Deep Active Survival Analysis approach on Prostate Cancer treatment
Principal Investigator
Name
Satyo Iswara
Degrees
B.S
Institution
University of Illinois At Urbana-Champaign
Position Title
Student
Email
About this CDAS Project
Study
PLCO
(Learn more about this study)
Project ID
PLCO-971
Initial CDAS Request Approval
Apr 25, 2022
Title
Deep Active Survival Analysis approach on Prostate Cancer treatment
Summary
This project is aim to analyze the effectiveness of prostate cancer treatment via deep learning network, specifically using deep active survival analysis method. The data I'm looking for are sequential data of prostate cancer treatment overtime and survival status of patient.
Survival analysis will be done by modeling two functions, survival function and hazard function. Survival function indicate probability of patient survival while hazard function predict the probability of patient mortality given previous event. Both of this function will be approximated using Cox Proportional Hazards model.
Data set will be ran through Survival analysis to provide suitable sample which then pass through deep learning network to further identify which treatment options chemotherapy, radiotherapy and surgery, will minimize patient mortality risk.
Survival analysis will be done by modeling two functions, survival function and hazard function. Survival function indicate probability of patient survival while hazard function predict the probability of patient mortality given previous event. Both of this function will be approximated using Cox Proportional Hazards model.
Data set will be ran through Survival analysis to provide suitable sample which then pass through deep learning network to further identify which treatment options chemotherapy, radiotherapy and surgery, will minimize patient mortality risk.
Aims
- Analyzing the most effective treatment for prostate cancer, especially the difference between chemotherapy, radiotherapy and surgery and related mortality risk in any of each intervention
- Identify the most significant feature that produce the most effective treatment for positive outcome
Collaborators
Jimeng Sun (course professor - jimeng@illinois.edu)