Pathologist and machine – the perfect partnership?
In this chapter, Dr. Williams explores opportunities and questions around the use of digital pathology (DP), whole slide imaging (WSI), and recent advances in artificial intelligence (AI). Dr. Williams explores how DP and AI can support pathologists – not replace them – and how DP and AI can be transformative for healthcare managers, achieving world-class outcomes for patients and a truly twenty-first century pathology service.
“Archives of high-resolution diagnostic images have the potential to provide real world training datasets that can be used to create AI apps. For example, an app might prescreen lymph node tissue to detect regions suspicious for tumour metastasis. Ideally, such apps fit into the workflow that is enabled by DP, allowing the app to become an important part of the diagnostic process. Some apps of this type have already been developed, executing tumour detection, diagnosis, grading as well as prognostic and predictive tasks.4 But there is much more to come. With the development of deep learning, we are starting to use AI to carry out more complex tasks in histology image recognition and classification. In deep learning, the systems created behave much more like the human brain, with layers of ‘neural’ (computational) networks that can develop their own algorithms without the need for algorithms to be inputted beforehand. Generating algorithms based on the large amount of data currently being gathered by research labs and hospitals also has the capacity to build our knowledge of cancer and cancer sub-types.”
4 Ibrahim A, Gamble P, Jaroensri R et al. Artificial intelligence in digital breast pathology: Techniques and applications. The Breast 2020;49:267–273.