Congress:
ECR25
Poster Number:
C-10312
Type:
Poster: EPOS Radiologist (scientific)
Authorblock:
L. J. Jensen, D. Kim, T. Elgeti, I. G. Steffen, L-A. Schaafs, M. Haas, L. J. Kurz, B. Hamm, S. N. Nagel; Berlin/DE
Disclosures:
Laura Jacqueline Jensen:
Nothing to disclose
Damon Kim:
Nothing to disclose
Thomas Elgeti:
Nothing to disclose
Ingo G. Steffen:
Nothing to disclose
Lars-Arne Schaafs:
Nothing to disclose
Matthias Haas:
Nothing to disclose
Lukas J. Kurz:
Nothing to disclose
Bernd Hamm:
Grant Recipient: Charité University Hospital Berlin, Department of Radiology
Sebastian N. Nagel:
Nothing to disclose
Keywords:
Genital / Reproductive system male, MR, Computer Applications-3D, Computer Applications-Detection, diagnosis, Segmentation, Cancer
Ten map-derived features, listed in Figure 5, differed significantly between the csPCa and non-csPCa group (AUCs 0.70-0.84). The diagnostic performance for TZ lesions (AUC 0.83-1.00) was superior to PZ lesions (AUC 0.74-0.85) in the univariate analysis. Parametric maps of distinctive features in the PZ are shown in Figure 6; for the TZ, they are shown in Figure 7. In the bivariate models, performance improved overall (AUCs 0.90-0.94), as shown in Figure 8, particularly for the PZ lesions (AUCs 0.91-0.93), summarized in Figure 9.