Portal de Periódicos da CAPES

Improved PI-RADS v2 prediction from simultaneous PET/MRI CNNs

2022; Schattauer Verlag; Linguagem: Inglês

10.1055/s-0042-1746057

2567-6407

Esteban Solari, S. Schahoff, Isabel Rauscher, Wolfgang Weber, Nassir Navab, M. Eiber, Stephan G. Nekolla,

Radiomics and Machine Learning in Medical Imaging

Ziel/Aim Prostate-specific membrane antigen (PSMA)-targeted positron emission tomography (PET) and magnetic resonance imaging (MRI) are the main imaging techniques for prostate cancer (PCa) diagnosis. Previously, we showed that radiomics have the potential to generate an image-based PCa staging by combining information from both PSMA PET and MRI radiomics to predict Gleason scores. In this work, we investigate the power of deep neural networks to predict the PI-RADS v2 score, an MRI-based diagnostic score for PCa.