Related Experiment Video
Updated: Jan 7, 2026
06:11
Anterior Cervical Discectomy and Fusion in the Ovine Model
Published on: October 5, 2009
13.2K
Development and Validation of an Interpretable Machine Learning Model for Predicting Early Fusion After Anterior
Hong Wang1, Shuang Liang, Kangkang Huang
1Department of Orthopedics, West China Hospital of Sichuan University, Chengdu, Sichuan Province, China.
Spine
|December 12, 2025
Summary
A machine learning model can predict early fusion after anterior cervical discectomy and fusion surgery. Key factors influencing fusion include range of motion, glucose, and bone density, aiding personalized patient care.
Area of Science:
- Spine surgery outcomes
- Machine learning in medicine
- Radiographic fusion assessment
Background:
- Anterior cervical discectomy and fusion (ACDF) is common, with early fusion (EF) crucial for success.
- Predicting EF after ACDF is challenging due to patient variability.
- Current predictive methods for EF are limited.
Purpose of the Study:
- Develop and validate a machine learning (ML) model to predict EF after ACDF.
- Identify key factors influencing EF.
- Enhance clinical decision-making for ACDF patients.
Main Methods:
- Retrospective analysis of 1,039 surgical segments from 840 ACDF patients (2013-2020).
- Utilized nine ML algorithms, with Stochastic Gradient Boosting (SGB) showing highest predictive ability (AUC 0.884 training, 0.830 testing).
- Applied SHapley Additive exPlanations (SHAP) to identify influential factors.
Main Results:
- SGB model demonstrated strong predictive performance for EF.
- Key predictors included preoperative FSU ROM, ΔFSU height, FPG, Ca, LDL-C, surgical type, age, and femoral BMD.
- Factors like higher FSU ROM, FPG, LDL-C, and age decreased EF likelihood; optimal ΔFSU height, Ca, and BMD increased it.
Conclusions:
- An ML-based approach, particularly SGB, shows promise for predicting EF post-ACDF.
- Identified factors offer insights for personalized perioperative management.
- Further multicenter validation is recommended for clinical implementation.
