Your browser doesn't support javascript. This means that the content or functionality of our website will be limited or unavailable. If you need more information about Vinnova, please contact us.

Deep learning for localization of the prostata in MRI using gold markers for prostate cancer radiotherapy

Reference number
Coordinator Lunds universitet
Funding from Vinnova SEK 44 802
Project duration November 2018 - March 2019
Status Completed
Venture Personal mobility between societal sectors
Call Funding for staff exchange and artificial intelligence (AI)

Purpose and goal

We have used advanced AI methods to train a deep neural network to identify gold markers in MR images and distinguish these from calcifications and other common substances naturally occurring in the body. The gold markers are used to locate the prostate in external radiation therapy and are commonly used together with (Computer Tomography) CT. Since MR images are better at depicting the soft tissues in the body, one wants to find a robust method that is able to identify the gold markers also in MR images.

Expected results and effects

At the completion of this project, we have developed a working method for locating the prostate in MR images that has similar performance compared to locating the prostate using CT images. The method is analyzed and evaluated in a number of different measures. The result of this project is a scientific article and a basis for developing a new product that can be used clinically for prostate localization in MR images.

Planned approach and implementation

We have tested methods that have previously been shown to be successful in solving similar problems. These methods have been analyzed and possible improvements to the methods have been identified which have then been incorporated into the neural networks. An important aspect of the project has been to analyze and validate the method´s performance so that it meets the requirements that are set for this particular task. This has been validated in a series of different tests.

The project description has been provided by the project members themselves and the text has not been looked at by our editors.

Last updated 27 October 2018

Reference number 2018-04353

Page statistics