Evaluation of bone changes in patients born premature using mandibular indices and fractal dimension analysis on dental panoramic radiographs

Abstract

Our study aims to assess mandibular trabecular bone microarchitecture and cortical bone density in prematurely born children compared to term-born peers, using fractal dimension (FD) analysis and mandibular cortical index (MCI) and panoramic mandibular index (PMI) derived from panoramic radiographs. Our study included 152 panoramic radiographs obtained for diagnostic purposes from 76 premature group (30 girls and 46 boys) and 76 control group (30 girls and 46 boys). The mean age of the premature group was 7.35 ± 1.92, and that of the control group was 7.33 ± 1.9. Analyses revealed that premature individuals had a significantly higher prevalence of deteriorated cortical bone patterns (MCI Type 2/3) compared to controls (p < 0.001). Logistic regression analysis showed that increasing age (OR: 0.541; p = 0.002) and gestational age (OR: 0.696; p = 0.045) significantly reduced the risk of developing porous bone structure. FD values in the left condyle and right interdental region were statistically significantly higher in the control group than in the patients born premature group (p < 0.05). FD analysis shows partially compromised mandibular trabecular microarchitecture in patients born premature. MCI distribution reveals greater cortical bone thinning and irregularity. These findings suggest reduced quality and long-term fragility. Clinical caution is necessary regarding alveolar resorption. Monitoring the potential long-term effects of increased preterm survival on bone quality and jawbone structure represents an important clinical requirement from both pediatric and dental perspectives.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-43597-3.

Introduction

The World Health Organization (WHO) defines preterm birth as delivery before 37 weeks of gestation or a birth weight below 2500 g. Advances in neonatal health and intensive care have substantially improved the survival rates of premature infants relative to previous decades^1,2. Childhood and adolescence represent critical periods for skeletal growth, during which approximately 90% of adult bone mass is acquired. Attaining peak bone mass and optimal strength in early life is essential for minimizing the risk of osteoporosis and fractures in adulthood³. Increasing attention is being paid to newborn bone health due to its long-term impact on skeletal development⁴. Premature infants exhibit reduced skeletal mineralization compared to term peers, predisposing them to lower bone mineral density (BMD) throughout life⁵.

Dual-Energy X-ray Absorptiometry (DXA) is considered the gold standard for assessing low bone mineral density (BMD). Despite its clinical utility and widespread use, DXA is expensive and not universally accessible. A significant limitation of DXA is its inability to differentiate between cortical and trabecular bone, which may result in misinterpretation of findings⁶. Dental panoramic radiographs (DPRs), commonly used in dental practice, can reliably indicate BMD and serve as a screening tool for osteoporotic changes⁷. DPR enables evaluation of mandibular bone mass, encompassing both cortical and trabecular components, and is also valuable for screening osteopenia and osteoporosis⁸.

Indices such as the mandibular cortical index (MCI), mental index, panoramic mandibular index (PMI), gonial index, and antegonial index are used to assess bone quality and mineral density in DPR⁹. Recent studies have shown that BMD alone is insufficient for comprehensive bone quality evaluation; assessment of trabecular microarchitecture is also necessary¹⁰. Given the complex geometry, self-similarity, and anisotropy of bone trabeculae, fractal dimension (FD) analysis has become a preferred quantitative method for evaluating trabecular bone architecture¹¹. FD analysis provides a quantitative way to evaluate the structural complexity and apparent density of jawbone trabeculae. It has some limitations but is relatively insensitive to projection geometry and radiodensity. Its non-invasive and accessible nature has led to its wider use in medicine and dentistry¹².

Previous studies have investigated BMD at the lumbar spine and femoral neck in young adults with a history of preterm birth using DXA^13,14, and one study has evaluated mandibular cortical bone thickness¹⁵. A review of the literature revealed no studies evaluating mandibular cortical bone quality in premature infants using either the Mandibular Cortical Index (MCI) or the Panoramic Mandibular Index (PMI). Furthermore, a study combining these indices with Fractal Dimension (FD) analysis is also lacking. This study aims to fill this important gap in the literature by addressing these parameters with a holistic approach. Accordingly, the goal is to comparatively evaluate the mandibular trabecular bone microarchitecture and cortical bone density of premature and term-born children through FD analysis and mandibular indices applied to panoramic radiographs. We hypothesized that preterm children would exhibit thinner mandibular cortices and more irregular trabecular structure.

Materials and methods

This retrospective observational study was collaboratively conducted by the Department of Pediatrics, Faculty of Medicine, and the Department of Pediatric Dentistry, Faculty of Dentistry, at Tokat Gaziosmanpaşa University in 2025.

Ethical approval

This study was approved by the Clinical Research Ethics Committee of Tokat Gaziosmanpaşa University Faculty of Medicine (approval date: 15.10.2025, approval no. 25-MOBAEK-361). The study complied with ethical and scientific standards. All procedures followed the Declaration of Helsinki¹⁶.

Power analysis

Before the sample size was set, a prospective (a priori) power analysis was performed. The calculation was based on the expected difference in fractal dimension values between preterm and term-born children. An effect size of d = 0.61 from previous literature guided the analysis. Using this, the minimum required sample size was calculated with G*Power for independent samples, α = 0.05 and 80% power¹⁷. Regarding the region of interest used for the sample size calculation, the power analysis was specifically based on FD values obtained from mandibular 2nd premolar–1st molar region. The analysis showed that at least 42 participants were needed per group. In this study, both groups included 76 children. This yielded an estimated power of 96% (β = 0.04).

Patient selection

Children included in the study attended the pediatric dentistry clinic for routine examinations, underwent clinical dental assessment with a mouth mirror and probe, had indicated and obtained panoramic radiographs, and had parental consent for study participation and access to medical records.

Gestational age (GA), gender, and chronological age were extracted from medical records. The study group consisted of children born preterm (< 37 weeks of gestation) with low birth weight (< 2500 g). The control group comprised healthy children born at term (≥ 37 weeks of gestation) with normal birth weight (≥ 2500 g), matched to the preterm group by chronological age and gender.

Inclusion criteria: Children aged 5–12 years who agreed to participate in the study, children/parents who answered the questions completely, and panoramic radiographs taken with a Castellini (X-Radius Trio Plus, Italy) panoramic device showing clear images of the mental foramen, inferior mandibular cortex, and boundaries of both condyles. For the study group, children born before 37 weeks of gestation or with a birth weight below 2500 g; for the control group, children born at term with a birth weight 2500 g and above.

Exclusion criteria were: Children outside 5–12 years; parents or children unwilling to participate; those unable to answer questions; those with mental retardation; incomplete medical records; systemic diseases affecting bone metabolism; past or ongoing orthodontic treatment; pathology such as cysts or tumors in the maxillofacial region; a history of trauma or surgery; positioning errors in radiographs; inability to view the inferior mandible cortex; and poor image quality due to artifacts.

Panoramic radiography

All digital panoramic radiographs were acquired using the Castellini X-Radius Trio Plus (Italy) device, following the manufacturer’s instructions and standardized settings: 70 kVp, 10 mA, 8-second exposure, and ×1.34 magnification. The vertical reference line of the device was aligned with the patient’s sagittal plane, and the Frankfurt horizontal plane was maintained parallel to the floor. Images were stored as high-resolution JPEG files (2441 × 1149 pixel, 300 dpi resolution, 8-bit depth) using MetaSoft/Planmeca Romexis 3.8.3 (Helsinki, Finland). Quantitative measurements were performed using ImageJ software (version 1.38; NIH, Maryland, USA).

Radiomorphometric indexes

MCI and PMI assessments were conducted on digital panoramic radiographs from both prematurely and term-born children. PMI was measured separately for the right and left sides. Calibration was performed by measuring the mesio-distal width of the mandibular central incisors in each radiograph.

MCI evaluates the quality of the inferior mandibular cortical bone behind the mental foramen. It is known as the Klemetti index¹⁸. Indices are classified into three categories. Type 1 means sharp, well-defined, radio-opaque, and homogeneous cortical borders. Type 2 indicates a moderately deteriorated cortex with fragmented borders. Type 3 is a severely resorbed and porous cortical bone with dense trabecular-like structures (Fig. 1)¹⁹.

Fig. 1.

Klemetti classification for morphological analysis of the mandibular inferior cortex (a) Type 1, (b) Type 2, (c) Type 3.

PMI was measured by the method of Benson and colleagues²⁰. It is the ratio between the distance from the inferior border of the mental foramen and the tangent along the mandible’s inferior border (Fig. 2).

Fig. 2.

Measurement of PMI in Mandibular Radiomorphometric Analysis. a: Thickness of the cortical bone on the perpendicular line drawn from the midpoint of the mental foramen to the tangent defining the lower border of the mandible. b: The distance between the lower border of the mandible and the mental foramen.

Fractal dimension analysis

FD analysis was performed on digital panoramic radiographs of children born prematurely and at term. An experienced examiner reviewed digitized images using ImageJ software (version 1.45s; NIH, Maryland, USA). The images were processed following White and Rudolph’s method²¹. The fractal dimension of the skeletonized image was calculated with the box- counting method (Fig. 3). For FD analysis, six regions of interest (ROI) were chosen in each radiograph. Each ROI measured 25 × 25 pixels referencing previous studies^8,19 that performed FD analysis on pediatric radiographs, and included the left and right mandibular angulus, the left and right mandibular second premolar-molar region, and the left and right mandibular condyle (Fig. 4). During selection, anatomical structures like cortical bone, tooth roots, and the mandibular canal were excluded from analysis.

Fig. 3.

Stages of fractal analysis. (a) Duplication of the original image. (b) Blurring with a Gaussian filter. (c) Subtraction of the blurred image from the original image. (d) Addition of 128 grayscale levels. (e) Conversion to a binary image using a threshold value of 128. (f) Erosion process. (g) Dilation process. (h) Inversion of the image. (i) Extraction of the skeletal structure of the contours.

Fig. 4.

Six selected ROI in panoramic radiography.

All analyses were performed by a single researcher (M.T.). Intra-observer reliability was assessed by repeating 20% of the measurements after one month.

Statistical analysis

Data were analyzed using the R program V4.4.1. The normality of the data distribution was examined using the Shapiro-Wilk and Kolmogorov-Smirnov Tests. The Mann-Whitney U test was used to analyze data that did not conform to normality in two groups. The Independent Samples T-test was used to compare data from two groups that conformed to a normal distribution. Robust Regression Analysis using the MASS package was used to identify independent variables affecting the dependent variable that did not conform to a normal distribution. The effects of independent variables on the probability of having porous bone were examined using binary logistic regression. The Monte Carlo Corrected Fisher’s Exact Test, Fisher’s Exact Test, and Yates Correction were used to examine the relationship between categorical variables, and multiple comparisons were performed using the Bonferroni Corrected Z-Test. Mean ± standard deviation and median (minimum: maximum) were used to present quantitative data. Categorical data were presented as frequencies (percentages). The significance level was set at p < 0.05.

Results

Demographic data

The study sample comprised panoramic radiographs from 152 children: 76 premature (30 girls and 46 boys) and 76 healthy term (30 girls and 46 boys). The mean age of the premature group was 7.35 ± 1.92, and that of the control group was 7.33 ± 1.9. No statistically significant difference was observed in the mean age between the groups (p = 0.945) (Table 1).

Table 1.

Examination of demographic characteristics.

	Total	Premature group	Control group	Test statistics	p
Sex
Girl (n%)‎	60(39.4)	30(39.4)	30(39.4)	0.000	1.000 *
Boy (n%)	92(60.5)	46(60.5)	46(60.5)
Age (Mean ±SD)	7.34±1.9	7.35±1.92	7.33±1.9	0.005	0.945**

^*Pearson Chi-Square Test; ^**Significance test of the difference between two means (Independent Samples t test) n(%); Mean ± Standard Deviation.

Intraclass correlation data

A good level of agreement was observed for PMI values, as indicated by intraclass correlation coefficients (p < 0.001). Moderate agreement was found for FD condyle, FD interdental, and MCI values (p < 0.001), while weak agreement was noted for FD angulus values (p < 0.001) (Table 2).

Table 2.

Intraclass correlation coefficients.

	ICC (%95 CI)	p
FD condil	0,552 (0,3 - 0,62)	<0,001
FD angulus	0,410 (0,3 - 0,58)	<0,001
FD interdental	0.507 (0,3 - 0,614)	<0,001
MCI	0.631 (0,78 - 0,921)	<0,001
PMI	0,834 (0,81 - 0,910)	<0,001

ICC: Intraclass correlation coefficient (< 0.50: Weak; 0.5–075: Moderate compliance; 0.75–0.9: Good level of fit; > 0.9: Perfect fit) (95% confidence interval).

Mandibular cortical index (MCI) data

In the MCI assessment, a significantly higher proportion of Type 1 cortical bone pattern was observed in the control group (94.7%) compared with the premature group (65.8%). This difference between the groups was statistically significant (p < 0.001) (Table 3).

Table 3.

Comparison of MCI types and PMI and FD values between groups.

Variables		Total n(%)	Premature group n(%)	Control group n(%)		Test statistics	p1
MCI
	Type 1	122(80.3)	50(65.8)	72(94.7)		20.174	<0.001
	Type 2	29(19.1)	25(32.9)	4(5.3)
	Type 3	1(0.7)	1(1.3)	0(0)
	Side	Premature group		Control group			p2
		Mean ± SD	Median [Q1-Q3]	Mean ± SD	Median [Q1-Q3]
PMI
	Right	0.36±0.091	0.341 [0.3-0.412]	0.382±0.089	0.397 [0.313-0.437]	1.555	0.120
	Left	0.418±0.106	0.401 [0.331-0.492]	0.44±0.101	0.454 [0.355-0.493]	1.235	0.217
FD
Condyle	Right	1.601±0.003	1.602 [1.601-1.602]	1.601±0.002	1.602 [1.601-1.602]	0.279	0.780
	Left	1.599±0.007	1.602 [1.6-1.602]	1.601±0.002	1.602 [1.602-1.602]	1.966	0.049
Angulus	Right	1.6±0.004	1.602 [1.6-1.602]	1.601±0.002	1.602 [1.601-1.602]	1.438	0.150
	Left	1.6±0.004	1.602 [1.6-1.602]	1.601±0.002	1.602 [1.6-1.602]	0.587	0.557
Interdental	Right	1.598±0.012	1.601 [1.599-1.602]	1.6±0.003	1.602 [1.598-1.602]	2.373	0.018
	Left	1.599±0.016	1.602 [1.601-1.602]	1.6±0.005	1.602 [1.602-1.602]	1.572	0.116

Bold values in the tables indicate statistically signifi cant results (p< 0.05).

p₁: Pearson chi-square test was used. p₂: Mann-Whitney U test was used. Q1: Quartile 1, Q3: Quartile 3.

When the distribution of MCI types by gender was examined, no statistically significant difference was observed between groups (p > 0.05). However, Type 2 cortical bone pattern was observed at a higher rate in girls (58.6%) compared to boys (41.4%); while Type 1 pattern was more common in boys (64.8%) compared to girls (35.2%). The most advanced stage, Type 3 cortical bone pattern, was detected in only one male case (Table 4).

Table 4.

Comparison of MCI types by gender.

		Total n(%)	Premature group n(%)	Control group n(%)	Test statistics	p
MCI
Type 1	Girl	43(35.2)	16(32)	27(37.5)	0.391	0.532
	Boy	79(64.8)	34(68)	45(62.5)
Type 2	Girl	17(58.6)	14(56)	3(75)	0.513	0.474
	Boy	12(41.4)	11(44)	1(25)
Type 3	Girl	0(0)	0(0)	0(0)	-	-
	Boy	1(100)	1(100)	0(0)

Bold values in the tables indicate statistically signifi cant results (p< 0.05).

Pearson Chi-Square Test was used. n(%).

Panoramic mandibular index (PMI) data

No statistically significant difference was found between right and left PMI values in term and patients born premature (p > 0.005). However, the control group showed higher mean values in both the right and left PMIs than the study group (Table 3).

Fractal dimension (FD) analysis data

According to FD analysis, values for the left condyle and the right interdental region were statistically significantly higher in the control group than in the preterm group (p < 0.005). No statistically significant differences were observed between the groups in FD values for the right condyle, left interdental, and right-left angulus regions (p > 0.05) (Table 3).

Regression analysis data

The effects of independent variables on MCI types in the premature group were examined using univariate and multivariate binary logistic regression models. When examining the univariate model results, a one-unit decrease in age increases the probability of Type 2,3 bone by 1.647 times (p = 0.004). A decrease of one birth week increases the probability of Type 2,3 bone occurrence by 1.225 times (p = 0.048). The effect of other independent variables on the probability of Type 2,3 bone occurrence was not found to be statistically significant (p > 0.050) (Table 5).

Table 5.

Results of binary logistic regression analysis for the probability of osteoporotic bone in the preterm birth group.

	MCI		Univariate		Multiple
	Healthy bone	Osteoporotic bone	OR (%95 CI)	p	OR (%95 CI)	p
Age	7,85 ± 2,01	6,40 ± 1,30	0,607 (0,433: 0,852)	0,004	0,541 (0,368: 0,794)	0,002
Gestation at birth (weeks)	32,70 ± 2,21	31,50 ± 2,72	0,816 (0,667: 0,998)	0,048	0,696 (0,489: 0,991)	0,045
Birth weight (g)	2053,60 ± 593,69	1880 ± 617,22	1 (0,999: 1)	0,235	1 (0,999: 1,002)	0,534
Sex
Female	16 (53,3)	14 (46,7)	2,479 (0,937: 6,562)	0,068	1,252 (0,363: 4,323)	0,722
Male	34 (73,9)	12 (26,1)	Reference

Bold values in the tables indicate statistically signifi cant results (p< 0.05).

OR (95% CI): Odds ratio (95% confidence interval).

When examining the multiple model results for the premature group, a one-unit decrease in age increases the likelihood of Type 2,3 bone by 1.848 times (p = 0.002). A one-unit decrease in gestational age increases the likelihood of Type 2,3 bone by 1.436 times (p = 0.045). The effect of other independent variables on the probability of porous bone was not found to be statistically significant (p > 0.050) (Table 5).

In the control group, the effect of independent variables on MCI types was found to be statistically insignificant (p > 0.005) (Supp.).

Independent variables affecting the right PMI score in the premature group were examined using Robust Regression Analysis, and the model was found to be statistically significant (F = 4.497; p = 0.003). A one-unit increase in age reduces the right PMI score by 1.865 units (p < 0.001). No statistically significant effect of other independent variables on the right PMI score was found (p > 0.05) (Table 6).

Table 6.

Analysis of the effect of independent variables on the right PMI score in the preterm birth group using robust regression analysis.

	β1(%95 CI)	Std. error	β2	Test statistic	p	VIF
Constant	43,031 (9,154: 76,908)	16,990	–	2,533	0,014	–
Age	-1,865 (-2,807: -0,923)	0,472	-0,433	-3,949	<0,001	1,072
Gestation at birth(weeks)	0,284 (-0,819: 1,388)	0,553	0,083	0,514	0,609	2,325
Birth weight (kg)	-0,002 (-0,006: 0,003)	0,002	-0,116	-0,750	0,456	2,132
Sex (Female)	0,373 (-3,883: 4,629)	2,134	0,022	0,175	0,862	1,398

Bold values in the tables indicate statistically signifi cant results (p< 0.05).

F = 4,497; p = 0,003; R² =%20, 21; Durbin-Watson = 2,054; ¹: Unstandardized beta coefficient; ²: Standardized beta coefficient; 95% CI: Bootstrap confidence interval; VIF: Variance Inflation Factor.

The effect of independent variables on the left PMI score in the premature group was found to be statistically insignificant (p > 0.005) (Supp.).

Independent variables affecting the right PMI score in the control group were examined using Robust Regression Analysis, and the model was found to be statistically significant (F = 22.067; p < 0.001). A one-unit increase in age reduces the right PMI score by 1.364 units (p = 0.001). A one-unit increase in gestational age increases the right PMI score by 2.7 units (p < 0.001). The PMI score for women is 9.365 units higher than for men (p < 0.001). No statistically significant effect of other independent variables on the Right PMI score was found (p > 0.05) (Table 7).

Table 7.

Analysis of the effect of independent variables on the right PMI score in the term birth group using robust regression analysis.

	β1(%95 CI)	Std. error	β2	Test statistic	p	VIF
Constant	-63,629 (-118,434: -8,823)	27,486	–	-2,315	0,024	–
Age	-1,364 (-2,121: -0,608)	0,379	-0,297	-3,595	0,001	1,088
Gestation at birth(weeks)	2,7 (1,3: 4,1)	0,702	0,320	3,846	<0,001	1,102
Birth weight (kg)	0,001 (-0,003: 0,004)	0,002	0,034	0,411	0,682	1,112
Sex (Female)	9,365 (6,529: 12,201)	1,422	0,526	6,584	<0,001	1,017

Bold values in the tables indicate statistically signifi cant results (p< 0.05).

F = 22,067; p < 0,001; R² =%55,42; Durbin-Watson = 2,449; ¹: Unstandardized beta coefficient; ²: Standardized beta coefficient; 95% CI: Bootstrap confidence interval; VIF: Variance Inflation Factor.

The independent variables affecting the left PMI score of the control group were examined using Robust Regression Analysis, and the model was found to be statistically significant (F = 74.539; p < 0.001). A one-unit increase in age reduces the Left PMI score by 2.618 units (p < 0.001). A one-unit increase in birth weight increases the Left PMI score by 0.008 units (p < 0.001). Women’s Left PMI score is 9.967 units higher than men’s (p < 0.001). No statistically significant effect of other independent variables on the Left PMI score was found (p > 0.05) (Table 8).

Table 8.

Analysis of the effect of independent variables on the left PMI score in the term birth group using robust regression analysis.

	β1(%95 CI)	Std. Error	β2	Test statistic	p	VIF
Constant	67,3 (27,14: 107,46)	20,141	–	3,341	0,001	–
Age	-2,618 (-3,129: -2,107)	0,256	-0,558	-10,221	< 0,001	1,098
Gestation at birth (weeks)	-0,871 (-1,902: 0,161)	0,517	-0,093	-1,684	0,097	1,122
Birth weight (kg)	0,008 (0,006: 0,01)	0,001	0,373	6,732	< 0,001	1,136
Sex (Female)	9,967 (7,93: 12,005)	1,022	0,513	9,756	< 0,001	1,020

Bold values in the tables indicate statistically signifi cant results (p< 0.05).

F = 74,539; p < 0,001; R² =%80,77; Durbin-Watson = 1,810; ¹: Unstandardized beta coefficient; ²: Standardized beta coefficient; 95% CI: Bootstrap confidence interval; VIF: Variance Inflation Factor.

Independent variables affecting FB right-left condyle, FB right-left angle, and FB right-left interdental scores were examined using Robust Regression Analysis, and the model was found to be statistically insignificant (p > 0.005) (Supp.).

The median FB right-left condyle, FB right-left angle, FB right-left interdental, and right-left PMI values obtained in premature infants did not differ according to the gestational age factor (p > 0.005). When MCI was examined in the group under 31 weeks, the Type 1% was 46.4%; in the group over 32 weeks, it was 87.9%. When the relationship between MCI and gestational age was examined, a statistically significant relationship was found (p < 0.001) (Table 9).

Table 9.

Comparison of quantitative variables according to gestation at birth (weeks) groups in preterm children.

	< 32 weeks	≥ 32 weeks	Total	Test statistic	p
FB right condyle	12 (6: 12)	12 (1: 12)	12 (1: 12)	1874,500	0,416x
FB right angulus	13 (3: 13)	13 (1: 13)	13 (1: 13)	1829,000	0,588x
FB right interdental	12 (2: 13)	13 (1: 13)	13 (1: 13)	1384,500	0,070x
FB left condyle	14 (2: 14)	14 (1: 14)	14 (1: 14)	1976,500	0,154x
FB left angulus	13 (2: 13)	13 (1: 13)	13 (1: 13)	1668,000	0,713x
FB left interdental	12,5 (3: 13)	13 (1: 13)	13 (1: 13)	1429,500	0,076x
RIGHT PMI	0,34 (0,25: 0,54)	0,37 (0,19: 0,57)	0,36 (0,19: 0,57)	1450,500	0,176x
LEFT PMI	0,4 ± 0,12	0,43 ± 0,1	0,43 ± 0,1	-1,486	0,139y
MCI
Type 1	13 (46,4)a	109 (87,9)b	122 (80,3)	22,413	< 0,001z
Type 2	15 (53,6)a	14 (11,3)b	29 (19,1)
Type 3	0 (0)	1 (0,8)	1 (0,7)

Bold values in the tables indicate statistically signifi cant results (p< 0.05).

^x Mann–Whitney U test; ^y Independent-samples t test; ^z Fisher’s exact test with Monte Carlo correction Median (minimum–maximum); Mean ± standard deviation; Frequency (percentage); ^a−b No significant difference between groups sharing the same letter.

Discussion

Since the early 1990s, advances in perinatal and neonatal care have significantly reduced mortality rates among premature infants and increased their chances of survival²². The increase in survival rates and life expectancy among premature infants has sparked interest in researching both systemic health and skeletal-bone-dental development. Today, an increasing number of premature individuals are reaching adulthood, and low bone mineral density, osteopenia, and osteoporosis risk are becoming more prevalent in this population²³. Our study found that mandibular cortical index values were significantly lower in premature children, as indicated by radiomorphometric analysis, reflecting lower bone mineral density in the mandible.

Implications of mandibular cortical index (MCI)

It has been shown that at least 80% of the calcium, phosphorus, and magnesium in the fetal skeleton accumulates during the last three months of pregnancy^24,25. As a result, skeletal mineralization is not fully achieved in low-birth-weight and premature infants, and bone geometry and bone quality are severely impaired²⁶. Regardless of birth weight, early birth has also been found to be associated with lower BMD in infancy and early childhood⁵. Numerous studies in the literature have shown that the Mandibular Cortical Index defines the porosity of the mandible, is related to mandibular bone mineral density, and is a useful method for osteoporosis screening^27–31. Consistent with the literature, our findings show that the Type 2 MCI cortical bone pattern was more frequently observed in the preterm group, while the Type 1 MCI pattern was predominant in the control group. According to the results of multivariate logistic regression analysis, age is a significant predictor of MCI type. Our study found that for every 1-unit decrease in age among prematurely born children, the risk of osteoporotic Type 2–3 MCI bone disease increased by approximately 46%.

Premature birth increases the risk of porous MCI in young individuals, but cortical porosity decreases with age. Although studies have shown that bone mineral density (BMD) gradually increases in premature infants with nutrition and mineral supplementation, even reaching normal levels, comparisons with healthy controls have reported no significant difference in BMD, yet the average BMD of premature infants is lower than that of healthy controls^32,33. One of the most notable findings of our study is the significant effect of gestational age on mandibular cortical bone quality. According to our findings, low gestational age is an independent risk factor for Type 2–3 MCI, with each week of gestational age reduction increasing the risk by approximately 30%. Our findings are consistent with the fact that a large part of bone mineralization occurs in the last trimester of the intrauterine period. Furthermore, preterm birth disrupts bone development, affecting the radiological appearance of the cortical layer.

Age and sex differences in panoramic mandibular index (PMI)

The panoramic mandibular index (PMI) is the ratio of the thickness of the mandibular cortex to the distance between the mental foramen and the inferior mandibular cortex. A thin mandibular cortex width (MCW) has been shown to be associated with decreased skeletal BMD^34,35. Whether mandibular indices can be used as a screening tool for assessing bone health in children has been a focus of researchers for years. To this end, a study examining the relationship between mandibular cortical measurements, such as the Panoramic Mandibular Index (PMI) and Mental Index (MI), and total body bone mineral density excluding the head (TBLH-BMD) determined by DEXA. The study revealed a strong, significant positive correlation between the Panoramic Mandibular Index (PMI) and bone density³⁴. Each standard deviation increase in total body bone density (TBLH-BMD) was associated with a significant increase in MI and PMI values. Genetic analyses revealed that the ODF3/BET1L/RIC8A/SIRT3 locus on chromosome 11 is associated with these mandibular indices and that this region has also been previously linked to overall bone density. Furthermore, children with a genetic predisposition for high bone density were found to have significantly higher mandibular index values, proving that the two traits share common biological pathways³⁶.

Aktuna Belgin et al.³⁷ found no significant PMI difference between the corticosteroid and healthy groups, though healthy subjects tended to have higher PMI. Other studies show PMI decreases with age in adults³⁸, and is higher in children under 10 ages³⁹. Temur et al.⁴⁰ found no significant age-based differences, similar to our study, where premature and control groups were comparable. However, premature infants had slightly lower mean PMI than controls. Our regression analysis, consistent with the literature, showed each year of age was linked to a 1.865-unit decline in right PMI. In controls, age, gender, and gestational age predicted right PMI, while birth weight had a minimal effect on left PMI compared to age or gender.

The adult-child difference in PMI is due to its calculation: PMI is the ratio of cortical thickness to the distance between the mandible’s lower edge and the mental foramen⁴¹. Peak bone mass (PBM), a major osteoporosis risk factor, is achieved by early adulthood⁴². In children, bone and mandibular height are still developing, so the shorter distance between the mandible and mental foramen leads to higher PMI values. As bone height increases faster than cortical thickness with age, PMI scores fall. This greater increase in mandibular height over cortical thickness explains the age-related drop in PMI. Lower average PMI in the premature group versus controls may indicate delayed cortical bone development, matching our MCI findings. Regression analysis links the age-related decrease in right PMI among the premature group to bone adaptation from right-dominant chewing, in line with Wolff’s Law⁴³.

Microstructural asymmetry reflected by fractal dimension (FD)

FA has been utilized in dentistry to investigate the effects of systemic diseases on jawbone structure³⁹. It has been applied as an auxiliary method to assess the impact of conditions such as familial Mediterranean fever¹⁹, Molar Incisor Hypomineralization³⁹, Cleft Lip and Palate⁴⁴, beta thalassemia major⁴⁵, sleep bruxism⁴⁶, congenital heart disease⁴⁷, celiac disease⁴⁸, and diabetes mellitus⁴⁹. These studies highlight FA’s role in early disease diagnosis, particularly in pediatric and adolescent populations³⁹.

FD values measure the complexity of trabecular bone; lower FD in the premature group suggests a simpler, more irregular bone structure. FD is positively correlated with BMD^50,51. Our study found a notable decrease in left condyle FD values in the premature group, while right condyle values were similar. Gulec et al.⁵² linked higher condyle FD in adults to unilateral chewing. Most people chew predominantly on the right side^53,54. We believe right-dominant chewing helps maintain bone mass and quality in the right mandibular condyle, preventing significant differences there. In contrast, the lack of similar stimulation on the left side may contribute to greater bone changes.

Our study found significantly lower FD values in the premature group’s right interdental region, indicating bone adaptation may work differently here. While right-side chewing loads appear to stabilize bone mass in robust areas like the condyle, they may overload the more delicate interdental trabeculae. Normally, Wolff’s Law suggests bone strengthens under load⁴³, but low initial mineralization from preterm birth may reverse this effect. As a result, the already limited bone capacity cannot increase its complexity under stress and instead develops fatigue and irregularities. Thus, the functional stimulus that protects the condyle may actually trigger trabecular breakdown in the sensitive interdental region.

Implications of premature birth on the jawbone

The literature indicates that preterm birth and low birth weight are associated with reduced bone mineral density (BMD), particularly during the growth and development period, and that these children are at risk for osteopenia and osteoporosis later in life⁵⁵. When the data from our study are evaluated as a whole, it is evident that premature birth can affect not only systemic bone maturation but also the microarchitectural features of the mandible. These findings support the idea that jawbone development follows a biological trajectory parallel to that of the skeletal system as a whole¹⁵. Therefore, careful radiographic assessment of bone quality and cortical structure during routine dental examinations in children with a history of preterm birth, monitoring of alveolar bone development, and inclusion of preventive approaches in treatment planning are critically important for long-term clinical follow-up.

The study’s contribution to the literature

Our study is unique in that it comprehensively examines mandibular cortical bone quality in patients born premature for the first time in the literature using morphometric indices such as MCI and PMI. Unlike previous studies that have often focused on single indices, our approach combines these indices with FD analysis, which provides information about bone microarchitecture and enables the simultaneous assessment of both qualitative and quantitative changes in bone structure. This multi-metric approach surpasses prior single-metric studies by providing a more holistic understanding of bone health. Furthermore, the analysis of the effects of critical clinical variables such as age, gestational age, birth weight, and gender on bone indices and the likelihood of osteoporotic bone using multivariate regression models reinforces the methodological strength and reliability of our findings.

Limitations of the study

The first limitation of our study is its single-center design and the lack of DXA (Dual-Energy X-ray Absorptiometry) data, which is considered the gold standard for bone density measurement. The second limitation is the inherent limitations of a retrospective design, such as the inability to control for confounding factors like nutritional status, vitamin D levels, and physical activity. The third limitation is the limitations of DPR itself. Even though FD is relatively insensitive to projection, 2D projection, magnification, and distortion have an effect on PMI and FD measurements. The fourth limitation of our study is the age range of 5–12 years. This age range is a period of rapid craniofacial growth; interpretation of the results requires caution, and bone change patterns may differ among sub-age groups. Therefore, future studies should adopt multi-center, prospective designs with larger sample sizes and sub-age groups to more clearly reveal the long-term effects of preterm birth on bone tissue. Furthermore, it is recommended to use a wide range of parameters, including multiple imaging modalities, to provide a more comprehensive assessment of bone microarchitecture and density.

Conclusion

Our study’s MCI, PMI, and FD analyses demonstrate that reduced bone mineral density in patients born premature is also evident at the mandibular level. In particular, the increased cortical bone porosity and pronounced microarchitectural irregularities observed in fractal analysis support the notion that premature birth may exert lasting effects on mandibular bone tissue maturation. These findings have significant clinical implications, as they highlight the need for early identification and monitoring of individuals with a history of preterm birth for potential bone health issues. Clinicians should consider incorporating regular mandibular assessments in routine follow-ups for preterm children to ensure timely intervention where necessary. This proactive approach can help mitigate the risk of dental and skeletal complications later in life, thereby improving patient outcomes. In this context, monitoring the potential long-term effects of increased preterm survival on bone quality and jawbone structure represents an important clinical requirement from both pediatric and dental perspectives.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

AC and HA designed the study. Data collection was performed by AC, MT, KD, and ST. MT and MDD conducted the fractal analysis. AC, MT, and MDD wrote the first draft of the manuscript. All authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Data availability

The data sets can be shared with researchers who wish to conduct studies upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

This study was approved by the Clinical Research Ethics Committee of Tokat Gaziosmanpaşa University Faculty of Medicine (approval date: 15. 10. 2025 approval no. 25-MOBAEK-361), in line with ethical and scientific standards. All study procedures adhered to the ethical principles outlined in the Declaration of Helsinki and were conducted in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.

Consent to participate

Written informed consent was obtained from all participating children and their parents.

Consent to publish

The authors affirm that human research participants provided informed consent for publication of the images/data in this manuscript.