Related Experiment Video
Updated: Aug 23, 2025
07:15
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
6.9K
Predicting Bone Health Using Machine Learning in Patients undergoing Spinal Reconstruction Surgery.
Yong Shen1, Zeeshan M Sardar1, Herbert Chase2
1Department of Orthopedic Surgery, Columbia University Medical Center, The Spine Hospital at New York Presbyterian.
Spine
|October 27, 2022
Summary
Machine learning accurately predicts bone health status in adult spinal reconstructive surgery patients. This approach identifies key risk factors, improving preoperative patient assessment and care for better surgical outcomes.
Area of Science:
- Orthopedics
- Data Science
- Medical Informatics
Background:
- Bone health significantly impacts outcomes in adult spinal reconstructive (ASR) surgery.
- Surgeons currently lack tools for preoperative risk stratification of bone health in ASR patients.
- Machine learning (ML) can analyze complex data to identify patients at risk for poor bone health.
Purpose of the Study:
- To develop a predictive machine learning model for preoperative bone health status in adult patients undergoing ASR surgery.
- To classify bone health into healthy, osteopenia, and osteoporosis categories.
- To identify patient features associated with poor bone health.
Main Methods:
- Retrospective review of 211 adult patients undergoing ASR surgery with dual-energy X-ray absorptiometry (DXA) scans.
- Data collected from electronic health records.
- Weka software used to build and evaluate ML models (random forest) for bone status classification based on WHO criteria.
Main Results:
- Prevalence of osteoporosis (OPO) was 23.22%, and osteopenia was 52.61%.
- The random forest model showed high performance on the training set (AUC 0.96) and moderate performance on the test set (AUC 0.69).
- Predictive features included BMI, insurance type, serum sodium, serum creatinine, bariatric surgery history, and SSRI use.
Conclusions:
- Machine learning can effectively predict bone health status in ASR patients.
- ML-driven data mining can uncover novel risk factors for bone health in this population.
- This predictive capability can aid in preoperative risk stratification and patient management.
