Novel imaging technologies and artificial intelligence in digital pathology for improved cancer diagnostics

Our areas of research

Artificial intelligence
The high and constantly increasing cancer incidence rates cause significant burden to the pathology laboratories and diagnostic delays for the patients. Digitalization enables the use of artificial intelligence (AI) solutions in digital pathology. Deep learning enables the implementation of computational image analysis, which provides a potential to increase diagnostic precision, reduce diagnosis time and decrease the interpathologist variability. In this project several thousands of unidentified digitalized pathology images of primary tumors and metastases are collected for development of the AI algorithms. We aim to develop tools for diagnostics of primary tumors, detection of metastases and assessment of molecular and prognostic markers with our main focus on skin cancer diagnostics.

Ex-vivo fluorescence confocal microscopy
A novel promising technology for the pathology diagnostics field is ex vivo fluorescence confocal microscopy (FCM) where fresh unfixed tissue specimens are optically scanned with a laser producing images mimicking conventional histopathology images in a few minutes enabling “bedside” diagnostics. FCM makes it possible to produce images of fresh tissue in a few minutes without interfering in any way with the subsequent routine histopathological investigations. Several hundreds of fresh tissue samples from different types of cancer tissue will be collected for FCM analyses. The FCM images will be compared with conventional histopathology. The fast pre- and intraoperative histopathological evaluation of cancers with RCM may offer speed and accuracy in the clinical setting, which would decrease operating costs and waiting time for the patient. Unnecessary revisits could also be avoided. Knowing the extremely high incidence rates of cancers, this could remarkably reduce health care costs.

ToF-SIMS mass spectrometry
In recent years, changes in the lipid metabolism of cancer cells have gotten increased attention. Lipid composition changes can occur much faster than changes in proteins and may be the first indicator of phenotypic alterations in cells. Thus, lipidomics is one of the main current focuses of cancer research. One way of examining and characterising lipids in cancer tissues is secondary ion mass time-of flight spectrometry (ToF-SIMS). ToF-SIMS offers a unique imaging technique with high-resolution chemical mapping of histopathological specimens. This enables side-by-side comparison with routinely stained histopathological slides, and thus detection of alterations at a cellular level. In this manner, chemical changes in healthy vs. cancerous tissue as well as differences between different cancer tissues can be studied. The increasing knowledge of lipidomics could be of great importance in improving earlier and more accurate biomarkers for cancers and serve as a target for novel therapeutic approaches.

Current collaborators:
Prof John Fletcher, Department of Chemistry and Molecular Biology, University of Gothenburg

Prof John Paoli, Department of Dermatology, Institute of Clinical sciences, Sahlgrenska Academy

Prof Roger Olofsson Bagge, Department of Surgery, Institute of Clinical sciences, Sahlgrenska Academy

Associate professor Kari Nielsen, Department of Clinical Sciences, Lund University

Prof Ilkka Pölönen, Department of Information Technology, University of Jyväskylä 

Associate professor Ida Häggström, Chalmers University of Technology and Gothenburg University

PhD Gabriele Campanella, Senior data scientist, Memorial Sloan Kettering Cancer Center, NYC