|
Activity Number:
|
414
- Risk Modeling and Regression Techniques
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 12, 2021 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
SSC (Statistical Society of Canada)
|
|
Abstract #318562
|
|
|
Title:
|
Group-Based Trajectory Models for Change Points: Application to Radiotherapy Schedule for Breast Cancer
|
|
Author(s):
|
Depeng Jiang* and Chendong Li and Jianwei Gou and Yemao Xia
|
|
Companies:
|
University of Manitoba and Nanjing Forestry University and Nanjing Forestry University and Nanjing Forestry University
|
|
Keywords:
|
Change points;
semi-parametric group-based trajectory;
piece-wise mixed effect models;
exhaustive listing methods;
Radiotherapy schedule
|
|
Abstract:
|
Determining the optimal schedule for radiation therapy treatments (RT) for cancer patients is complex and challenging. In RT, a patient is prescribed a series of daily fractions. In this paper, we propose to use the piece-wise mixture model to examine the patterns of fraction durations over a course of treatment. We first use the semi-parametric group-based approach to identify the distinctive groups with different patterns of fraction duration. Then we conduct the piece-wise mixed-effect models to identify the group-specific distinct phases of durations. The change points between the distinct phases are estimated using the exhaustive listing methods in terms of the goodness of fit and quality of prediction. Finally, we examine the patient characterizes associated with each pattern of fraction durations. With application to RT scheduling of breast cancer, our methods will greatly improve the efficiency of making treatment plans.
|
Authors who are presenting talks have a * after their name.
