Abstract：
Low bone mineral density and impaired bone quality have been shown to be important prognostic factors for curve progression in adolescent idiopathic scoliosis (AIS). There is no evidence-based integrative interpretation method to analyze high-resolution peripheral quantitative computed tomography (HR-pQCT) data in AIS. This study aimed to (1) utilize unsupervised machine learning to cluster bone microarchitecture phenotypes on HR-pQCT parameters in girls with AIS, (2) assess the phenotypes\' risk of curve progression and progression to surgical threshold at skeletal maturity (primary cohort), and (3) investigate risk of curve progression in a separate cohort of girls with mild AIS whose curve severity did not reach bracing threshold at recruitment (secondary cohort). Patients were followed up prospectively for 6.22 ± 0.33 years in the primary cohort (n = 101). Three bone microarchitecture phenotypes were clustered by fuzzy C-means at time of peripubertal peak height velocity (PHV). Phenotype 1 had normal bone characteristics. Phenotype 2 was characterized by low bone volume and high cortical bone density, and phenotype 3 had low cortical and trabecular bone density and impaired trabecular microarchitecture. The difference in bone quality among the phenotypes was significant at peripubertal PHV and continued to skeletal maturity. Phenotype 3 had significantly increased risk of curve progression to surgical threshold at skeletal maturity (odd ratio [OR] = 4.88; 95% CI, 1.03-28.63). In the secondary cohort (n = 106), both phenotype 2 (adjusted OR = 5.39; 95% CI, 1.47-22.76) and phenotype 3 (adjusted OR = 3.67; 95% CI, 1.05-14.29) had increased risk of curve progression ≥6° with mean follow-up of 3.03 ± 0.16 years. In conclusion, 3 distinct bone microarchitecture phenotypes could be clustered by unsupervised machine learning on HR-pQCT-generated bone parameters at peripubertal PHV in AIS. The bone quality reflected by these phenotypes was found to have significant differentiating risk of curve progression and progression to surgical threshold at skeletal maturity in AIS.
Adolescent idiopathic scoliosis (AIS) is an abnormal spinal curvature that commonly presents during puberty growth. Evidence has shown that low bone mineral density and impaired bone quality are important risk factors for curve progression in AIS. High-resolution peripheral quantitative computed tomography (HR-pQCT) has improved our understanding of bone quality in AIS. It generates a large amount of quantitative and qualitative bone parameters from a single measurement, but the data are not easy for clinicians to interpret and analyze. This study enrolled girls with AIS and used an unsupervised machine-learning model to analyze their HR-pQCT data at the first clinic visit. The model clustered the patients into 3 bone microarchitecture phenotypes (ie, phenotype 1: normal; phenotype 2: low bone volume and high cortical bone density; and phenotype 3: low cortical and trabecular bone density and impaired trabecular microarchitecture). They were longitudinally followed up for 6 years until skeletal maturity. We observed the 3 phenotypes were persistent and phenotype 3 had a significantly increased risk of curve progression to severity that requires invasive spinal surgery (odds ratio = 4.88, p = .029). The difference in bone quality reflected by these 3 distinct phenotypes could aid clinicians to differentiate risk of curve progression and surgery at early stages of AIS.
Adolescent idiopathic scoliosis (AIS) is an abnormal spinal curvature that commonly presents during puberty growth. Evidence has shown that low bone mineral density and impaired bone quality are important risk factors for curve progression in AIS. High-resolution peripheral quantitative computed tomography (HR-pQCT) has improved our understanding of bone quality in AIS. It generates a large amount of quantitative and qualitative bone parameters from a single measurement, but the data are not easy for clinicians to interpret and analyze. This study enrolled girls with AIS and used an unsupervised machine-learning model to analyze their HR-pQCT data at the first clinic visit. The model clustered the patients into 3 bone microarchitecture phenotypes (ie, phenotype 1: normal; phenotype 2: low bone volume and high cortical bone density; and phenotype 3: low cortical and trabecular bone density and impaired trabecular microarchitecture). They were longitudinally followed up for 6 years until skeletal maturity. We observed the 3 phenotypes were persistent and phenotype 3 had a significantly increased risk of curve progression to severity that requires invasive spinal surgery (odds ratio = 4.88, p = .029). The difference in bone quality reflected by these 3 distinct phenotypes could aid clinicians to differentiate risk of curve progression and surgery at early stages of AIS.
摘要：
低骨密度和骨质量受损已被证明是青少年特发性脊柱侧凸（AIS）曲线进展的重要预后因素。没有基于证据的综合解释方法来分析AIS中的高分辨率外周定量计算机断层扫描（HR-pQCT）数据。这项研究旨在（a）利用无监督机器学习对AIS女孩的HR-pQCT参数进行骨骼微结构表型聚类，(b)评估骨骼成熟度时曲线进展和进展至手术阈值的表型风险(主要队列),(c)在招募时曲线严重程度未达到支撑阈值的轻度AIS女孩的单独队列(次要队列)中,调查曲线进展的风险.在主要队列中，对患者进行了6.22±0.33年的前瞻性随访（N=101）。在青春期峰高速度（PHV）时，通过模糊C均值对三种骨微结构表型进行聚类。表型-1具有正常的骨特征。表型-2的特点是低骨体积和高皮质骨密度，表型3的皮质和小梁骨密度低，小梁微结构受损。在青春期PHV中，表型之间的骨质量差异显着，并持续到骨骼成熟。表型3在骨骼成熟时曲线进展至手术阈值的风险显着增加（奇数比（OR）=4.88；95％置信区间（CI）：1.03-28.63）。在次要队列中(N=106)，表型-2(校正OR=5.39;95CI:1.47~22.76)和表型-3(校正OR=3.67;95CI:1.05~14.29)的曲线进展风险均增加≥6°,平均随访时间为3.03±0.16年.总之,三种不同的骨微结构表型可以通过无监督机器学习对AIS中青春期PHV的HR-pQCT生成的骨参数进行聚类。发现这些表型反映的骨骼质量在AIS中具有明显的曲线进展和进展到骨骼成熟度的手术阈值的风险。
青少年特发性脊柱侧凸（AIS）是青春期生长过程中常见的异常脊柱弯曲。证据表明，低骨密度和骨质量受损是AIS曲线进展的重要危险因素。高分辨率外周定量计算机断层扫描（HR-pQCT）提高了我们对AIS中骨质量的理解。它从一次测量中产生大量的定量和定性骨参数，但是这些数据对于临床医生来说并不容易解释和分析。这项研究招募了AIS女孩，并使用无监督的机器学习模型在首次临床就诊时分析她们的HR-pQCT数据。该模型将患者分为3种骨骼微结构表型（即表型1：正常，表型2：低骨量和高皮质骨密度，和表型3：皮质和小梁骨密度低，小梁微结构受损）。纵向随访6年，直到骨骼成熟。我们观察到这三种表型是持续的，和表型-3有显著增加的风险曲线进展到严重程度,需要侵入性脊柱手术(赔率比=4.88,P=0.029).这3种不同表型反映的骨质量差异可以帮助临床医生在AIS的早期阶段区分曲线进展和手术的风险。
青少年特发性脊柱侧凸（AIS）是青春期生长过程中常见的异常脊柱弯曲。证据表明，低骨密度和骨质量受损是AIS曲线进展的重要危险因素。高分辨率外周定量计算机断层扫描（HR-pQCT）提高了我们对AIS中骨质量的理解。它从一次测量中产生大量的定量和定性骨参数，但是这些数据对于临床医生来说并不容易解释和分析。这项研究招募了AIS女孩，并使用无监督的机器学习模型在首次临床就诊时分析她们的HR-pQCT数据。该模型将患者分为3种骨骼微结构表型（即表型1：正常，表型2：低骨量和高皮质骨密度，和表型3：皮质和小梁骨密度低，小梁微结构受损）。纵向随访6年，直到骨骼成熟。我们观察到这三种表型是持续的，和表型-3有显著增加的风险曲线进展到严重程度,需要侵入性脊柱手术(赔率比=4.88,P=0.029).这3种不同表型反映的骨质量差异可以帮助临床医生在AIS的早期阶段区分曲线进展和手术的风险。
