Imaging study of pelvic morphology in adolescent idiopathic scoliosis: a retrospective study

Abstract

Purpose

This study aimed to examine pelvic morphology differences between adolescents with and without AIS, investigate variations among AIS patients with different scoliosis degrees and segments, and identify factors influencing pelvic rotation and tilt in relation to scoliosis.

Methods

Imaging data from 213 AIS patients and 187 normal adolescents were analyzed using the Caldwell-Moloy pelvic morphology classification. Differences in pelvic morphology were compared, and Cobb angle, vertebral rotation, pelvic rotation, and coronal pelvic tilt were measured and correlated.

Results

The non-scoliotic group had the highest percentage of Gynecoid-type pelvises (35.3%) and the lowest Platypelloid-type pelvises (14.5%). The AIS group had the highest Gynecoid-type pelvises (52.6%) and the lowest Anthropoid-type pelvises (8.9%). Significant differences were found between the groups (P < 0.001). Pelvic morphology varied among AIS patients with different scoliosis degrees (P = 0.006) but not segments (P = 0.554). Cobb angle correlated with vertebral rotation (P < 0.001).

Conclusion

Pelvic morphology significantly differs between AIS and non-scoliotic adolescents. These descriptive findings suggest a possible relationship between spinal deformity and pelvic structure. Further longitudinal or mechanistic studies are needed to determine whether pelvic morphology plays a role in AIS screening, progression monitoring, or treatment planning.

Introduction

Adolescent idiopathic scoliosis (AIS) is one of the most prevalent spinal abnormalities among adolescents, with an estimated incidence rate of approximately 2%-3% in this demographic [1]. AIS is characterized by a three-dimensional deformity in spinal morphology, affecting the coronal, horizontal, and sagittal planes [2, 3]. Its etiology is widely considered to be a multifactorial condition involving genetic, biomechanical, neuromuscular, and hormonal factors [4]. Severe scoliosis often requires surgical intervention, which carries risks and potential complications [5]. Therefore, early diagnosis and treatment are crucial in the management of AIS.

As a connecting structure that transmits longitudinal pressure and gravity from the spine to the lower limbs, changes in the shape and position of the pelvis will affect the sagittal curvature of the lumbar spine [6], walking and standing posture [7], as well as the weight-bearing capacity of the lower limbs [8], thereby affecting the overall development of the adolescent musculoskeletal system. In addition to correction of the thoracic or thoracolumbar curve in the coronal plane, successful treatment of AIS is contingent on adequate restoration of balance and contour in the sagittal plane [9]. Previous studies have mostly focused on the influence of the sagittal plane position of the pelvis on scoliosis by affecting the physiological curvature of the lumbar spine [10, 11], resulting in almost no research on the relationship between pelvic morphology and AIS. The growth and development of bones are influenced by stress, and with the development of society, changes in people’s dietary habits, nutritional status, transportation methods, and sitting time have led to corresponding changes in human skeletal development. These lifestyle changes, especially in the daily lives of adolescents, the reduction of weight-bearing and standing time, the extension of sitting time, and changes in poor posture habits can to some extent affect the morphology and structure of the pelvis [12, 13]. Considerable number of studies have shown that the skeleton of patients with scoliosis exhibits completely different deformation from that of normal adolescents [14, 15]. The same goes for the pelvis [16]. Therefore, we hypothesize that pelvic morphology is also different between AIS patients and non-scoliotic adolescent patients. The purpose of this study is to observe the pelvic morphology in AIS patients and the differences among different types of scoliosis.

Method

Participants

This retrospective case study analyzed AIS patients who had full-length spinal X-rays taken at our hospital from January 2018 to November 2024. Inclusion criteria were as follows: (1) meeting the diagnostic criteria of the Society on Scoliosis Orthopedic and Rehabilitation Treatment (SOSORT) 2016 Guidelines for the orthopedic and rehabilitation treatment of adolescent idiopathic scoliosis during growth spurt [17]; (2) aged between 10 and 18 years old; (3) All included images were taken when the patients were diagnosed with scoliosis for the first time. Exclusion criteria were: (1) History of spinal surgery, orthopaedic surgery, or spinal tumour; (2) Structural and functional abnormalities of the pelvis and hip joints; (3) History of lower limb disorders; (4) Incomplete and unclear imaging data. (5) Patients with no idiopathic causes of spinal deformity were excluded. The control group included non-scoliotic adolescents who underwent full-length spine X-rays at our hospital during the same period.

Radiographic metrics

During the X-ray image acquisition, the patient were positioned standing upright under standardized conditions: eyes gazing forward, arms relaxed at their sides, feet together and toes forward. The technician performed exposure imaging from an anterior–posterior position. After image acquisition, using post-processing software, the images are stitched into a complete spine x-ray image. The pelvic X-ray is taken in the pelvic inlet position, with the central ray directed to the midpoint of the line connecting the anterior superior iliac spines and inclined 35°-45° toward the foot side for exposure. All data were measured using Surgimap software. Pelvic X-ray is taken by the same radiologist under the same X-ray machine in an inlet view. All imaging data were measured by experienced specialists in a standardized manner to ensure consistency and minimize measurement errors [18, 19].

The classification of the pelvis was mostly based on Caldwell Moloy’s [20]: Gynecoid: The pelvic entrance is circular or elliptical in shape, with a transverse diameter equal to or slightly larger than the anterior posterior diameter, and the entrance is wide at both ends (Fig. 1a); Android: The pelvic entrance is wedge-shaped or heart-shaped, with a narrow posterior entrance and a triangular anterior entrance(Fig. 1b); Platypelloid: The pelvic entrance is horizontally elliptical, with a transverse diameter greater than the anterior posterior diameter, and the anterior and posterior parts of the entrance are almost equal (Fig. 1c); Anthropoid: The pelvis is elongated and elliptical in shape, with a greater anterior posterior diameter than the transverse diameter at the entrance. The transverse diameter is either absolutely or relatively shortened, and the two parts of the entrance are deeper and narrower (Fig. 1d). During the information processing, 3 reviewers conducted The determination of pelvic types independently. In cases of evaluation disagreement, 3 reviewers did separate reassessments and discussions until the conclusions were reached.

Fig. 1.

Pelvic morphology classification

The following parameters were additionally measured in the full-length standing spinal X-rays for all 213 AIS patients:

1. Cobb angle: the angle of the two intersecting lines drawn along the edge of the top and bottom vertebras of the curve. On the top vertebra, the line starts at the high side, is drawn along the top edge and slopes downward according to the angle of the vertebra. Similarly, on the bottom vertebra, the line starts on the low side, is drawn along the bottom edge and will slope in an upward direction.

2. Apex-vertebral rotation (AR): The evaluation of AR on radiographs was made using the Raimondi method [21].

3. Pelvic rotation (PR): the absolute value of the difference between the left-right ratio and 1 (|1-L/R|) of horizontal distance between the anterior superior iliac spine and the inferior ilium at the sacroiliac joint on the same side [22].

4. Coronal pelvic tilt (CPT) defined as the angle between the line connecting bilateral eyebrow arch of acetabulum and horizontal line (Fig. 2).

Fig. 2.

Cobb angle, apex-vertebral rotation and pelvic parameters measurement

Cobb angle severity was classified as follows: mild (10° to 20°), moderate (> 20° to 40°), and severe (> 40°). The curve type was defined based on the location of the apical vertebra: curves with an apex between T2 and T11 were classified as thoracic, those with an apex between T12 and L1 as thoracolumbar, and those with an apex distal to L1–2 as lumbar.

Statistics

Statistical analysis were performed using SPSS Statistics 25 (IBM Corp, Armonk, New York, USA). The normality of all continuous data was assessed using the Shapiro-Wilk test. The Spearman test was used to test the correlation between Cobb angle, AR, PR and CPT. The inter group differences between basic characteristics of subjects, pelvic morphology were analyzed using the Chi-square test, while the inter group comparisons of Cobb angle, AR, PR and CPT were conducted using the Kruskal-Wallis test. The significance level for two-tailed tests was set at α = 0.05.

Results

Basic characteristics

The present study included 400 subjects. There were 187 adolescents (127 females and 60 males) in the non-scoliotic group and 213 AIS patients (148 females and 65 males) in the AIS group. There is no significant difference in gender distribution between the two groups (P > 0.05). The mean age of the AIS group was 14.05 ± 2.26 years, and the mean age of the non-scoliotic group was 13.57 ± 2.45 years. There is a statistical difference in the stage of the Risser sign (P < 0.05; Table 1) between the AIS group (2.76 ± 1.79) and the non-scoliotic group (2.45 ± 1.81). The mean major Cobb angle was 23.65 ± 13.09 (range, 10.0–84.8°).

Table 1.

Basic characteristics of subjects [n (%)]

Parameters		AIS	Non-AIS	X2	P
Gender	Male	60(32.1)	65(30.5)	0.114	0.736
	Female	127(67.9)	148(69.5)
Risser sign	0	54(28.9)	46(21.6)	14.334	0.013
	1	5(2.7)	12(5.6)
	2	17(9.1)	22(10.3)
	3	54(28.9)	40(18.8)
	4	28(14.9)	54(25.4)
	5	29(15.5)	39(18.3)
Age	10	22(11.8)	7(3.3)	19.073	0.014
	11	24(12.8)	31(14.5)
	12	30(16.0)	24(11.3)
	13	17(9.1)	29(13.6)
	14	27(14.4)	25(11.7)
	15	21(11.2)	37(17.4)
	16	16(8.6)	27(12.7)
	17	17(9.1)	15(7.0)
	18	13(7.0)	18(8.5)

Comparison of pelvic morphology between males and females among all subjects

Among males, the most common type was Android (44.8%), followed by Gynecoid (28.8%), Anthropoid (15.2%), and Platypelloid (11.2%). In contrast, the Gynecoid type was predominant in females (51.3%), followed by Platypelloid (18.9%), Anthropoid (16.4%), and Android (13.4%). A statistically significant difference in pelvic morphology was found between genders (P < 0.001; Table 2). This significant gender difference in morphology distribution was maintained when analyzing the AIS (P < 0.001) and non-scoliotic control (P < 0.001) groups separately.

Table 2.

Comparison of pelvic morphology between genders [n (%)]

Groups	Pelvic Morphology	Male	Female	X2	P
All subjects	Gynecoid	36(28.8)	141(51.3)	49.29	0.000
	Android	56(44.8)	37(13.4)
	Platypelloid	14(11.2)	52(18.9)
	Anthropoid	19(15.2)	45(16.4)
Scoliosis group	Gynecoid	11(18.3)	55(43.3)	30.15	0.000
	Android	31(51.7)	19(15.0)
	Platypelloid	5(8.3)	22(17.3)
	Anthropoid	13(21.7)	31(24.4)
Non-scoliotic group	Gynecoid	25(38.5)	86(58.0)	19.84	0.000
	Android	25(38.5)	18(12.2)
	Platypelloid	9(13.8)	30(20.3)
	Anthropoid	6(9.2)	14(9.5)

Comparison of pelvic morphology between the two groups

In the AIS group, the distribution was as follows: Gynecoid (52.1%), Android (20.2%), Platypelloid (18.3%), and Anthropoid (9.4%). In the non-scoliotic group, the distribution was Gynecoid (35.3%), Android (26.7%), Platypelloid (14.5%), and Anthropoid (23.5%). There was a statistically significant difference in pelvic morphology between the two groups (P < 0.001; Table 3). The pelvic morphology showed a significant difference between the group of females with and without AIS (P = 0.005). The inter-group comparison of males also revealed a significant difference (P = 0.028).

Table 3.

Comparison of pelvic morphology between two groups [n (%)]

Pelvic Morphology	AIS	Non-AIS	X2	P
Gynecoid	111(52.1)	66(35.3)	21.55	0.000
Android	43(20.2)	50(26.7)
Platypelloid	39(18.3)	27(14.5)
Anthropoid	20(9.4)	44(23.5)

Comparison of pelvic morphology between Cobb degrees

The study population comprised 107 patients with mild AIS, 84 with moderate AIS, and 22 with severe AIS. Among mild AIS patients, pelvic morphology distribution was as follows: Gynecoid (49.5%), Android (22.4%), Platypelloid (15.0%), and Anthropoid (13.1%). The moderate AIS group demonstrated a distribution of 56.0% Gynecoid, 22.6% Android, 14.3% Platypelloid, and 7.1% Anthropoid. In contrast, the severe AIS group exhibited only two morphological types: Gynecoid (50.0%) and Platypelloid (50.0%). There was a significant difference in pelvic morphology between mild, moderate and severe AIS groups (P = 0.002). Furthermore, the statistically significant difference in pelvic morphology remained evident when comparing patients with scoliosis of the same gender but varying degrees of curvature (Table 4).

Table 4.

Comparison of pelvic morphology between Cobb degrees [n (%)]

		Gynecoid	Android	Platypelloid	Anthropoid	X2	P
Total	Mild	53(49.5)	24(22.4)	16(15.0)	14(13.1)	20.45	0.002
	Moderate	47(56.0)	19(22.6)	12(14.3)	6(7.1)
	Sever	11(50.0)	0(0.0)	11(50.0)	0(0.0)
Female	Mild	34(52.3)	9(13.8)	12(18.5)	10(15.4)	14.29	0.027
	Moderate	42(64.6)	9(13.8)	10(15.4)	4(6.2)
	Sever	10(55.6)	0(0.0)	8(44.4)	0(0.0)
Male	Mild	19(45.3)	15(35.7)	4(9.5)	4(9.5)	15.96	0.014
	Moderate	5(26.4)	10(52.6)	2(10.5)	2(10.5)
	Sever	1(25.0)	0(0.0)	3(75.0)	0(0.0)

Comparison of pelvic morphology among different segmental scoliosis

The study included 100 patients with thoracic scoliosis, 61 with thoracolumbar scoliosis, and 52 with lumbar scoliosis. The distribution of pelvic morphology in the thoracic AIS group was 50.0% Gynecoid, 20.0% Android, 22.0% Platypelloid, and 8.0% Anthropoid. The thoracolumbar AIS group distribution was 47.6% Gynecoid, 24.6% Android, 18.0% Platypelloid, and 9.8% Anthropoid, while the lumbar AIS group distribution was 61.5% Gynecoid, 15.5% Android, 11.5% Platypelloid, and 11.5% Anthropoid. Statistical analysis identified no significant difference (P = 0.554) in pelvic morphology distribution among the different segmental scoliosis. Intergroup analyses stratified by gender similarly demonstrated no statistically significant differences in pelvic morphology distribution among the various curve types (Table 5).

Table 5.

Comparison of pelvic morphology among different segmental scoliosis [n (%)]

		Gynecoid	Android	Platypelloid	Anthropoid	X2	P
Total	Thoracic	41(57.7)	7(9.9)	17(23.9)	6(8.5)	2.90	0.830
	Thoracolumbar	22(53.7)	7(17.1)	8(19.5)	4(9.7)
	Lumbar	23(63.9)	4(11.1)	5(13.9)	4(11.1)
Male	Thoracic	9(31.1)	13(44.8)	5(17.2)	2(6.9)	4.16	0.680
	Thoracolumbar	7(35.0)	8(40.0)	3(15.0)	2(10.0)
	Lumbar	9(56.3)	4(25.0)	1(6.2)	2(12.5)
Female	Thoracic	50(50.0)	20(20.0)	22(22.0)	8(8.0)	4.948	0.554
	Thoracolumbar	29(47.6)	15(24.6)	11(18.0)	6(9.8)
	Lumbar	32(61.5)	8(15.5)	6(11.5)	6(11.5)

Comparison of the Cobb angle, AR, PR, and CPT among different segmental scoliosis and Cobb degrees

Comparative analysis across different segmental scoliosis, no significant differences were observed in Cobb angle (P = 0.059), AR (P = 0.524), PR (P = 0.988) and CPT (P = 0.189). In contrast, analysis across different Cobb degrees, only Cobb angle and AR showed significant differences (P < 0.001), while no significant differences were observed in PR (P = 0.558) and CPT (P = 0.053; Table 6).

Table 6.

Comparison of parameters among different segmental scoliosis and different Cobb degrees

Groups		Cobb	AR	PR	CPT
Topographic	H	5.651	1.224	0.023	3.337
	P	0.059	0.542	0.988	0.189
Cobb degrees	H	171.904	66.591	1.166	5.857
	P	0.000	0.000	0.558	0.053

Correlations between Cobb angle, AR, PR and CPT in the AIS group

The Cobb angle has a significant correlation with AR(P < 0.001), but no significant correlation with PR (P = 0.555) and CPT (P = 0.297), There is also no significant correlation between AR and CPT (P = 0.881, Table 7).

Table 7.

Correlations between Cobb angle, AR, PR and CPT

Parameters		Cobb	AR	PR
AR	r	0.595
	P	0.000
PR	r	-0.041	-0.002
	P	0.555	0.978
CPT	r	0.072	-0.010	0.009
	P	0.297	0.881	0.896

Comparison of body mass index (BMI) across different pelvic morphology types

The mean BMI values across pelvis morphology types were as follows: Gynecoid (18.77 ± 2.99 kg/m²), Android (18.56 ± 3.17 kg/m²), Platypelloid (18.69 ± 2.64 kg/m²), and Anthropoid (17.31 ± 2.94 kg/m²). No significant difference in BMI was observed among the Gynecoid, Android, and Platypelloid pelvic types (p > 0.05). However, the Anthropoid type demonstrated a significantly lower BMI compared to each of these three types (p < 0.05; Table 8).

Table 8.

Comparison of BMI among different pelvic morphology

		Android	Platypelloid	Anthropoid
Gynecoid	t	0.492	0.165	3.174
	P	0.623	0.869	0.002
Android	t		-0.249	2.385
	P		0.803	0.018
Platypelloid	t			2.632
	P			0.010

Discussion

Evolutionary responses to selection for bipedalism and encephalization has substantially sculpted the human pelvis, resulting in a morphology substantially distinct from that in apes. Although key morphological adaptations are present by birth, the developmental-genetic mechanisms governing pelvic shape remain largely unknown. Natural selection has markedly sculpted the human pelvic morphology to accommodate the biomechanical demands unique of bipedal locomotion and the obstetric requirements associated with increased neonatal brain size [23] and other factors [24, 25]. The human pelvis is markedly derived, with its morphology evident before birth [26] and likely under a complex regulatory architecture to accommodate different demands. Currently, there is little understanding of the transcriptional and epigenetic mechanisms governing human skeletal, let al.one pelvic development [27].

AIS, the most prevalent form of structural scoliosis, is characterized by a three-dimensional spine deformation, which consisted of coronal deviation, alterations in sagittal curvature, and rotation of vertebral bodies in the horizontal plane [28, 29]. Previous studies have demonstrated variation in AIS incidence across different age groups. Huang et al. [30] reported incidence rates of 0.17% n grades 1–3, 0.82% in grades 4–6, 2.64% in middle school students, and 4.00% in high school students. Similarly, Kamtsiuris et al. also found AIS incidence rate of 6.5% in adolescents aged 11–13, which increased to 11.1% for 14–17 years old [31]. Another study showed that the highest positive rate was found in the age group of 15–16 [32]. Collectively, these studies indicate a higher incidence of scoliosis among older adolescents. Our findings are consistent with this trend, as the AIS patients were older than non-scoliotic adolescents with the same gender ratio (P = 0.014). Furthermore, our analysis confirmed the presence of sexual dimorphism in pelvic morphology among adolescents, as statistical differences in pelvic morphology was observed between adolescents of different genders. The Android type was most prevalent among male participants (44.8%), whereas the Gynecoid type was predominant among females (51.3%). This is consistent with the conclusions of previous studies [33, 34], demonstrating that fundamental differences in pelvic structure are already evident during adolescence.

A significant difference in pelvic morphology was observed between adolescents with AIS and non-scoliotic controls. Although the Gynecoid type was the most prevalent morphology in both groups, its frequency was significantly higher in the AIS group (52.1%) than in the non-scoliotic group (35.3%). Conversely, the Android type was more common in the non-scoliotic group (26.7%) compared to the AIS group (20.2%). The AIS group showed proportions of 18.3% Platypelloid and 9.4% Anthropoid, whereas the non-scoliotic group demonstrated proportions of 14.5% Platypelloid and 23.5% Anthropoid. Abitbol [12] proposed that morphological variations in pelvic development may be influenced by biomechanical factors during early development. Specifically, the Android pelvic type was reported to be associated with exposure to strenuous physical activity during adolescence. Conversely, the Anthropoid type was more frequently observed in individuals with delayed acquisition of erect posture beyond the typical age of 14 months, while the Platypelloid type showed higher prevalence in those who attained erect posture before 14 months. Fischer et al. suggested that the pattern of human pelvic dimorphism is conservative compared to the highly variable magnitude of the sex differences, which may depend partly on environmental factors such as mechanical loadings [35]. In a complementary finding, Abitbol reported that the female pelvis shows greater plasticity and susceptibility to body mass than the pelvic girdle of similarly aged males before the end of maturation [36]. Consequently, the divergent distribution of pelvic morphology observed between AIS patients and non-scoliotic adolescents in this study may be attributable to differences in lifestyle and physical activity patterns. This interpretation is supported by previous research indicating that individuals with AIS tend to engage in less vigorous physical and sporting activity compared to their unaffected peers, a behavioral difference that may have implications for overall health and quality of life [37, 38].

The observed divergence in pelvic morphology may also be associated with endocrine and metabolic factors, such as hormonal profiles and body composition. Steinetz et al. proposed that the quantity and temporal pattern of sex hormone secretion—particularly estrogen and androgen—serve as primary drivers of sexually dimorphic developmental trajectories by mediating pelvic bone remodeling. This mechanism may be particularly relevant in AIS, as patients have been shown to exhibit higher systemic estrogen levels compared to unaffected adolescents [39]. Furthermore, the presence of estrogen receptors ESR1 and ESR2 has been confirmed within the paravertebral skeletal musculature of individuals with idiopathic scoliosis, suggesting a localized pathway for estrogen-mediated influence on musculoskeletal development. Elevated expression and asymmetric distribution of estrogen receptors in deep paravertebral musculature have been associated with greater magnitude of scoliotic deformity and an increased risk of curve progression [40]. Conversely, lower circulating androgen levels may exacerbate aberrant bone development in AIS patients [41]. Furthermore, alterations in melatonin secretion concentration and circadian rhythm, which differ from patterns observed in healthy individuals, may contribute to the pathogenesis of adolescent idiopathic scoliosis and associated pelvic morphological changes by modulating the activity of osteoclasts and chondrocytes [42]. Adolescent females with AIS demonstrate lower body weight, body mass index (BMI), skeletal muscle mass, and body fat percentage compared to their unaffected peers [43]. In contrast, a longitudinal study reported that adolescents with higher body fat percentages tend to develop greater pelvic width [44]. This association may be partly explained by the influence of elevated estrogen levels associated with increased adiposity, which can promote wider pelvic development. Additionally, biomechanical stress resulting from increased body mass, as well as modifiable factors such as physical activity levels and gait patterns, may further contribute to variations in skeletal morphology. In comparison of pelvic morphology between Cobb degrees, we also found a significant difference observed across mild, moderate and severe AIS groups (P < 0.008). The Gynecoid morphology was more prevalent in moderate AIS cases (56.0%) than in mild cases (49.5%), whereas the proportion of the Android type showed no substantial difference between these severity groups. Notably, among the 22 patients with severe AIS, only Gynecoid and Platypelloid were present and accounted for a similar proportion (50%). As previously discussed, pelvic type is the result of the combined effects of hormones, body composition, physical activity and other factors. Body mass index, lean muscle mass index, and estimated bone mass index were significantly lower in the severe scoliosis group compared with those in the moderate scoliosis group and were negatively correlated with Cobb angle [45]. Analysis of patients stratified by Cobb angle severity (mild, moderate, and severe) revealed a tendency toward delayed menarche with increasing curve severity [46]. In addition, Kurnik et al. [47] found that a larger lumbar lordosis may lead to abnormal loads on the sacrum. Studies have confirmed that the lumbar lordosis is greater in patients with lumbar scoliosis [48]. These findings demonstrated that the lumbar lordosis angles may indirectly influence the pelvic morphology. We therefore speculate that the differences in pelvic types between AIS patients and non-scoliotic adolescents arise from a multifactorial etiology, integrating biomechanical forces, hormonal influences, and developmental factors.

Skeletal growth is both a biological process and a physical force, as highlighted by D’Arcy Thompson [49]. It can be influenced by physical stress, as seen in the Hueter–Volkmann Law [50], where bone growth is inhibited by compression and enhanced by tension. Crijns et al. [51] demonstrated that differential growth in the spine first causes hypokyphosis and mild lateral bending, eventually leading to scoliosis, as observed in AIS. They proposed that AIS arises from restricted differential growth between the vertebral column and surrounding muscles and ligaments. The strength of ligaments is linked to collagen cross-links, which must break under tension to allow growth [52, 53]. In AIS, this process may be impaired due to insufficient or aberrant dynamic mechanical loading, thereby hindering the necessary remodeling of collagen cross-links and restricting physiological growth [54]. This pathophysiological hypothesis is consistent with the divergent distribution of pelvic morphology observed between the AIS and non-scoliotic cohorts. The aberrant biomechanical mechanisms implicated in scoliosis pathogenesis may similarly influence pelvic development in affected individuals, potentially contributing to the morphological variations identified in this study.

No significant difference (P = 0.554) was found in the distribution of pelvic morphology types across different segmental scoliosis. It is important to note that the Caldwell-Moloy classification system primarily based on the coronal plane morphology and proportional dimensions of the pelvic inlet and does not incorporate sagittal alignment parameters. Consequently, this classification may inadequately characterize three-dimensional pelvic morphology, particularly variations in the sagittal plane. Furthermore, it is plausible that pelvic morphology is governed by systemic regulatory mechanisms rather than being solely determined by local musculoskeletal adaptations. This latter hypothesis, however, necessitates further investigation to elucidate the relative contributions of systemic versus local factors.

Currently, there is no universal consensus on the causes of idiopathic scoliosis. Burwell et al. proposed a general theory of AIS etiology where a pelvic rotation-inducing system transfers this rotation to the spine. altered transverse pelvic morphology could modify the loads acting on the spine and influence the progression of AIS as suggested by Mac-Thiong et al. [48]. So, this study assumed that there should be a certain degree of correlation between the axial rotation and tilt of the pelvis and the parameters of scoliosis. After statistical analysis, we found that there is only a significant correlation between Cobb angle and AR(P < 0.001). Comparison of parameters among different segmental scoliosis and different Cobb degrees in pelvic parameters also found no significant differences in pelvic parameters. This is roughly consistent with the conclusions of previous studies. Stylianides et al. found no statistical difference was observed between AIS patients and non-scoliotic teenagers for all 12 angles describing pelvic rotation and tilts. Gum et al. identified no significant correlations between the major Cobb angle and left/right pelvic ratios within any scoliosis subgroup. There was only a suggestive correlation of 0.3259 (P = 0.0736) between the largest Cobb angle and the left/right pelvic width ratios for the AIS Lenke 1A1 thoracic sub-group [55]. Wang et al. [56] confirmed that adolescents with AIS, regardless of whether the major curve was thoracic or left thoracolumbar/lumbar, consistently demonstrated compensatory rightward axial rotation of the pelvis (i.e., toward the direction of the thoracic curve). This compensatory rotation was more prevalent in patients with major thoracic curves than in those with major lumbar curves. The authors further reported that the direction and magnitude of apical vertebral rotation significantly influenced these compensatory pelvic rotations. However, the Cobb angles consisted exclusively of AIS patients with severe curvature (Cobb angles 40–70°); consequently, their findings may not be generalizable to individuals with mild or moderate AIS. Based on this limitation and the results of this study, we propose that pelvic rotation is unlikely to be a primary etiological factor in scoliosis onset. Instead, it may represent a secondary compensatory mechanism adopted by the pelvis to maintain overall truncal stability in response to the spinal deformity.

Conclusion

In summary, this study identified a statistically significant difference in pelvic morphology between adolescents with AIS and non-scoliotic controls. Specifically, the AIS group exhibited a significantly higher prevalence of the Gynecoid type and a significantly lower prevalence of the Android type compared to the non-scoliotic group. These morphological variations may be attributable to several associated factors prevalent in AIS, including endocrine abnormalities, diminished physical activity, and altered body composition. However, the precise causal relationships and mechanistic pathways underlying these associations require further empirical validation. For moderate to severe cases, three-dimensional imaging is recommended to obtain a more accurate assessment of AR, as evaluation relying solely on two-dimensional radiographs may be insufficient. Changes in AR measurement may serve as a potentially indicator for assessing the risk of curve progression, thereby warranting more vigilant monitoring and potentially justifying early intervention strategies.

Additionally, observed PR and CPT may be associated with compensatory mechanisms secondary to lower limb length discrepancy or primary hip joint pathology rather than being directly attributable to scoliosis. Consequently, the detection of significant PR or CPT in AIS patients should prompt clinical evaluation to rule out these alternative underlying etiologies. Notably, the severe AIS group exhibited a restricted morphological profile, consisting exclusively of Gynecoid or Platypelloid pelvic types, and demonstrated significant differences in AR. Preoperative evaluation of pelvic morphology in surgical candidates (particularly female patients) may provide valuable insights into anticipating postoperative compensatory mechanisms and spinopelvic balance. For patients presenting with lumbar scoliosis and a Gynecoid pelvic morphology, rehabilitation strategies emphasizing lumbopelvic stabilization training may represent a beneficial therapeutic approach to mitigate compensatory curve progression. However, the efficacy of this specific intervention requires validation through future prospective, controlled interventional studies.

Collectively, these findings highlight descriptive associations between specific pelvic morphology patterns and AIS. However, their functional biomechanical implications and clinical significance of these associations warrant further investigation through longitudinal observational studies and mechanistic research.

Abbreviations

AIS

Adolescent idiopathic scoliosis

AR

Apex-vertebral rotation: the degree of rotation of the scoliosis apex vertebra in the axial plane measured by the Raimondi method

PR

Pelvic rotation: the absolute value of the difference between the left-right ratio and 1 (|1-L/R|) of horizontal distance between the anterior superior iliac spine and the inferior ilium at the sacroiliac joint on the same side.

CPT

Coronal pelvic tilt: the angle between the line connecting bilateral eyebrow arch of acetabulum and horizontal line

BMI

Body Mass Index

Author contributions

Conceptualization: YXM and NTH; Data curation: SY, PYS; Formal analysis: SY; Investigation: SY, QFP and NTH; Methodology: SY, QFP and NTH; Project administration: YXM; Resources: YXM; Software: not applicable; Supervision: YXM; Validation: YXM; Visualization: SY; Writing -original draft: SY; Writing-review & editing: YXM.

Funding

This work was supported by the special program of scientific research for hospitals associated with Zhejiang Chinese Medical University in 2023(Grant No. 2023FSYYZY16) and Key Discipline Project of High level TCM of National Administration of Traditional Chinese Medicine (GJXK2023-85). The funding sources were not involved in the design, analysis, or writing process.

Data availability

The datasets analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Approval of the study was obtained from the ethics committee of our hospital. The research has been carried out in accordance with the World Medical Association Declaration of Helsinki. All methods were performed in accordance with the relevant guidelines and regulations and used for patients’ diagnosis and therapeutic assessment, as part of routine treatment. As the data obtained in this study were anonymous and the study was retrospective, the ethics committee waived the need to obtain informed consent.

Consent for publication

As the data obtained in this study were anonymous and the study was retrospective, the ethics committee waived the need to obtain consent for publication.

Competing interests

The authors declare no competing interests.