Image-based precision diagnostics for cancer and infectious diseases
One of our major research areas is image-based diagnostics, with special focus on tissue diagnostics. We are exploring AI-assisted methods for automated analysis of digitized cancer tissue and microbiological samples. Methods for computerized profiling of digitized tumor samples are applied in analysis of comprehensive series of breast, prostate and colorectal cancers together with detailed clinical data. The goal is to achieve more precise, efficient and safe diagnostic profiling of the individual patient and to provide information to be used in decisions regarding treatment.
Personalized prediction of cancer outcome
The group has developed an online risk calculator - "Prognomics" - for indivualized prediction of cancer outcome. The risk calculator is an extension of our previously published case-match method and is connected to a database with molecular, as well as clinical and outcome data on more than 10 million cases from 15 countries. Survival estimates can be retrieved for all major cancers and the extensive database allows for estimation of risk even in rare subgroups. The case-match approach is also used for explorative analysis of novel biomarkers and is linked to our image analysis system described above.
The MoMic Project – Point-of-care digital microscopy at the front line
Our research group has together with Karolinska Institutet, University of Oulu and a microelectronics company (Laser Probe LP Ltd, Oulu) developed a prototype for a low-cost miniaturized microscope that can capture highly magnified images of patient samples and transfer these for diagnostic support (e.g. cancer and parasite diagnostics, biomarker readout) over wireless mobile networks.
Virtual microscopy and cloud-based artificial intelligence allows the diagnostic process to be performed remotely, by a human observer in combination with automated computer vision analysis. To evaluate the feasibility of the method, field studies are carried out in Tanzania and Kenya in collaboration with partner universities. The methods will aid diagnostics at the point-of-care and aims to empower local experts to deliver high-quality and affordable diagnostic services.