Predictive visual analytics for prostate cancer therapy

Master Thesis
1

Description

Prostate cancer radiotherapy requires meticulous planning prior to treatment, where the treatment plan is optimized with organ delineations on a pre-treatment scan of the patient. Records from retrospective imaging cohorts with complete data can be employed in a visual analytics environment, to predict expected changes of new incoming patients with incomplete data, using generative models. In our previous work, we have only investigated the incorporation of in the predictions. We would like to further investigate the incorporation of additional non-imaging information of the patients.

Tasks

Design appropriate visual analytics strategies for the incorporation of non-imaging cohort information into an existing predictive visual analytics framework. This should support the exploration and prediction of patient outcome during the upcoming treatment, and the assessment of treatment strategies, with respect to the anticipated changes.

Requirements

  • Interest and knowledge in biomedical visualization and visual analytics.
  • Interest and knowledge in machine learning.
  • Some experience with machine learning frameworks.
  • Good programming skills.
  • Creativity and enthusiasm.

Environment

To be discussed (depending on the background of the student). 

References

[1] https://renataraidou.com/vapor-visual-analytics-for-the-exploration-of-pelvic-organ-variability-in-radiotherapy/
[2] https://renataraidou.com/lessons-learnt-from-developing-visual-analytics-applications-for-adaptive-prostate-cancer-radiotherapy/
[3] https://renataraidou.com/bladder-runner-visual-analytics-for-the-exploration-of-rt-induced-bladder-toxicity-in-a-cohort-study/

Responsible

For more information please contact Renata Raidou.