RADar – Learning & Innovation Center

ProCaTS

[Prostate Cancer Treatment Support]

ProCaTS aims to demonstrate the feasibility of a data/A.I.-driven decision support tool for personalized and optimal treatment choice in clinical practice in prostate cancer. The project envisions developing an explainable A.I. and combining it with a dashboard tailored to clinicians’ needs.

Objective

The aim of this project is to support physicians in treatment decision for prostate cancer patients using an AI algorithm. First, patient data is extracted from different data systems within the hospital and put into the OMOP common data model. To improve patient care and recurrence detection, Patient Reported Outcome Measures (PROMS) are also added to the dataset. Using all these data, an AI algorithm is designed to support the physician in treatment decision. Within this step, focus relies on explainable AI such that algorithm predictions are substantiated. As a final part, the machine learning model results are incorporated into a graphical user interface which will be used by the physicians during treatment decision at multidisciplinary oncological consults (MOC).

Methodology

  1. Unlocking data and put into OMOP CDM
  2. Training explainable ML algorithms to predict patient outcome
  3. Develop a decision support tool to inform physicians about predicted patient outcome

Impact and future directions

Acceptance by the physicians of the graphic user interface that incorporating a validated AI model. Aiming of > 50% of patients to be consulted this way. For the future, anticipating the further development of the software tool, within the clinical pathway, and planning to have this solution medically approved so that it can be rolled out to other hospitals.
(benefits for patients and future plans or next steps for the project)

General info & contact

Keywords (#): Prostate cancer, treatment decision support, AI-driven, software tool

RADar project research lead: Dr. Nathalie Mertens

RADar project researchers: Prof. Dr. Peter De Jaeger, Pieter-Jan Lammertyn, Dr. Siel Depestele

Principal investigator: MD Wim Van Haute

Timeline: Jan ‘22- Mar ‘24

Status: Ended, next steps and validation will be performed in the IBIS-pro project.

Publications / presentations:

Abstract oral presentation at RSNA. Title: To biopsy or not to biopsy? Feasibility of predicting Gleason score as a pre-biopsy gatekeeper in prostate cancer patients using multimodal patient features. Presented 29 nov 2023.

Partners: Siemens Healthineers + ML6 (Skyhaus)

Funding: Vlaio (HBC.2021.0964)