Project: Resilient predictive models based on quantitative imaging to guide prostate cancer treatment
In progress (2021-today)
- Maxence Larose1,2,3(2021-aujourd’hui)
- Louis Archambault1,2(2021-aujourd’hui)
- Martin Vallières3(2021-aujourd’hui)
1 Physics, physics engineering and optics department, Université Laval, Québec (QC), Canada
2 CHU de Québec, Québec (QC), Canada
3 Computer science department, Université de Sherbrooke, Sherbrooke (QC), Canada
Objective: To develop resilient predictive models based on quantitative imaging and clinical features to guide treatment for advanced prostate cancer using images from different modalities such as computed tomography (CT) and positron emission tomography (PET).
3 specific objectives:
- Generation of synthetic data to impute missing data.
- Automatic segmentation of organs of interest and detection of failed segmentations.
- Robust prediction of clinical outcomes based on stable radiomic features.