Abstract #2767

Validation of a diagnostic transition zone prostate cancer model trained in a 1.5T MRI scanner on an independent cohort of patients scanned on a 3T MRI scaner.

Nikolaos Dikaios¹, Ed William Johnston¹, Hashim Ahmed², Lucy, Simmons³, Alex Freeman⁴, Clare Allen⁵, David Atkinson¹, and Shonit Punwani¹

¹Centre of Medical Imaging, UCL, London, United Kingdom, ²Department:Research Department of Urology, UCL, London, United Kingdom, ³Div of Surgery & Interventional Sci, UCL, London, United Kingdom, ⁴Histopathology, UCL, London, United Kingdom, ⁵Radiology, UCL, London, United Kingdom

Diagnostic models for classifying prostate cancer within the transition zone based on multi-parametric magnetic resonance imaging (mp-MRI) have been proposed to improve radiologist’s performance. Such diagnostic models are trained on mp-MRI data acquired at 1.5T or 3T MRI scanners, but to the best of our knowledge no-one has yet examined whether these magnet specific models are interchangeable. This work applied a previously published diagnostic model trained on mp-MRI data acquired at a 1.5T scanner, on an independent cohort of patients with mp-MRI data acquired at a 3T and found that the performance of the model doesn’t drop significantly.

This abstract and the presentation materials are available to members only; a login is required.

Join Here