Related Experiment Video
Updated: Jun 14, 2025
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
Assessment of outcomes and machine Learning-based models to predict local failure risk following stereotactic
Sreenija Yarlagadda1, Yanjia Zhang2, Anshul Saxena2
1Department of Radiation Oncology, Miami Cancer Institute, Baptist Health South Florida, 8900 N Kendall Drive, Miami, FL, 33176, USA.
Introduction:
We assessed the outcomes of stereotactic radiosurgery (SRS) for small intact brain metastases (SBM) (≤ 2 cm) and developed machine learning (ML) algorithms to predict the probability of local failure (LF).
Methods:
Consecutive patients with SBM treated with SRS between January 2017 and July 2022 were included. Propensity score matching (PSM) was performed with related factors to enhance balance for comparison. Variable selection and three time-varied generalized estimating equations (GEE) were used to create predictive models.
Results:
1503 SBMs in 235 patients treated over 358 SRS courses were analyzable. The actuarial 1-year cumulative rate of LF was lower in lesions treated with 24 Gy (5.9%, 95% CI: 4.2-8.2%) or 22 Gy (7.7%, 95% CI: 5.3-11.0%) compared to 20 Gy (25.3%, 95% CI: 18.1-34.7%) (p < 0.001). 22 Gy and 24 Gy were associated with a 63% and 74% reduction in risk in LF compared to 20 Gy (HR: 0.37; 95% CI: 0.24-0.57; p < 0.005 and HR: 0.26; 95% CI: 0.17-0.39; p < 0.005, respectively). The generated models could recommend the best dose with an individualized percentage probability of LF with each dose at 6 months, 1 year, and 2 years with a minimum AUC of 0.75. The 1-year model had the highest AUC (0.88), accuracy (88%), and specificity (91%), while the 2-year model had the highest sensitivity (89%).
Conclusion:
The ML models developed predict LF as a function of dose which could aid in clinical decision-making to select an appropriate dose for SBM to optimize tumor control outcomes and schedule appropriate follow-up.
Insights
Higher doses of stereotactic radiosurgery (SRS) significantly reduce local failure (LF) in small brain metastases (SBM). Machine learning models predict LF probability, aiding personalized treatment decisions for SBM.
Area of Science:
- Neurosurgery
- Radiation Oncology
- Medical Machine Learning
Background:
- Stereotactic radiosurgery (SRS) is a key treatment for small brain metastases (SBM).
- Predicting local failure (LF) is crucial for optimizing SBM treatment outcomes.
- Machine learning (ML) offers potential for personalized prediction of LF.
Purpose of the Study:
- To assess SRS outcomes for SBM (≤2 cm).
- To develop ML algorithms for predicting LF probability after SRS for SBM.
- To guide clinical decision-making for optimal SBM dose selection.
Main Methods:
- Analysis of consecutive SBM patients treated with SRS (Jan 2017 - Jul 2022).
- Propensity score matching (PSM) for balanced comparison of treatment factors.
- Development of predictive models using variable selection and generalized estimating equations (GEE).
Main Results:
- Lower 1-year LF rates observed with 24 Gy (5.9%) and 22 Gy (7.7%) compared to 20 Gy (25.3%).
- Higher SRS doses (22 Gy, 24 Gy) significantly reduced LF risk (63-74%) versus 20 Gy.
- ML models achieved high performance (AUC ≥ 0.75) in predicting LF at 6 months, 1 year, and 2 years.
Conclusions:
- Higher SRS doses are associated with significantly reduced local failure in SBM.
- Developed ML models accurately predict LF risk based on SRS dose.
- These models can assist clinicians in selecting optimal SRS doses for SBM, improving tumor control and follow-up scheduling.
