Growth and Development of the Pediatric Knee Ligaments: A Quantitative Magnetic Resonance Imaging Study

Abstract

Background

Pediatric knee ligaments undergo substantial morphologic change throughout growth and development, yet normative magnetic resonance imaging (MRI)–based data across all 4 major knee ligaments remain limited. The aim of this study was to characterize developmental changes in the morphology of the anterior cruciate ligament (ACL), posterior cruciate ligament (PCL), medial collateral ligament (MCL), and lateral collateral ligament (LCL) in pediatric patients using MRI-based measurements of length, width, area, and volume.

Methods

Knee MRI examinations performed between 2015 and 2025 in patients aged 1-18 years were retrospectively reviewed. Measurements were obtained in standardized sagittal, coronal, and axial planes. For all ligaments, derived area and volume were calculated. Three raters performed all measurements, and mean values were used for analysis. Associations between ligament morphology and age, sex, height, and body mass index (BMI) were evaluated using mixed-effect ​regression.

Results

In total, 160 MRI scans met inclusion criteria (70 males and 90 females). Across age groups, all four ligaments demonstrated increases in absolute size. Mean ACL sagittal length increased from 19.7 ​mm in the 1- to 3-year group to 37.8 ​mm in the 16- to 18-year group, while PCL sagittal length increased from 27.5 ​mm to 43.4 ​mm. Total coronal MCL length increased from 45.0 ​mm to 62.5 ​mm, and total coronal LCL length increased from 37.6 ​mm to 59.7 ​mm.

Conclusions

Height was the most consistent independent predictor of ligament size, with significant positive associations across nearly all ACL, PCL, MCL, and LCL measures. BMI was also positively associated with several ligament dimensions. After adjustment, chronological age showed more limited independent effects, while sex differences were variable and ligament specific. Pediatric knee ligament morphology increases substantially throughout growth, but developmental changes are more strongly associated with somatic growth than with chronological age alone. These findings provide MRI-based reference data for the ACL, PCL, MCL, and LCL across the pediatric age spectrum and may aid interpretation of pediatric knee imaging and surgical planning.

Key Concepts

• (1)Pediatric anterior cruciate ligament (ACL), posterior cruciate ligament (PCL), medial collateral ligament (MCL), and lateral collateral ligament (LCL) dimensions increase substantially across childhood and adolescence, providing magnetic resonance imaging (MRI)–based reference data for all four major knee ligaments.
• (2)Height is the most consistent independent predictor of ligament size, suggesting that pediatric knee ligament morphology is more closely tied to somatic growth than chronological age alone.
• (3)Body mass index is positively associated with several ligament dimensions, including multiple ACL, PCL, MCL, and LCL size measures.
• (4)Developmental changes are specific to ligament ​and measure, with some cruciate and collateral ligament dimensions showing relative plateauing during adolescence.
• (5)These normative morphometric data may improve pediatric knee MRI interpretation and support individualized surgical planning in skeletally immature patients.

Level of Evidence

IV

Introduction

The stabilizing architecture of the pediatric knee is dynamic. Throughout growth, the cruciate and collateral ligaments develop within a joint undergoing physeal maturation, epiphyseal enlargement, and remodeling of surrounding osseous geometry [1]. Therefore, age-specific normative data are essential for accurate magnetic resonance imaging (MRI) interpretation and anatomically informed treatment planning in skeletally immature patients [2,3]. The need for developmental reference data has become more clinically important as the burden of pediatric knee ligament injury has risen. Population-based studies have shown that anterior cruciate ligament (ACL) tears and reconstructions in children and adolescents increased significantly over the past two decades [4,5]. In this setting, normative ligament morphometry is no longer just descriptive anatomy, but rather it has direct relevance to MRI interpretation, graft selection, tunnel planning, and physeal-respecting reconstruction strategies in patients whose anatomy is still changing [[5], [6], [7]]. Furthermore, studies have established normative growth and development for the knee bony morphology [8]; however, comprehensive pediatric ligament-focused research is limited.

Most of the existing developmental MRI literature has focused on the ACL, with early work demonstrating that the ligament undergoes age-dependent morphologic changes, including changes in angulation angle, throughout growth and development [3]. Subsequent studies refined this by showing that ACL length and width increase with age in a predictable but not strictly uniform manner [9], that ACL morphology changes differently in males and females during growth and maturation [10], and that ACL area may plateau earlier than linear dimensions, suggesting that growth of ligament size is not homogeneous across all morphometric variables [11]. Taken together, these studies establish that the pediatric ACL follows measurable developmental trajectories, but they also highlight substantial heterogeneity in cohorts, outcome variables, and measurement conventions across the literature.

Far fewer pediatric MRI data are available for the posterior cruciate ligament (PCL) and for comparative cruciate ligament development. Studies have reported that intact PCL dimensions may help predict native ACL size in skeletally immature knees and have described age-dependent changes in pediatric PCL surgical anatomy that showed that key PCL-related relationships vary across skeletal development [12,13]. More recently, age-stratified MRI measurements of both the ACL and PCL found that cruciate growth is not uniform across childhood, with the most pronounced changes occurring between 4 and 12 years of age [14]. Despite these advances, the literature on pediatric cruciate development remains weighted toward selected surgical variables or isolated ligament-specific questions, rather than a unified assessment of all major stabilizing ligaments of the knee.

The gap is even greater for the collateral ligaments. Pediatric MRI studies of the medial collateral ligament (MCL) and lateral collateral ligament (LCL) have primarily focused on the relationship of their femoral origins to the distal femoral physis ​because those relationships are critical for physeal-sparing and anatomic reconstruction. Studies have shown that the distance between the MCL origin and the femoral physis changes with age [15], while others demonstrated that both collateral ligament origins remain epiphyseal during growth and that their distances from the physis increase with age and body size [16]. However, these studies do not provide a broad description of whole-ligament morphology. As a result, the available pediatric collateral ligament literature is anatomically informative for reconstruction planning but insufficient for establishing comprehensive normative morphometric reference values.

A comprehensive MRI-based analysis for all four major knee ligaments across growth and development would address a meaningful gap in the pediatric orthopaedic and imaging literature. Such data could improve differentiation of normal developmental variation from pathology on MRI, support more individualized planning for pediatric ligament reconstruction, and provide age-aware geometric inputs for biomechanical modeling of the growing knee. The purpose of this study was to characterize developmental changes in the morphology of the ACL, PCL, MCL, and LCL on pediatric knee MRI using a standardized measurement framework. Specifically, we evaluated sagittal ACL/PCL length and sagittal/coronal widths, ACL inclination angle, coronal length components and width for the MCL and LCL, and derived area and volume for all four major ligaments.

Material and methods

Study design

This retrospective observational morphometry study was conducted after institutional review board approval. Clinically indicated knee MRI examinations obtained between January 2015 and June 2025 were screened for eligibility. Examinations were eligible for inclusion if the patient was 1-18 years of age at the time of imaging. The lower age bound of 1 year was selected because too few eligible diagnostic-quality knee MRI examinations that met the inclusion criteria were available in patients younger than 1 year to support meaningful analysis. The upper age bound of 18 years reflected the institution’s definition of pediatric care and was not intended to represent a threshold for skeletal maturity. Chronological age therefore defined the study’s pediatric sampling frame, whereas physeal status was not an inclusion criterion. Radiology reports and image review were used to exclude congenital musculoskeletal abnormalities, prior ipsilateral knee surgery, inadequate visualization of prespecified ligamentous or osseous landmarks, age outside the study range, and any diagnosed injury involving the collateral and cruciate ligaments. Studies were required to contain diagnostic-quality multiplanar sagittal, coronal, and axial images sufficient to visualize the cruciate and collateral ligaments and the landmarks necessary for the prespecified measurements. Demographic variables collected from the electronic medical record included age, sex, height, and weight. The study was conducted at a single center, and recorded race and ethnicity were also abstracted from the electronic medical record as documented in the clinical record.

All MRI examinations were acquired as part of routine clinical care. Therefore, scanner platforms and sequence parameters were heterogeneous. All measurements were performed in the picture archiving and communication system using built-in digital calipers. Linear measurements were recorded in millimeters (mm), estimated areas in square millimeters (mm²), estimated volumes in cubic millimeters (mm³), and angular measurements in degrees (°). Three investigators independently measured each parameter, and the mean of the 3 measurements was used for primary analyses.

ACL measurements

ACL measurements, including length, sagittal and coronal widths, and inclination angle, were obtained on the images that best depicted the full ligament course (Fig. 1). These measurements were selected on the basis of previously published pediatric MRI methods used to characterize ACL growth and morphology [9,14].

Figure 1.

MRI-based measurement protocol for the ACL. (a) Sagittal ACL length measured from the most anterior to the most posterosuperior aspect of the ligament. (b) Sagittal ACL width measured orthogonal to the midpoint of ligament length. (c) Coronal ACL width measured at the corresponding midsubstance location after cross-referencing between planes. (d) ACL inclination angle measured in the sagittal plane between a tangent to the anterior ACL border and a tangent to the lateral tibial plateau. ACL, anterior cruciate ligament; MRI, magnetic resonance imaging.

PCL measurements

Sagittal PCL length was measured as a two-segment curvilinear length (Fig. 2). Sagittal and coronal PCL widths were measured perpendicular to the ligament fibers at the midsubstance/crimp region (Fig. 3a-b). These methods were consistent with published pediatric MRI descriptions of PCL morphometry [7,14].

Figure 2.

MRI-based measurement protocol for the sagittal PCL length. Sagittal PCL length was calculated as a two-segment curvilinear measurement equal to the sum of segment a (femoral origin to crimp) and segment b (crimp to distal insertion). MRI, magnetic resonance imaging; PCL, posterior cruciate ligament.

Figure 3.

MRI-based measurement of PCL width. (a) Sagittal PCL width and (b) coronal PCL width. MRI, magnetic resonance imaging; PCL, posterior cruciate ligament.

MCL and LCL measurements

The MCL and LCL were identified using established pediatric MRI descriptions of the collateral ligaments and their femoral origins relative to the distal femoral physis [15,16]. The proximal origins and distal attachments of the MCL and LCL were estimated on the coronal sequence and triangulated to the axial image for confirmation. Due to the bowed structure of the MCL (Fig. 4) and LCL (Fig. 5) seen on the coronal image, their respective lengths were taken as the sum of (A+B). The “A” segment was measured parallel with its fibers from the distal femoral physis to the first bow point. The “B” segment was measured parallel with its fibers from the termination of the “A” segment to the ligament’s most distal tibial insertion point. The distal portion of the MCL was typically at the level of the proximal tibial growth plate. The resulting A+B measurement represented a standardized two-segment estimate of the visible coronal-plane ligament course rather than a reconstructed 3-dimensional centerline length.

Figure 4.

MRI measurement protocol for the MCL. (a) Total coronal-plane MCL length was measured as A+B to account for the bowed configuration of the ligament. (b) Coronal MCL width was measured at the cross-referenced proximal-to-mid ligament level. (c) Axial MCL width was measured at the widest proximal-to-mid portion of the ligament. MCL,medial collateral ligament; MRI, magnetic resonance imaging.

Figure 5.

Coronal MRI measurement protocol for the LCL. (a) Due to the bowed nature of the LCL, total coronal length was measured as A+B, where segment A extends from the distal femoral physis to the first bow point and segment B extends from the first bow point to the distal ligament insertion. (b) Coronal LCL length was measured at the cross-referenced proximal-to-mid ligament level. (c) Axial LCL width was measured at the widest proximal-to-mid portion of the ligament. LCL, lateral collateral ligament; MRI, magnetic resonance imaging.

Collateral ligament width was measured independently from ligament length. The proximal-to-mid ligament level at which the MCL and LCL appeared widest was first identified on the axial images. This anatomic level was then cross-referenced on the coronal images, and ligament width was measured at the corresponding level in both planes. The axial and coronal measurements represent approximately orthogonal diameters of the ligament and were used to estimate cross-sectional area (CSA); they were not used to determine ligament length.

CSA and volume

For each ligament, CSA ​was estimated assuming an elliptical cross section using orthogonal width measurements from two imaging planes (Eq. (1)):

$$CSA=\pi \ast \left(0.5\ast widthincoronal\right)\ast \left(0.5\ast widthinsagittal\right)$$ (Eq.1)

Estimated ligament volume was then calculated as (Eq. (2)):

$$Volume=CSA\ast ligamentlength$$ (Eq.2)

This geometric approach has been used previously in pediatric MRI cruciate morphometry and provides a pragmatic approximation of ligament size when direct 3-dimensional segmentation is not available [14].

Statistical analyses

Continuous variables were summarized as mean ​± ​standard deviation. For descriptive analyses, the mean of the three rater’s measurements was calculated for each morphometric variable within each subject and then summarized overall and by prespecified age strata (1-3, 4-6, 7-9, 10-12, 13-15, and 16-18 years). These chronological age strata were selected as a descriptive framework to represent developmental intervals across the pediatric age spectrum while maintaining adequate resolution to visualize changes during childhood, early adolescence, and late adolescence. These categories were intended for descriptive presentation only and were not considered surrogate markers of skeletal maturity or biological developmental stages. Associations between morphometric outcomes and age, sex, height, and body mass index (BMI) were evaluated using linear mixed-effect ​regression models with random intercepts for subject and rater when identifiable. Age was modeled continuously in regression analyses, whereas categorical age strata were used primarily for descriptive visualization of developmental patterns. Results were reported as β coefficients with 95% confidence intervals (CIs). Inter-rater reliability was assessed using average-measure intraclass correlation coefficients derived from a two-way random-effect, absolute-agreement framework. All ligament morphology measurements were classified as good: 0.75-0.90 [17,18].

Secondary exploratory analyses evaluated associations between recorded race/ethnicity and each of the ligament outcomes. Race analyses were limited to observations recorded as White or Black. Primary race models adjusted for age stratum, sex, and height and used subject-clustered robust standard errors. Secondary trajectory models used a spline for age with a race-by-age interaction and an omnibus test of the race terms. Benjamini-Hochberg false discovery rate (FDR) correction was applied across the 21 outcomes, with a q ​value <0.05 considered statistically significant.

Results

A total of 282 MRI examinations were screened, and 160 met the inclusion criteria (70 males and 90 females). Patients were distributed across the following age groups: 1-3 years (n ​= ​11), 4-6 years (n ​= ​30), 7-9 years (n ​= ​30), 10-12 years (n ​= ​30), 13-15 years (n ​= ​29), and 16-18 years (n ​= ​30) (Table 1). Because younger pediatric MRI examinations meeting strict inclusion criteria were less common, the youngest age group contained fewer participants than the remaining groups. Recorded race and ethnicity were available for secondary analysis. Because subgroup sizes were small and unevenly distributed across age strata, these variables were evaluated in exploratory models rather than incorporated into the primary developmental analyses.

Table 1.

Demographic characteristics of the included pediatric knee MRI cohort by age group.

Age group (years)	Sex	N	Height (cm)	Weight (kg)	BMI (kg/m2)
1-3	Female	4	86.8	13.2	17.5
	Male	7	103.8	18.0	16.7
4-6	Female	17	114.6	24.9	18.9
	Male	13	114.5	23.1	17.7
7-9	Female	18	130.4	31.3	18.4
	Male	12	132.8	33.7	19.1
10-12	Female	15	148.8	48.8	22.0
	Male	15	154.1	57.5	24.2
13-15	Female	21	162.9	64.8	24.4
	Male	8	173.6	75.2	24.9
16-18	Female	15	164.5	76.4	28.2
	Male	15	175.6	82.3	26.7

BMI, body mass index; MRI, magnetic resonance imaging.

ACL morphology

ACL dimensions increased substantially across age strata (Table 2). Mean sagittal ACL length increased from 19.7 ​± ​2.7 ​mm in the 1- to 3-year group to 37.8 ​± ​3.6 ​mm. Sagittal width increased from 3.3 ​± ​0.5 ​mm to 8.4 ​± ​1.2 ​mm, while coronal ACL width increased from 4.4 ​± ​1.0 ​mm in the youngest group to 7.5 ​± ​0.8 ​mm at 10-12 years, with relative stability thereafter. ACL inclination angle also increased from 41.2 ​± ​2.9° to 49.7 ​± ​4.7°, with the most pronounced increase occurring between 10-12 and 13-15 years. ACL area increased from 11.8 ​± ​3.8 ​mm² to 51.9 ​± ​8.5 ​mm², and ACL volume increased from 241.8 ​± ​84.1 ​mm³ to 1,998.1 ​± ​374.5 ​mm³.

Table 2.

Age-stratified ACL morphometry.

ACL
Age group (years)	N	Length (mm)	Sagittal width (mm)	Coronal width (mm)	Area (mm2)	Volume (mm3)	Inclination angle (°)
1-3	11	19.7 (2.7)	3.3 (0.5)	4.4 (1.0)	11.8 (3.8)	241.8 (84.1)	41.2 (2.9)
4-6	30	23.8 (3.6)	4.5 (0.9)	5.5 (0.9)	20.6 (6.1)	528.8 (221.1)	42.2 (4.3)
7-9	30	30.0 (3.4)	5.9 (0.9)	7.4 (1.0)	35.5 (8.6)	1,116.4 (351.3)	42.8 (3.8)
10-12	30	34.5 (3.7)	7.3 (0.9)	7.5 (0.9)	47.4 (7.1)	1,680.7 (387.4)	44.3 (4.1)
13-15	29	36.3 (3.4)	7.9 (0.9)	7.6 (1.0)	47.5 (7.1)	1,755.8 (370.1)	49.5 (2.8)
16-18	30	37.8 (3.6)	8.4 (1.2)	7.7 (0.8)	51.9 (8.5)	1,998.1 (374.5)	49.7 (4.7)

ACL, anterior cruciate ligament.

Sagittal ACL length, sagittal width, coronal width, inclination angle, estimated cross-sectional area, and estimated volume are summarized by age group. Data are presented as mean ​± ​standard deviation.

In adjusted analyses, height was positively associated with all ACL size measures (all P ​< ​.001). BMI was also positively associated with sagittal length, sagittal width, area, and volume. By contrast, age remained an independent predictor only for ACL inclination angle (β ​= ​0.682°/year, 95% CI: 0.300 to 1.06; P ​< ​.001) and sagittal width (β ​= ​0.105 ​mm/year, 95% CI: 0.023 to 0.187; P ​= ​.012). Male sex was not independently associated with ACL size measures, although males demonstrated a lower ACL inclination angle than females (β ​= ​−1.53°, 95% CI: −2.86 to −0.198; P ​= ​.025). The relationships between height and ACL length, sagittal width, and coronal width, stratified by sex, are shown in Fig. 6.

Figure 6.

Relationship between height and ACL morphometry, stratified by sex. Subject-level mean values are plotted for (a) sagittal ACL length, (b) sagittal ACL width, (c) coronal ACL width, and (d) ACL inclination angle. Solid lines with circle markers represent female patients, and dashed lines with triangle markers represent male patients. ACL, anterior cruciate ligament.

PCL morphology

PCL dimensions increased across pediatric development (Table 3). Mean sagittal PCL length increased from 27.5 ​± ​3.2 ​mm to 43.4 ​± ​3.9 ​mm. Sagittal PCL width increased from 2.6 ​± ​0.4 ​mm to 4.7 ​± ​0.8 ​mm, whereas coronal width increased from 5.4 ​± ​1.4 ​mm to 10.5 ​± ​1.3 ​mm by 10-12 years and then remained relatively stable through adolescence. PCL area increased from 11.3 ​± ​2.8 ​mm² to 36.4 ​± ​7.2 ​mm², and PCL volume increased from 327.5 ​± ​106.1 ​mm³ to 1,597.3 ​± ​411.3 ​mm³. Overall, the greatest gains in PCL size occurred between early childhood and early adolescence, with relative plateauing of coronal width, area, and volume thereafter.

Table 3.

Age-stratified PCL morphometry.

PCL
Age group (years)	N	Length (mm)	Sagittal width (mm)	Coronal width (mm)	Area (mm2)	Volume (mm3)
1-3	11	27.5 (3.2)	2.6 (0.4)	5.4 (1.4)	11.3 (2.8)	327.5 (106.1)
4-6	30	31.1 (3.4)	2.9 (0.6)	6.4 (1.3)	15.1 (4.8)	490.6 (192.0)
7-9	30	37.4 (3.6)	3.4 (0.7)	9.2 (1.4)	24.1 (6.8)	913.1 (290.3)
10-12	30	41.4 (3.6)	4.3 (0.6)	9.8 (1.3)	35.5 (7.4)	1,496.0 (401.0)
13-15	29	42.6 (4.0)	4.5 (0.6)	10.0 (1.4)	35.8 (7.4)	1,529.1 (383.2)
16-18	30	43.4 (3.9)	4.7 (0.8)	10.5 (1.3)	36.4 (7.2)	1,597.3 (411.3)

PCL, posterior cruciate ligament.

Sagittal PCL length, sagittal width, coronal width, estimated cross-sectional area, and volume. Data are presented as mean ​± ​standard deviation.

In mixed-effect ​regression, height and BMI were positively associated with all PCL size measures (all P ​< ​.001). Age was not an independent predictor of PCL length, width, area, or volume after adjustment for sex, BMI, and height. Male sex was independently associated only with a greater sagittal PCL length (β ​= ​1.49 ​mm, 95% CI: 0.446 to 2.54; P ​= ​.005), whereas no significant sex differences were observed for the remaining PCL measures. The relationships between height and PCL length, sagittal width, and coronal width, stratified by sex, are shown in Fig. 7.

Figure 7.

Relationship between height and PCL morphometry, stratified by sex. Subject-level mean values are plotted for (a) sagittal PCL length, (b) sagittal PCL width, and (c) coronal PCL width. Solid lines with circle markers represent female patients, and dashed lines with triangle markers represent male patients. PCL, posterior cruciate ligament.

MCL morphology

MCL morphology demonstrated progressive increases across pediatric development (Table 4). Total coronal MCL length (A+B) increased from 45.0 ​± ​4.0 ​mm to 62.5 ​± ​5.8 ​mm. MCL axial width increased from 1.5 ​± ​0.2 ​mm to 2.7 ​± ​0.4 ​mm, and coronal width increased from 1.8 ​± ​0.3 ​mm to 2.8 ​± ​1.2 ​mm. Estimated MCL area and volume increased from 2.3 ​± ​0.7 ​mm² to 6.0 ​± ​1.4 ​mm² and from 101.1 ​± ​33.2 ​mm³ to 378.3 ​± ​80.8 ​mm³, respectively.

Table 4.

Age-stratified MCL morphometry.

MCL
Age group (years)	N	Length A+B (mm)	Axial width (mm)	Coronal width (mm)	Area (mm2)	Volume (mm3)
1-3	11	45.0 (4.0)	1.5 (0.2)	1.8 (0.3)	2.3 (0.7)	101.1 (33.2)
4-6	30	52.1 (2.5)	1.7 (0.2)	2.0 (0.4)	2.9 (0.9)	142.4 (52.2)
7-9	30	54.2 (5.4)	2.0 (0.4)	2.2 (0.4)	3.6 (1.3)	195.7 (82.6)
10-12	30	60.0 (5.0)	2.4 (0.4)	2.4 (0.4)	4.5 (1.0)	272.6 (80.3)
13-15	29	60.8 (5.0)	2.6 (0.4)	2.6 (0.3)	5.3 (1.0)	324.9 (78.3)
16-18	30	62.5 (5.8)	2.7 (0.4)	2.8 (1.2)	6.0 (1.4)	378.3 (80.8)

MCL, medial collateral ligament.

Total coronal MCL length (A+B), axial width, coronal width,andestimated area and volume are summarized by age group. Data are presented as mean ​± ​standard deviation.

In mixed-effect ​regression, height was positively associated with total coronal MCL length A+B (β ​= ​0.304 ​mm/cm, 95% CI: 0.208 to 0.401; P ​< ​.001) and axial width (β ​= ​0.011 ​mm/cm, 95% CI: 0.005 to 0.016; P ​< ​.001). BMI was positively associated with total coronal MCL length A+B (β ​= ​0.257 ​mm per kg/m², 95% CI: 0.135 to 0.380; P ​< ​.001) ​and axial width (β ​= ​0.008 ​mm per kg/m², 95% CI: 0.001 to 0.015; P ​= ​.021). Age remained independently associated with greater MCL area (β ​= ​0.211 ​mm²/year, 95% CI: 0.045 to 0.377; P ​= ​.013) and volume (β ​= ​10.9 ​mm³/year, 95% CI: 0.389 to 21.4; P ​= ​.042), whereas total coronal MCL length A+B demonstrated a small inverse association with age after adjustment (β ​= ​−0.625 ​mm/year, 95% CI: -1.20 to −0.054; P ​= ​.032). Male sex was associated with greater total coronal MCL length A+B (β ​= ​2.91 ​mm, 95% CI 0.917 to 4.91; P ​= ​.004), while coronal width was not independently associated with age, sex, BMI, or height. The relationships between height and total coronal MCL length, axial width, and coronal width, stratified by sex, are shown in Fig. 8.

Figure 8.

Relationship between height and MCL morphometry, stratified by sex. Subject-level mean values are plotted for (a) total coronal MCL length measured as A+B, (b) axial MCL width, and (c) coronal MCL width. Solid lines with circle markers represent female patients, and dashed lines with triangle markers represent male patients. MCL, medial collateral ligament.

LCL morphology

LCL morphology increased across pediatric development (Table 5). Total coronal LCL length (A+B) increased from 37.6 ​± ​3.7 ​mm to 59.7 ​± ​6.5 ​mm. LCL axial width increased from 1.9 ​± ​0.4 ​mm to 3.1 ​± ​0.4 ​mm, and coronal width increased from 1.8 ​± ​0.4 ​mm to 3.1 ​± ​0.4 ​mm. Estimated LCL area and volume increased from 3.0 ​± ​1.1 ​mm² to 7.7 ​± ​1.8 ​mm² and from 112.4 ​± ​39.4 ​mm³ to 449.5 ​± ​138.5 ​mm³, respectively. Overall, descriptive findings showed progressive increases in LCL length, width, area, and volume across age strata, with mild plateauing of some segmental length measures in later adolescence.

Table 5.

Age-stratified LCL morphometry.

LCL
Age group (year)	N	Length A+B (mm)	Axial width (mm)	Coronal width (mm)	Area (mm2)	Volume (mm3)
1-3	11	37.6 (3.7)	1.9 (0.4)	1.8 (0.4)	3.0 (1.1)	112.4 (39.4)
4-6	30	43.3 (5.7)	2.2 (0.4)	2.1 (0.4)	3.8 (1.4)	170.1 (80.5)
7-9	30	48.5 (5.5)	2.4 (0.4)	2.4 (0.5)	4.6 (1.5)	224.7 (92.8)
10-12	30	55.7 (6.9)	2.6 (0.5)	2.8 (0.5)	6.1 (1.9)	332.8 (114.3)
13-15	29	57.3 (5.2)	3.0 (0.3)	3.0 (0.4)	7.2 (1.6)	410.4 (107.6)
16-18	30	59.7 (6.5)	3.1 (0.4)	3.1 (0.4)	7.7 (1.8)	449.5 (138.5)

LCL, lateral collateral ligament.

Total coronal LCL length (A+B), axial width, coronal width,andestimated area and volume are summarized by age group. Data are presented as mean ​± ​standard deviation.

In mixed-effect ​regression, height was positively associated with total coronal LCL length A+B (β ​= ​0.398 ​mm/cm, 95% CI: 0.333 to 0.462; P ​< ​.001), axial width (β ​= ​0.016 ​mm/cm, 95% CI: 0.010 to 0.022; P ​< ​.001), coronal width (β ​= ​0.014 ​mm/cm, 95% CI: 0.007 to 0.020; P ​< ​.001), area (β ​= ​0.057 ​mm²/cm, 95% CI: 0.034 to 0.081; P ​< ​.001), and volume (β ​= ​5.00 ​mm³/cm, 95% CI: 3.59 to 6.41; P ​< ​.001). BMI was also positively associated with total coronal LCL length A+B, axial width, coronal width, area, and volume (all P ​≤ ​.011). Age was not independently associated with most LCL measurements after adjustment, although total coronal LCL length A+B demonstrated a small inverse association with age (β ​= ​−0.757 ​mm/y, 95% CI: -1.14 to −0.375; P ​< ​.001). Male sex was not independently associated with LCL size measurements. The relationships between height and total coronal LCL length, axial width, and coronal width, stratified by sex, are shown in Fig. 9.

Figure 9.

Relationship between height and LCL morphometry, stratified by sex. Subject-level mean values are plotted for (a) total coronal LCL length measured as A+B, (b) axial LCL width, and (c) coronal LCL width. Solid lines with circle markers represent female patients, and dashed lines with triangle markers represent male patients. LCL, lateral collateral ligament.

Race and ethnicity analyses

The primary race analysis included 140 MRI observations (123 White and 17 Black) from patients aged 4-18 years. No Black observations were available in the 1- to 3-year group, and other recorded racial categories were too sparse for separate modeling. The restricted ethnicity analysis included 56 observations from White females aged 4-15 years (51 non-Hispanic and 5 Hispanic), with all Hispanic observations occurring between 4 and 13 years of age. Across the 21 ligament outcomes, no association with recorded race or ethnicity remained significant after FDR ​correction (Table 6). ACL inclination angle was nominally lower in Black than in White observations (adjusted difference: −2.76°; 95% CI: −5.41° to −0.10°; P ​= ​.042; q ​= ​0.879), and LCL axial width was nominally greater in Hispanic than in non-Hispanic observations in the restricted model (adjusted difference: 0.89 ​mm; 95% CI: 0.17-1.61 ​mm; P ​= ​.016; q ​= ​0.341). Omnibus race/age trajectory tests for total MCL and LCL length were suggestive but did not meet the FDR threshold (both q ​= ​0.073).

Table 6.

Exploratory analyses of recorded race and ethnicity.

Analysis	Analytic sample	Key outcome/test	Adjusted result (95% CI)	P; FDR q
Primary race	White vs ​Black, ages 4-18 years; n ​= ​140 (123/17)	ACL inclination angle	Black vs White: 2.76° (−5.41 to −0.10)	0.0419; 0.8788
Race trajectory	Same race sample	MCL length A+B omnibus race/age test	–	0.0069; 0.0725
Race trajectory	Same race sample	LCL length A+B omnibus race/age test	–	0.005; 0.0725
Restricted ethnicity	Non-Hispanic vs Hispanic White females, ages 4-15 years; n ​= ​56 (51/5)	LCL axial width	Hispanic vs non-Hispanic: 0.89 ​mm (0.17–1.61)	0.0162; 0.3408

ACL, anterior cruciate ligament; CI, confidence interval; FDR, false discovery rate; LCL, lateral collateral ligament.

Primary race models were adjusted for age stratum, sex, and height; ethnicity models were adjusted for age stratum and height. Benjamini-Hochberg false discovery rate correction was applied across 21 outcomes. No result met q ​< ​0.05. Race and ethnicity were treated as recorded social/demographic categories, and findings should be interpreted as associative and exploratory.

Discussion

This study provides an MRI-based description of how the major stabilizing ligaments of the pediatric knee change across growth and development using a unified morphometric framework. Across the ACL, PCL, MCL, and LCL, absolute ligament dimensions increased with age, but the pattern of change was not uniform across all measures. In the adjusted analyses, height was the most consistent predictor of ligament size, while chronological age had a more limited independent association once body size was considered. Taken together, these findings suggest that pediatric ligament development is better understood as part of overall somatic growth rather than as a process explained by age alone. This question has become increasingly relevant as pediatric ligament injury and reconstruction have become more common and as treatment has shifted toward earlier, anatomy-conscious, physeal-respecting strategies in appropriately selected patients [4,5,19]. Importantly, the cohort was defined according to the institution’s pediatric age range rather than by physeal status. Therefore, the chronological age groups should not be interpreted as biologically homogeneous maturity categories. In particular, the 16- to 18-year stratum may include both patients with remaining skeletal growth and patients with closing or closed physes.

The cruciate ligament findings were broadly consistent with established trends in the pediatric MRI literature. Prior work has shown that the ACL becomes longer, thicker, and more vertically oriented throughout childhood and adolescence ​while also suggesting that growth is not uniform across all dimensions and that some morphometric differences become more apparent during adolescence, including sex-related variation in selected measurements [3,9,10,20,21]. Our ACL findings align with these broader trends: linear size increased substantially across development, ACL inclination increased with growth, and some cross-sectional measures appeared to stabilize later than others. Importantly, prior MRI studies have demonstrated that ACL dimensions and associated intercondylar notch morphology do not increase uniformly throughout childhood, with certain parameters demonstrating slowing or plateauing during adolescence despite continued somatic growth [9,11,14,22]. These findings suggest that ligament development does not necessarily parallel overall body growth and may reflect distinct timeline patterns of local maturation and remodeling.

The present study adds to that body of work by showing that once height and BMI are considered, many apparent age-related ACL differences are attenuated. This suggests that chronological age alone may be an incomplete proxy for expected ACL morphology in a growing knee [3,9,10,20,21].

A similar interpretation applies to the PCL, although the available pediatric literature is much smaller. Existing MRI studies suggest that the PCL also undergoes substantial developmental change and that these changes may have practical importance for reconstruction planning and for estimating native cruciate dimensions in skeletally immature patients [7,12,14]. The results of this study support that general pattern, showing progressive increases in PCL morphology across pediatric development. However, as with the ACL, many of these differences were more strongly related to body size than to age after adjustment. This is an important distinction ​because it suggests that normative interpretation of pediatric PCL morphology should move beyond simple age-based expectations and instead account for the child’s overall stage of somatic growth.

The collateral ligament results may be the most important contribution of the present study. In contrast with the cruciate ligaments, published pediatric MRI work on the MCL and LCL has largely emphasized the relationship of the femoral origins to the distal femoral physis, reflecting the reconstructive importance of physeal anatomy in the growing knee [15,16]. Much less attention has been given to whole-ligament morphometry, such as segmental length, midsubstance width, derived area, or volume. That gap is meaningful because collateral ligament pathology remains clinically relevant in pediatric and adolescent athletes, even if isolated collateral injuries are less common [23,24]. By providing standardized measurements for both the MCL and LCL, this study extends the literature from attachment-site anatomy to broader structural characterization of the collateral ligaments themselves.

These findings also have several practical implications. First, they provide developmental reference data for MRI interpretation in a population where normal anatomy changes continuously across growth. For example, when MRI findings are equivocal, clinicians may use patient-specific developmental context, including age and body size, to determine whether a measured ligament dimension is consistent with expected variation or represents an abnormal finding. A smaller ligament in a shorter child may represent normal developmental anatomy rather than pathology, while dimensions outside the expected range for a patient’s developmental stage may prompt closer evaluation.

Second, they may assist with surgical planning, particularly in procedures where graft sizing, tunnel position, and physeal avoidance must be balanced against restoration of native anatomy. For example, a taller adolescent patient near skeletal maturity and a younger child of substantially smaller stature may have markedly different expected native ligament dimensions despite overlapping chronological considerations. Incorporating patient-specific size information may therefore improve understanding of expected anatomy during preoperative planning. However, these measurements represent native ligament morphology in uninjured pediatric knees and should not be interpreted as direct recommendations without further validation. The relationship between preinjury native anatomy and reconstruction parameters remains to be established, particularly in the setting of ligament injury where associated anatomic adaptations may occur. Existing pediatric MRI literature has already emphasized the value of imaging-based morphometry for estimating native ligament dimensions and informing reconstruction strategy, and broader pediatric ACL literature continues to stress the importance of accounting for remaining growth potential when operative treatment is considered [6,7,19,25]. Our findings reinforce that concept and suggest that body size should be incorporated into preoperative expectations whenever possible. In practice, this could include developing height-based normative charts or predictive models that allow clinicians to compare an individual patient’s ligament dimensions with expected values for similarly sized peers rather than relying exclusively on chronological age. Such approaches may be particularly useful in patients whose physical size and chronological age are discordant, such as shorter adolescents or taller younger patients. Finally, these data may provide developmentally appropriate geometric inputs for future biomechanical models of the growing knee [26]. Rather than scaling adult ligament dimensions to represent pediatric anatomy, future simulations may incorporate MRI-derived age- and size-specific ligament parameters to better reflect the mechanical environment experienced by children and adolescents.

An important implication of this study is that chronological age alone did not fully account for variation in pediatric knee ligament morphology. Height was the strongest and most consistent predictor across nearly all ligaments, while chronological age was comparatively less informative after adjustment. These findings suggest that future pediatric ligament reference standards may benefit from incorporating body size–based parameters, particularly height, rather than relying solely on chronological age categories. Because formal skeletal maturity and physeal status were not directly measured, these findings should be interpreted as evidence that somatic size provides information beyond chronological age rather than as a direct comparison of somatic growth and biological maturation. Height should therefore be viewed as a readily available clinical correlate of ligament dimensions, not as a substitute for formal skeletal maturity assessment.

Limitations

Several limitations should be considered when evaluating our results. This was a retrospective, cross-sectional study, so differences between age groups reflect between-subject comparisons rather than true longitudinal growth. Although age strata were used to facilitate visualization of developmental trends, sample sizes were not identical across groups. In particular, the youngest age group contained fewer participants, which may increase variability and reduce confidence in estimates for early childhood. However, regression analyses incorporated age as a continuous variable and adjusted for anthropometric factors, reducing reliance on categorical comparisons alone. Future studies with larger longitudinal cohorts across the entire pediatric age spectrum will be important for validating these developmental trajectories. In addition, the cohort was derived from clinically indicated MRIs rather than asymptomatic volunteers, which may introduce referral bias despite exclusion of knees with ligament injury, prior surgery, or major confounding pathology. Additionally, because injured knees were excluded, these findings characterize native ligament morphology rather than the anatomy encountered during reconstruction. Although normative measurements may provide context for understanding expected ligament dimensions, the native size of an injured ligament cannot be assumed from these reference values alone, and future studies are needed to determine how these measurements translate to graft selection and reconstruction strategies.

The single-center cohort also had limited racial and ethnic representation. The race analysis included only 17 Black observations, no Black observations in the youngest age stratum, and too few observations in other racial categories for separate analysis. The ethnicity analysis included only five Hispanic observations, all among White females aged 4-13 years. These imbalances limited statistical power and prevented reliable characterization of subgroup-specific developmental trajectories; therefore, the absence of FDR-significant findings should not be interpreted as evidence that ligament morphology is equivalent across racial or ethnic groups. Larger, more diverse, multicenter cohorts are needed to evaluate these associations.

In addition, area and volume were estimated from linear measurements rather than direct 3-dimensional segmentation. Formal skeletal maturity could not be determined consistently across the cohort. Most participants did not have contemporaneous hand/wrist imaging suitable for formal bone-age assessment or additional knee radiographs suitable for an independent assessment of local physeal status. Because the available additional imaging represented a limited and clinically selected subset, a subset comparison could have introduced substantial selection and timing bias and was therefore not performed. We also did not impute skeletal maturity from chronological age, height, sex, or other anthropometric measures. The 16- to 18-year age group may consequently include patients with open, closing, and closed physes, and physeal status may also vary within the younger age strata. This limitation prevented direct assessment of the independent relationship between skeletal maturation and ligament morphology. Accordingly, the finding that height was more consistently associated with ligament dimensions than chronological age should not be interpreted as evidence that height is more informative than formal bone age and physeal status. In this study, the term “pediatric” denotes the institution’s age-defined pediatric care population and does not imply that all included patients were skeletally immature.

Conclusions

The findings of this study demonstrated that height is the most consistent independent predictor of ligament size, with significant positive associations across nearly all major ligament measures. BMI was found to also be positively associated with several ligament dimensions. After adjustment, chronological age showed more limited independent effects, while sex differences were variable and ligament specific. Pediatric knee ligament morphology increases substantially throughout growth, and within this cohort, ligament dimensions were more consistently associated with somatic size, particularly height, than with chronological age alone. These findings support future development of body size–based reference tools, such as height-adjusted references or nomograms, to complement age-based interpretation of pediatric knee MRI. However, because physeal status and formal skeletal maturity were unavailable, these findings should not be interpreted as direct open-versus-closed physeal reference values. This study adds meaningful value to the pediatric knee literature by applying a common morphometric approach across the major ligaments and establishing a foundation for developmentally informed MRI interpretation, surgical research, and biomechanical modeling of the growing knee. Overall, the findings support that pediatric knee ligaments develop substantially throughout growth ​but that these changes are best interpreted in relation to somatic growth and biologic development rather than chronological age alone.

Author contributions

Aidan Scheible: Writing – review & editing, Methodology, Data curation, Conceptualization. Joseph A. Magoline: Writing – review & editing, Methodology, Formal analysis, Data curation. Julienne Ryland: Writing – review & editing, Methodology, Formal analysis. Alexandria Rundell: Writing – review & editing, Data curation. Michael Oravec: Writing – review & editing, Methodology, Formal analysis, Data curation. Kristen Smith: Writing – review & editing, Resources, Methodology, Formal analysis, Data curation, Conceptualization. Prateek Madaan: Writing – review & editing, Investigation, Formal analysis, Data curation, Conceptualization. Tracey Mehlman: Writing – review & editing, Formal analysis, Data curation, Conceptualization. Alexandria Mallinos: Writing – review & editing, Writing – original draft, Supervision, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization.

Ethics approval and consent

The authors ​declare that no patient consent was necessary as no images or identifying information is included in the article.

The authors ​declare that this retrospective study was approved by the institutional review board (IRB #2025-058). A waiver of informed consent was granted.

Funding

No funding was received in support of this research.

CRediT authorship contribution statement

Aidan Scheible: Writing – review & editing, Methodology, Data curation, Conceptualization. Joseph A. Magoline: Writing – review & editing, Methodology, Formal analysis, Data curation. Julienne Ryland: Writing – review & editing, Methodology, Formal analysis. Alexandria Rundell: Writing – review & editing, Data curation. Michael Oravec: Writing – review & editing, Methodology, Formal analysis, Data curation. Kristen Smith: Writing – review & editing, Resources, Methodology, Formal analysis, Data curation, Conceptualization. Prateek Madaan: Writing – review & editing, Investigation, Formal analysis, Data curation, Conceptualization. Tracey Mehlman: Writing – review & editing, Formal analysis, Data curation, Conceptualization. Alexandria Mallinos: Writing – review & editing, Writing – original draft, Supervision, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization.

Declaration of competing interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.