Digital pathology in a nutshell
Diagnostic pathology is of crucial importance for health care, especially cancer care. Pathologists analyze tissue, histology, and cell samples, cytology, from the patient. This knowledge is used to find the correct diagnosis and therapy. Digitization of the pathology images, whole-slide imaging, provides a possibility to increase both efficiency and quality of the diagnostic work. In order to fulfill this potential, cross-disciplinary research efforts are needed that can combine clinical expertise with knowledge in artificial intelligence (AI) and human-computer interaction.
Ground Truth for AI Models
The CMIV research agenda in digital pathology tackles several image analysis challenges. An important groundwork is to create systematically annotated collections of imaging data to be used as ground truth for training of AI models. CMIV has developed several large data collections, large also with international standards, that are shared for research purposes through the AIDA Data hub.
CMIV is also a leading partner of the EU project BIGPICTURE, creating an international data sharing platform with millions of slides.
Clinical applications of AI
A common ground for the CMIV digital pathology research projects is the focus on making AI-powered solutions work in the clinical setting. A key aspect is how to combine less-than-perfect AI with interactions from human users, at an effort level feasible for high-throughput clinical settings.
Our research agenda includes increasing diagnostic precision and efficiency, for several cancer types. We work within interactive machine learning to accelerate training and finetune during application. We also develop methodology and tools to validate and monitor AI solution performance in clinical production.
The CMIV pathology group consists both of medical and technical researchers from the university, pathologists and lab assistants from the clinical pathology department at the hospital, and researchers from industry.
The group is steadily expanding. If you think working in our group would interest you, please reach out to Adjunct Professor Claes Lundström (see below for contact).