Evolutie van autonome diagnostische workflows in kankerbeeldvorming

Over the past decade, image analysis has changed fundamentally. In radiology and, increasingly, in histopathology, we have moved from hand-crafted features and simple pattern recognition to deep models that extract clinically meaningful information directly from pixels. At the same time, imaging has become even more central to oncology: almost every important diagnostic and therapeutic decision in cancer care involves a computed tomography (CT), magnetic resonance imaging (MRI), positron emission tomography (PET), mammogram, or histopathology slide.