Related Experiment Video
Updated: May 23, 2026
Modified Posterior Vertebral Column Resection for Patients with Thoracolumbar Kyphotic Deformity
Published on: September 16, 2022
Development and Validation of a Predictive Model Incorporating Paravertebral Muscle Quality for Progressive Kyphosis
Xiaolong Li1, Lei Liu, Zhiyang Xie
1Department of Spine Center, Zhongda Hospital, Medical School, Southeast University, China.
Study Design:
Retrospective cohort study.
Objective:
To investigate risk factors for progressive kyphosis (PK) following percutaneous kyphoplasty (PKP) and develop a validated nomogram for individualized risk prediction.
Summary Of Background Data:
PK after PKP significantly impacts patient outcomes, yet its pathogenesis remains incompletely understood. While paravertebral muscle (PVM) degeneration has been implicated in spinal pathology, its independent contribution to PK after PKP has not been systematically quantified, and no clinical prediction model incorporating muscle quality exists.
Methods:
This study enrolled 330 elderly patients (aged ≥60 y) who underwent single-level PKP for acute or subacute osteoporotic vertebral compression fractures (2013-2022), with a minimum follow-up of 24 months (median 39 mo). PK was defined as an increase in local Cobb angle >10° from immediate postoperative to final follow-up. Patients were randomly divided into training (n=231, 70%) and validation (n=99, 30%) sets. LASSO regression was used for variable selection, followed by multivariable logistic regression to identify independent risk factors and construct a nomogram. PVM fat infiltration (FI) was quantified on axial T2-weighted MRI at the L4/5 level using ImageJ. Model performance was assessed by AUC, calibration plots, and decision curve analysis.
Results:
Four independent predictors were identified: age (OR=1.107, P=0.012), preoperative paravertebral muscle fat infiltration (OR=1.116, P<0.001), preoperative Cobb angle (OR=1.227, P=0.001), and black line signal on MRI (OR=3.251, P=0.015). The nomogram showed excellent discrimination in training (AUC=0.885) and validation (AUC=0.881) sets, with good calibration and net benefit. An online dynamic nomogram was developed for clinical use (https://dynamicnomogramlee.shinyapps.io/DynNomApp/).
Conclusion:
The nomogram incorporating age, paravertebral muscle fat infiltration, preoperative Cobb angle, and black line signal provides accurate, individualized prediction of progressive kyphosis after kyphoplasty, enabling early identification of high-risk patients for targeted preventive strategies.