The integrated solution is meant to increase MR imaging’s status as an alternative to PSA screenings, which are currently the standard test for identifying prostate cancer early but lack specificity. As a result, patients often undergo biopsies and experience anxiety while waiting long periods for results. The algorithm will operate seamlessly in the background, immediately processing acquired images upon their arrival in the PACS system.

"The ongoing development of clinical research studies seeks to establish links between the quantitative features extracted from baseline prostate MR and critical clinical endpoints, such as biochemical recurrence. These endeavors hold the promise of yielding future predictive biomarkers that could more effectively stratify patients at risk and identify candidates for more intensive treatment regimens," said Herigault and Alberich-Bayarri.


The two caution that this potential is still a considerable distance from being a practical clinical application.

The solution will eventually be used to address other forms of cancer. The deep-learning-based automated gland segmentation and biomarker extraction processes have received 510(k), CE Mark, and UKCA regulatory clearances, but the integration of these solutions into an AI-based lesion detection algorithm is currently only intended for research purposes. This is because achieving higher specificity is also dependent on how AI models are trained and the quality of the data they process, say Herigault and Alberich-Bayarri.

Philips and Quibim will be showcasing their solutions at the Radiological Society of North America Annual Meeting between November 26 and 29 in Chicago.

Back to HCB News

MRI Homepage