one week later using 100 randomly selected images. Kappa statistics were used to evaluate intra- and inter-observer agreement, interpreted according to the Landis and Koch scale [9].

Each CBCT image was interpreted according to the following staging criteria, based on Angelieri’s theory [5]: First, the cursor was positioned at the patient’s midsagittal plane in the coronal and axial views. Then, in the sagittal view, the participant’s head was adjusted so that the anteroposterior long axis of the palate aligned with the horizontal line. The maturation stage of the MPS was then assessed using the central axial slice in the superoinferior dimension. For a curved palate, two central axial slices were obtained for classification: one along the axis of the posterior palate and another along the axis of the anterior palate, both contributing to the evaluation of the maturation stage.

Stage A: The MPS appears as a nearly straight, high-density line with little to no interdigitation.

Stage B: An irregular and scalloped suture maintaining a high-density appearance, with partly shown as two parallel, scalloped high-density lines separated by small low-density spaces.

Stage C: Two parallel, scalloped high-density lines separated by small low-density spaces within the maxillary and palatine bones. This pattern extends between the incisive foramen and the palatino-maxillary suture, as well as posterior to it. The suture may appear either straight or irregular.

Stage D: Para-sutural bone density is increased compared to the surrounding maxillary bone. Fusion has not yet fully occurred in the maxillary portion and remnants of two high-density lines separated by small low-density spaces may still be present.

Stage E: Complete fusion. The suture cannot be observed in at least part of the maxilla, and bone density matches that of the surrounding palatal regions.

Interestingly, during the image interpretation process, clinicians observed a higher number of cases at the transition between Stages D and E. Given the general clinical belief that the success rate of arch expansion declines after Stage D, this study further subdivides Stage D into two subtypes D1 and D2, each with distinct imaging characteristics.

The schematic diagram of the new staging and the representative CBCT cross-sectional images are shown below (Figs. 1 and 2).

Fig. 1.

Schematic diagram of imaging morphology of 6 MPS stages

Fig. 2.

Representative images of all six MPS maturation stages

In addition to the conventional five-stage classification system, two supplementary subtypes were incorporated into Stage D. Stage D1: The suture consists of two parallel, scalloped high-density lines separated by small low-density spaces within the maxillary bones. However, the posterior portion exhibits a pattern more characteristic of Stage D. In the palatine bones, the suture may no longer be visible, while parasutural bone density is increased compared to the surrounding palatine bone. Stage D2: The entire suture is no longer visible, resembling Stage E. Instead, the bone adjacent to the midline exhibits high-density linear bands extending from the incisive foramen to posterior palate, with significantly increased density compared to the surrounding maxillary bone.

Statistical analysis

Data are presented in Table 1 as absolute (number) and relative (proportion) frequencies.

Table 1.

The optimal age threshold for distinguishing the progression of palatal ossification based on Youden index

Stage	Group	Age cut off	Youden index	Sensitivity	Specificity
ABC&D	All	11.2012	0.6895	0.7794	0.9098
ABC&D	Male	11.2012	0.7270	0.7913	0.9358
ABC&D	Female	10.7513	0.6526	0.7369	0.9158

Using age to diagnose and distinguish maturational stages is a binary - classification problem. In biomedical research, ROC curves are used to evaluate disease - marker effectiveness. Each point on the ROC curve shows the relationship between sensitivity (q) and specificity (p) at a selected threshold [10].

Sensitivity (q) is the proportion of true positives correctly predicted, calculated as

.

where TP is true positives and FN is false negatives. Specificity (p) is the proportion of true negatives correctly predicted,

with TN as true negatives and FP as false positives. The Youden index (J) is used to find the optimal threshold in ROC curves:

When the biomarker of maturational stage and age are monotonically related, as the age cutoff rises, sensitivity drops while specificity rises, showing a trade - off. The optimal cutoff maximizes the Youden index, correctly classifying the most individuals without over- or under- diagnosing.

In all contingency tables, our age groups are classified as 0-<10, 10-<15, 15-<20, 20-<25, and 25-<30 (intervals left-open, right-closed). The subgroups of interest include: (1) individuals aged -<20 years; (2) combining A, B, and C into one group; (3) all groups diagnosed with D and E. The chi-squared tests utilized include the Pearson’s Chi-squared test and the Fisher exact test to examine the independence between age and maturational stages, as well as sex and maturational stages. When the proportion of frequencies less than 5 in the contingency table exceeds 20%, the Fisher exact test is employed. The Mann-Whitney U test was used to assess the potential variance in sex at the ages of 10–20. Additionally, given that the subgroups D1 and D2, as proposed in this paper, seem to exhibit some differences and similarities in distribution with E from previous sections, this section further tests the similarity in data distribution among subgroups using the Kolmogorov-Smirnov test, t-test(for mean comparation), and Wilcoxon test(for median comparation). A 5% significance level (P < 0.05) was adopted. All analyses were performed in R.

Results

Kappa coefficients for intra- and inter-interpreter agreement were 0.93 and 0.92, respectively, indicating nearly perfect reliability. Among 3162 CBCT cases (age range: 3–90 years, mean: 30.13 years), 1,472 were male (mean age: 28.16 years) and 1,690 were female (mean age: 31.83 years). No significant difference in age was observed between genders. (D = 0.0.060347, p = 0.168 (Asymptotic two-sample Kolmogorov-Smirnov test)) The distribution of maturation stages and age groups of the MPSs in the total sample is shown in Fig. 3.

Fig. 3.

Heat map for age and staging diagnosis. Figure 3A separates the classic D stage into two subgroups, D1 and D2, presenting their respective age distributions, while Fig. 3B illustrates the age distribution across the five classic stages

For clinical practice, MPS in stages A-C supports effective early rapid expansion. Using a large dataset, we calculated the Youden index and aimed to minimize the false positive rate in stage D, a balancing diagnostic sensitivity and specificity as well. We identified the key threshold for the transition of the MPS from stages A, B, and C to stage D, with an overall age of 11.2 years old, including 11.2 years old for males and 10.75 years old for females. (Table 1) In the ABC&D stage, the Youden indices for the entire group, males, and females were all relatively high (> 0.65), indicating that at these age cut-off points, the disease marker was able to differentiate between various maturational stages effectively.

From the age distribution in Fig. 4C and D, it is evident that the traditional D and E stages show similar distribution patterns. In figure with 6 stages, however, the D1 and D2 subgroups exhibit clearly distinct patterns while the D2 subgroup and E patterns look quite analogous. To assess the necessity for subgroup differentiation, three statistical tests were conducted between subgroups D1, D2, and E: the Kolmogorov-Smirnov test, mean t-test, and median Wilcoxon test. Across all three statistical tests, significant differences were observed in the age distributions among Stage D1, Stage D2, and Stage E (Table 2). This suggests that these three stages correspond to distinct degrees of MPS ossification, each exhibiting unique age-related characteristics. This finding indicates that the D-to-E staging continuum may harbor clinically relevant age inflection points, which have the potential to revise the current understanding of the age threshold for rapid maxillary expansion (RME). In turn, this conclusion corroborates the statistical validity and rationale for subdividing Stage D into D1 and D2, and may serve as a decision-making reference to guide the selection between the miniscrew-assisted rapid maxillary expansion (MARPE) and traditional surgically assisted rapid maxillary expansion (SARME) protocols.

Fig. 4.

Violin plot on age distribution of different diagnostic stages for different gender groups. F, female; M, male

Table 2.

Examinations of the age distributions of Stage D1, D2, and E using the Kolmogorov-Smirnov test, Mean test (t test), and Median test (Wilcoxon test). To investigate the age and gender differences between the six stages, considering clinical significance, data distribution balance, and statistical testing effectiveness, patients aged 0–30 years will be selected from the total sample for further analysis and testing. Altogether 1741 patients were selected, with 880 female and 861 male participants aged 0–30 years, presented in Table 3

D1, D2	the Kolmogorov-Smirnov test	Mean test(t test)	Median test(the Wilcoxon test)
	D = 0.26512, p-value < 2.2e-16	t = -5.487, df = 401.66, p-value = 7.25e-08	W = 78,329, p-value < 2.2e-16
D2, E	D = 0.22341, p-value = 1.592e-07	t = 5.2846, df = 421.3, p-value = 2.025e-07	W = 48,341, p-value = 4.276e-08
D1, E	D = 0.45582, p-value = 4.071e-06	t = 12.619, df = 629.58, p-value < 2.2e-16	W = 19,940, p-value < 1.848e-07
D, E	D = 0.35179, p-value < 2.2e-16	t = 10.784, df = 778.51, p-value < 2.2e-16	W = 126,670, p-value < 2.2e-16

Table 3.

Distribution of the MPS maturation stage classifications by age group and gender. Values are presented as n [%]. F, female; M, male

Age_ group	Sex_ group	A	B	C	D1	D2	E
0–10	F + M	86 [0.1321]	370 [0.5684]	148 [0.2273]	23 [0.0353]	7 [0.0108]	17 [0.0261]
0–10	F	34 [0.1181]	159 [0.5521]	66 [0.2292]	18 [0.0625]	4 [0.0139]	7 [0.0243]
0–10	M	52 [0.1433]	211 [0.5813]	82 [0.2259]	5 [0.0138]	3 [0.0083]	10 [0.0275]
10–15	F + M	17 [0.0338]	90 [0.1789]	127 [0.2525]	145 [0.2883]	69 [0.1372]	55 [0.1093]
10–15	F	10 [0.0418]	28 [0.1172]	71 [0.2971]	60 [0.251]	33 [0.1381]	37 [0.1548]
10–15	M	7 [0.0265]	62 [0.2348]	56 [0.2121]	85 [0.322]	36 [0.1364]	18 [0.0682]
15–20	F + M	2 [0.0093]	4 [0.0187]	35 [0.1636]	53 [0.2477]	45 [0.2103]	75 [0.3505]
15–20	F	1 [0.0083]	4 [0.0331]	15 [0.124]	25 [0.2066]	22 [0.1818]	54 [0.4463]
15–20	M	1 [0.0108]	0 [0]	20 [0.2151]	28 [0.3011]	23 [0.2473]	21 [0.2258]
20–25	F + M	0 [0]	2 [0.012]	5 [0.0299]	29 [0.1737]	36 [0.2156]	95 [0.5689]
20–25	F	0 [0]	0 [0]	2 [0.019]	13 [0.1238]	22 [0.2095]	68 [0.6476]
20–25	M	0 [0]	2 [0.0323]	3 [0.0484]	16 [0.2581]	14 [0.2258]	27 [0.4355]
25–30	F + M	0 [0]	0 [0]	4 [0.0194]	24 [0.1165]	45 [0.2184]	133 [0.6456]
25–30	F	0 [0]	0 [0]	3 [0.0236]	19 [0.1496]	24 [0.189]	81 [0.6378]
25–30	M	0 [0]	0 [0]	1 [0.0127]	5 [0.0633]	21 [0.2658]	52 [0.6582]

Figure 5 shows the distribution of six MPS maturation stages stratified by age (0–10, 10–15, 15–20, 20–25, 25–30 years) and sex. Stage B predominated (57%) in children aged 0–10 years. The 10–15 year group was dominated by Stage D1 (29%), followed by Stage C (25%). Stage E was the most frequent stage among participants aged 15–30 years. Females exhibited lower proportions of Stages A, B and C than males in the 10–15 and 15–20 year cohorts. All subgroups under 30 years yielded P < 0.05, confirming a significant correlation between age and MPS maturation stage (Supplementary Table 1).

Fig. 5.

Proportion of six MPS maturation stages across age groups and sexes. F, female; M, male

Furthermore, this association persists between males and females. Further statistical analysis was conducted on the differences in stage distribution between the two genders across 5 age groups. (Table 4) A significant difference was detected in the distribution of MPS maturation stage diagnostic based on sex among patients between 0 and 30 years. (P = 1.132e-10) Tracing the specific age subgroups where gender differences reveal, 0–10, 10–15 and 15–20 subgroups all exhibit significant distribution variations between sexes. (Table 5) The Mann-Whitney U test revealed that the maturation age in males was significantly lower than that in females (W = 12156, one-tailed p = 0.0066). With an effect size of r = 1.11, this finding indicates a substantially large practical significance of the observed difference.

Table 4.

Distribution of the improved MPS maturation staging by gender. Values are presented as n [%]. F, female; M, male

Sex	Stage A	Stage B	Stage C	Stage D1	Stage D2	Stage E	Total N
F	45[5.1]	191[21.7]	157[17.8]	135[15.3]	105[11.9]	247[28.1]	880
M	60[7]	275[31.9]	162[18.8]	139[16.1]	97[11.3]	128[14.9]	861

Table 5.

Differences in maturation stage distribution between genders across age groups

Age Groups	Stage Groups	Test	Statistics	P-value
0–30	A, B,C, D1,D2,E F, M	Pearson’s Chi-squared test	55.3	1.132e-10
0–10	A, B,C, D1,D2,E F, M	Fisher’s Exact Test		0.02691
10–15	A, B,C, D1,D2,E F, M	Pearson’s Chi-squared test	24.969	0.0001413
15–20	A, B,C, D1,D2,E F, M	Fisher’s Exact Test		0.00324
20–30	A, B,C, D1,D2,E F, M	Fisher’s Exact Test		0.2467

Discussion

Early maxillary expansion leverages the growth potential of unossified MPS, allowing timely intervention with minimal invasiveness. In general, children with sutures (stage A-C) are suitable for tooth-supported expansion, which avoids surgical intervention or bone-anchored devices while maintaining high efficiency [1, 11, 12]. Conversely, in cases of advanced ossification (Stages D-E), tooth-supported devices predominantly produce dental effects, carrying risks such as root resorption, excessive tipping, and periodontal compromise.In such cases, micro-implant-assisted skeletal expansion (MARPE) or surgically assisted rapid maxillary expansion (SARME) is required to achieve true skeletal widening [13]. This underscores the clinical imperative for accurate, pre-treatment assessment of MPS maturation.

Currently, the assessment of MPS maturity relies partly on age and gender, though these are considered less reliable indicators [14]. Histology remains the gold standard for determining the degree of ossification, but it is impractical for clinical prediction. Some researchers use the cervical spine (CVS), wrist bones, and dental jaw on radiographs to indirectly estimate the MPS mineralization [15, 16, 17]. While X-ray plain film examination offer a low radiation, the correlation between developmental stages across different regions remain limited.

CBCT offers advantages over traditional CT, including lower radiation dose and faster scanning time, making it widely used in dentistry. It overcomes limitations of occlusal radiographs by visualizing the complex 3D structure of the maxilla and palate. Despite minimal radiation, it constitutes an additional examination. Furthermore, interpretation of the morphology of the palatal sutures from CBCT images demands substantial professional acumen from the radiologist and it entails a considerable duration [18].

Our findings are align with previous studies, particularly regarding age distribution of stages A and B [19, 20, 21]. However, many prior studies relied on limited imaging data, especially for adolescents, due to stricter CBCT regulations in Western countries [11, 22, [23]. This study leveraged the comprehensive database of our pediatric medical center, analyzing all available CBCT scans from the past three years. This large-scale dataset encompassing a wide age range significantly enhances the study’s reliability and clinical significance.

The results indicate that children under 11.2 years, typically exhibit non-ossified palatal sutures (stages A-C), making tooth-supported expansion effective and CBCT unnecessary. Beyond 11.2 years, the maxilla progressively undergoes ossification, often entering stage D1. Therefore, patients over 11.2 years old may benefit from hybrid (tooth-bone) supported expanders. The orthodontist can then make a definitive diagnosis regarding suture status based on the initial expansion response and determine if CBCT is warranted. Our findings provide robust data support for validating clinical experience and offer concrete guidance for minimizing unnecessary CBCT radiation exposure in adolescents [6, 24]. For adolescents diagnosed (clinically depends on age) as stage D1 who exhibit a poor initial response to hybrid-support expansion appliances, characterized by an inactivated expander, intolerable pain, or extensive ulcers [25], we recommend CBCT for auxiliary diagnosis to determine the necessity of SARME [26, 27]. For SARME, micro-implant anchorage placement should be optimized, potentially positioned close to the palatal side, to achieve better parallel expansion. However, if posterior dentition width is well-coordinated, stage D1 patients may avoid invasive procedures. Exploiting the biomechanical advantage that the unossified mid-maxilla expands more readily than the ossified palatine bone can result in a triangular expansion pattern. In such scenarios, anterior expansion can often be effectively managed using a tooth-supported device alone [13].

For stage D2-E patients, surgical assistance is generally indicated to facilitate lateral arch expansion while minimizing stress on abutment teeth, reducing the risk of unnecessary tipping and displacement. Based on data from individuals under 30 years of age, the Youden index was calculated to determine the age threshold (Supplementary Table 2). Although the sensitivity and specificity of the results are less optimal than those between the ABC and DE stages, 15.3 years can serve as a reference age cutoff for the degree of anterior palatal suture ossification, similar to the previously reported research [28]. This indicates that the likelihood of complete ossification of the palatal suture is significantly increased in individuals over 15.3 years of age. To achieve ideal width gain and avoid adverse side effects, invasive surgery may be unavoidable during maxillary expansion for this population. Additionally, this finding provides further warnings to both clinicians and patients regarding treatment risks.

Regarding the association between gender and the degree of palatal suture ossification, this study found significant differences in the ossification age of the palatal suture between the two genders. This result is basically consistent with conclusions reported previously [21, 26, 29], namely that palatal suture ossification in boys occurs later than in girls. This significant difference is reflected both in the age staging thresholds and the distribution of age stages between the two genders: boys exhibit a higher cutoff age for complete ossification, and within the same age range, the palatal suture staging of boys tends to be in early stages such as A, B, and C. This conclusion suggests that rapid maxillary expansion should be used more cautiously when designing expansion plans for older girls, while the applicable age range for rapid maxillary expansion can be moderately widened in the male cohort.

It is noteworthy that a small proportion (15%) of adolescents were classified as stage D2 based on CBCT morphology. Whether CBCT morphology accurately reflects ossification degree and biomechanical behavior in these cases remains unclear and warrants prospective study.

Another approach involves measuring the bone mineral density (BMD) of the maxillary posterior skeleton (MPS). It has the ability to offer valuable insights into the skeletal conditions of the maxilla, which can significantly contribute to more accurate diagnosis, treatment planning, and outcome prediction in relevant fields. However, the high demands for imaging precision and accurate image segmentation are indeed obstacles that need to be overcome [30, 31, 32].

The rapid development of computer technology and artificial intelligence (AI) in recent years has injected new vitality into this field, spawning a series of innovative evaluation methods. Building on the use of computer-generated pseudo-color images (based on grayscale values) for coronal MPS staging assessment, the research team found that combining pseudo-color images with specific texture analysis algorithms enables more accurate identification of characteristic texture patterns across different MPS stages. The early non-ossification stage is characterized by sparse textures and low color contrast, while as ossification progresses, the textures gradually become denser and the color distribution more complex. Compared with evaluation using grayscale images alone, the diagnostic accuracy is improved by approximately 10%-15% [33].

Meanwhile, AI programs dedicated to the interpretation of palatal suture CBCT images have developed rapidly. Deep learning models represented by convolutional neural networks (CNN) have performed prominently. For example, a modified model based on ResNet50 achieves an accuracy of up to 99.74% in locating slices containing complete palatal suture structures from CBCT cross-sections. Multimodal fusion AI technology integrates CBCT morphological information, bone mineral density data, and clinical variables (age, gender) to achieve more comprehensive and accurate assessment of palatal suture closure. Compared with the subjective interpretation by clinicians, the consistency is improved by 20%-30%, significantly reducing human bias. In addition, new AI models based on the Transformer architecture have begun to be applied in this field. Their powerful global feature capture capability provides a new perspective for palatal suture closure assessment, showing great potential for clinical application [34, 35].

When using CBCT imaging to benchmark against the gold standard of histological morphology for staging, all of the aforementioned methods showed poor performance in estimation. The MPSs of individuals exhibited unique ossification patterns and forked morphologies, which were difficult to fully capture and present through imaging techniques [36]. Clinical cohort studies have further demonstrated that the success or failure of maxillary arch expansion is not directly related to staging; rather, it is likely influenced by a combination of factors. Additionally, this study lacks evidence supporting a correlation between imaging evaluation and clinical outcomes of arch expansion.

Based on a comprehensive literature review and the conclusions of this study, the selection of an appropriate arch expansion strategy should rely on a thorough evaluation of individual patient conditions. These include factors such as root length, abutment tooth morphology and position, periodontal health, maxillary bone thickness, pterygopalatine suture closure, overall health status, hormone use, and genetics. In children under 11.2 years of age, rapid maxillary expansion (RME) is recommended as the preferred strategy for maxillary width expansion without additional CBCT examination in the absence of specific indications.In adolescents over 11.2 years of age, CBCT assessment serves as an optional examination for selecting maxillary expansion strategies. However, the final decision-making process still needs optimization. It may be necessary to establish a multi-dimensional evaluation system integrating these various factors to conduct a more comprehensive analysis, thereby avoiding one-sided conclusions in the selection of expansion timing and methods.

Conclusion

This study establishes a critical age threshold of approximately 11.2 years for substantial MPS ossification. MPS maturation is an age-progressive phenomenon, occurring earlier in females than in males, particularly within the 0–20 year age range.

Based on imaging morphology, an improved staging subdividing Stage D into D1 and D2 is proposed. Stage D2 differs in morphology from D1 and E, but exhibits an age distribution closer to Stage E. Stage D2 and E share a comparable level of MPS maturation and should receive similar clinical interventions.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Zisheng Tang: Supervision. Resources and Project administrationXin Yang: Conceptualization, Methodology and Writing - Review & EditingQiaoqiao Ding: Software, Validation and Data CurationZiwei Chen: Formal analysis, Data Curation and VisualizationYidan Sun: Data curation, Writing- Original draft preparation.Kaijun Gu: Validation, Data curation and Writing - Original Draft.

Funding

This study was supported by the Shanghai Municipal Health Commission (Grant No. 20214Y0123) and the Construction Project of the "Discipline PeakClimbing Plan" of Xinhua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (Grant No.XKPF2024B5002).

Data availability

All data supporting the findings of this study are available within the paper.

Declarations

Competing interests

The authors declare no competing interests.