Next generation image analysis solutions - towards image-based diagnostics


Novel and accurate analysis techniques, which are well beyond the state-of-the-art, are required to harvest all available information from cells and tissues. These are of great importance for patient samples utilized in the translational research, both in academy and pharma. In addition, the automation of image analysis for pathology, cytology and hematology is a significant growth area for future clinical diagnostics.

In this project we combined the expertise of the FiDiPro Fellow Dr. Peter Horvath (ETH Zürich, Switzerland) in biomedical image analysis with the on-going translational biomedical research at the University of Helsinki (UH) as well as with a group of companies/actors from the fields of health and informatics. The involved partners represented the whole innovation chain in the health sector, from academic research groups and core facilities affiliated with the Institute for Molecular Medicine Finland (FIMM, UH) and the Biomedicum Imaging Unit (BIU, UH) to companies on biomedical image analysis, therapeutics development, diagnostics and clinical laboratory services. On the informatics side, the innovation chain was covered by partners involved e.g. in machine learning algorithm development and testing.

The results of this project have been novel open-source image analysis and machine learning tools applicable for example, for cancer tissue biomarker detection. First, the project focused on questions driven by translational biomedical research on patient samples. Solutions were customized and service-based. Secondly, the solutions were geared towards intelligent automated pipelines that can benefit industry, including Pharma in high content image-based drug screening and the clinical laboratories in the future. The project has created a significant Finnish knowledge cluster within currently emerging field of image-based clinical diagnostics.

As a result…


  1. We developed algorithms that analyze microscopy images of drug treated patient derived cells. We correct image intensities (Smith et al. 2015, Nature Methods), detect cells in very complex environments (Molnar et al., 2016, SciRep), and artificial intelligence-based methods to identify the types of cells in big data environment (Piccinini etal. Cell Systems 2017). Results are already utilized in the treatment of different diseases such as prostate cancer (Saaed etal, European Urology 2017, 2019 ), amd leukemia (Fristmantas et al. Blood 2017) 
  2. To most precisely understand the differences and role of single cells, we developed a machine learning method that can intelligently select and isolate cells for DNA sequencing from most complex environments (Brasko etal 2018, Nature Communications).


You can access the open software and tools developed: 




If you wish to read more, the top 5 publications of the project are:

Intelligent image-based in situ single-cell isolation. 

Brasko C., Smith K., Molnar, Cs., Farago N., Hegedus L., Balind A., Balassa, T., Szkalisity, A., Sukosd F., Kocsis K., Balint B., Paavolainen, L., Enyedi M.Z., Nagy I., Puskas L.G., Haracska L., Tamas, G., Horvath, P. (2018). Nature Communications.

Data-analysis strategies for image-based cell profiling.

Caicedo, J. C., ..., Molnar, Cs., ..., Horvath, P., Linington, R. G., Carpenter, A. E. (2017). Nature Methods.

ACC: Discovery software for phenotypic image analysis.

Piccinini, F., Balassa, T., Szkalisity, A., Molnar, Cs., Paavolainen, L., Kujala, K., Buzas, K., Sarazova, M., Pietiainen, V., Kutay, U., Smith K., Horvath, P. (2017). Cell Systems.

Role for formin-like 1-dependent acto-myosin assembly in lipid droplet dynamics and lipid storage.

Pfisterer, S., Gateva, G., Horvath, P., Pirhonen, J., Salo, V., Karhinen, L., Varjosalo, M., Ryhänen, S., Lappalainen, P., Ikonen, E. (2017). Nature Communications.

Screening out irrelevant cell-based models of disease.

Horvath, P., Aulner, N., Bickle, M., Davies, A., Del Nery, E., Ebner, D., Montoya, M., Ostling, P., Pietiainen, V., Price, L., Shorte, S., Turcatti, G., von Schantz, C., Carragher, N. (2016). Nature Reviews Drug Discovery.


Our partners in the health sector represent the whole innovation chain:

Academic research groups:

Core labs:

Clinical laboratory:

Company partners:


Steering group:

Peter Horvath (FIMM-UH), Juhani Luotola (chairman, originally representing SalweLtd), Vilja Pietiäinen (project manager, FIMM, UH), Heidi Arling (coordinator,  FIMM, UH), Johan Lundin, (FiDiPro fellow host, FIMM, UH), Pekka Ruusuvuori, QUVA Oy, Ari Ristimäki (HUSLAB), Elina Ikonen (BIU, UH), Caroline Heckman (FIMM, UH)


Absent: Krister Wennerberg (HTB-FIMM, UH), Lukasz Kuryk (Targovax), Jenni Mäki-Jouppila (Pharmatest Services Ltd).

Harri Ojansuu was representing TEKES/BusinessFinland until 10/2018, and Raimo Pakkanen (10-12/2018).

Last updated: 17.05.2019 - 09:11