Abstract #4949

Diagnostic Performance in MR-visible Prostate Cancer: Can a Quantitative Computer-aided Diagnosis System Be Superior to the Qualitative PI-RADS v2 Guideline?

Jing Wang¹, Yang Fan², and Yudong Zhang³

¹Center for Medical Device Evaluation, CFDA, Beijing, People's Republic of China, ²MR Research China, GE Healthcare, Beijing, People's Republic of China, ³Department of Radiology, the First Affiliated Hospital with Nanjing Medical University, Nanjing, People's Republic of China

A novel CAD system was developed for prostate cancer detection based on multi-parametric MRI, including textured T2w, DKI and Tofts-Ktrans. MR features were evaluated by using machine-assisted classification methods such as PCA and SVM analysis. The validation performed in 54 patients confirmed as PCa, to determine whether the CAD has the ability to correct diagnosis in MR-visible prostate cancer, as comparison with a proposed structured PI-RADS v2. Our results showed that the automatic PCa detection using CAD had significantly higher AUC than PI-RADS v2 in distinguishing cancer from normal prostate tissue.

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

Join Here