Precision medicine in digital pathology via image analysis and machine learning

In this chapter, we present the concept that digital pathology and artificial intelligence can add value and speed to a pathologist's diagnosis while striving toward precision medicine. We describe how image analysis and machine learning can segment images and compute object- and spatial-based data prior to analysis. This data analysis can identify patients who are at a high risk of succumbing to a disease, who may need more detailed clinical follow-up, or who will respond to specific therapy. We also describe how deep learning algorithms can learn complex morphological patterns from both human- and data-led input in order to perform diagnostic or prognostic tasks. Finally, we discuss the theory behind some commonly used machine learning algorithms and how they may attain regulatory approval.