Project

Personalizing Treatment of Prostate Cancer With Image-based Diagnostics Informed by Deep Learning and Physics-based Models

Short Summary

Prostate cancer is a common cancer in men. Intraductal carcinoma of the prostate is a relatively rare subtype of prostate cancer, but it is often found in high-risk disease cases. Even though intraductal carcinoma of the prostate is implicated in high-risk cases, it remains understudied and difficult to identify. Thus, in this project, we aim to use a combination of physics-based models and deep learning to study intraductal carcinoma of the prostate and develop new image-based diagnostic tools. These tools and insights will improve the precision prostate cancer diagnosis and may lead to the identification of new therapeutic targets.

Goals

The goal of this project is to improve the precision of prostate cancer diagnostics and make it easier for clinicians to tailor treatment to a patient’s personal cancer.

Significance

Prostate cancer is the most commonly diagnosed cancer in men in Europe and contributes the 4th most cancer-related deaths (2nd most cancer-related deaths in Switzerland). While extremely prevalent, many prostate cancers cases are slow growing and thus only require close monitoring. The more severe cases require aggressive therapies such as androgen deprivation therapy or radical prostatectomy. Thus, precise diagnostic tools are essential to prevent overtreatment of patients and maximize quality of life.

Background

Intraductal carcinoma of the prostate is a subtype of prostate cancer that makes up a relatively small fraction of total prostate cancer cases, but is often found in high risk or severe cases. Unfortunately, our current understanding of intraductal carcinoma of the prostate is limited due to challenges with identification using conventional methods. Currently, prostate cancer is graded based on images of small sections of the tumor. Recent advances in deep learning have enhanced our ability to detect specific cancer subtypes from images. Further, state of the art physics-based simulations provide insight into how tumors form. In this project, we will merge these two approaches to enhance the precision of prostate cancer diagnostics and study the features that underpin intraductal carcinoma of the prostate.

Publications

Patents / Startups

Publications

Patents / Startups

Transition Postdoc Fellowship Project

Co-Investigators

Prof. Dagmar Iber, ETH Zürich
Prof. Lukas Bubendorf, University Hospital Basel
Dr. Clémentine Le Magnen, University Hospital Basel
Prof. Cyrill Rentsch, University Hospital Basel

Consortium

Status

In Progress

5th Call, TPdF