Estimating Survival in Patients with Operable Skeletal Metastases
Author Information
Author(s): Forsberg Jonathan Agner, Eberhardt John, Boland Patrick J., Wedin Rikard, Healey John H.
Primary Institution: Memorial Sloan-Kettering Cancer Center
Hypothesis
Can Bayesian classifiers be developed to estimate survival in patients undergoing surgery for skeletal metastases?
Conclusion
A robust Bayesian model was developed to accurately estimate survival in patients with operable skeletal metastases.
Supporting Evidence
- The Bayesian model showed an area under the ROC curve of 0.85 for 3-month survival.
- The model's predictions were based on 15 candidate features related to survival.
- The study included a diverse patient population with various types of skeletal metastases.
Takeaway
Doctors can use a special computer model to better guess how long patients with bone cancer might live after surgery.
Methodology
A Bayesian belief network model was developed using historical data from patients who underwent surgery for skeletal metastases, with ten-fold cross-validation to assess accuracy.
Potential Biases
Potential bias from the subjective nature of the surgeon's estimate of survival.
Limitations
The study is a feasibility study with no external validation or comparisons to existing models.
Participant Demographics
Median age was 62 years, with 55.1% women and 85.2% white, non-Hispanic.
Statistical Information
P-Value
p<0.0001
Confidence Interval
95% CI: 0.80–0.93 for 3-month survival; 95% CI: 0.77–0.90 for 12-month survival
Statistical Significance
p<0.0001
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website