Comparing Models for Radiation-Induced Eye Damage
Author Information
Author(s): Vitali Moiseenko, William Y. Song, Loren K. Mell, Niranjan Bhandare
Primary Institution: British Columbia Cancer Agency
Hypothesis
To what extent are model predictions for radiation-induced optic neuropathy and retinopathy data-driven or dependent on the choice of the model?
Conclusion
The log-logistic model tends to lead to larger D50 and lower γ compared to other models, indicating that model choice can significantly affect predictions.
Supporting Evidence
- The log-logistic model consistently yields larger D50 and smaller γ compared to logistic and probit models.
- Incidence of optic neuropathy was 5 in 101 patients treated twice-daily and 19 in 172 patients treated once daily.
- For retinopathy, the incidence was 7 in 78 for patients treated twice daily and 23 in 108 for patients treated once daily.
Takeaway
This study looks at how different mathematical models predict eye damage from radiation therapy, showing that the choice of model can change the results a lot.
Methodology
Four dose-response models were tested for their ability to describe data on radiation-induced optic neuropathy and retinopathy.
Potential Biases
Model parameters are sensitive to the choice of model and the range of complication incidence.
Limitations
The study is limited by the small number of events leading to broad confidence intervals for model parameters.
Participant Demographics
Patients treated for head and neck cancers between 1964 and 2000.
Statistical Information
Confidence Interval
95%
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website