Hematoscope Lab

Thank you - a wrap up of 2022

Oscar Brück December 12, 2022
research fundingresearchbig datateam members

The year 2022 is coming to an end. The Hematoscope Lab is thrilled on the progress made during our time in operation 2021-2022. Here’s a sneak peak on what we have been doing and where we are now.

Our team

Our team includes 10 members.

  • Johanna Heikkinen and Anni Dohlen are the hard-working technicians responsible for creating the high-resolution image data our lab is established on.
  • Otso Brummer, Stas Tatun, and Mikko Purhonen are the genious minds behind our algorithms. They have employed various types of convolutional neural networks to mimick as much as possible the thinking of the hematopathologists. Otso and Mikko finalized their engineer’s theses in deep learning-guided medical image analysis in 2022.
  • Joonas Palosuo is the talent behind the frontend of the Hematoscope application, the tool that links the algorithms to clinical use.
  • Jaana Friman has instructed others on cytomorphological analysis to accurately annotate images.
  • Sanna Siitonen and Kimmo Porkka have helped me to coordinate the team, vision the computational platform in routine diagnostics and integrate EHR-wide clinical data to furure algorithms.

    Collaboration

    We are happy to lead two international multi-center studies in acute and chronic myeloid leukemias with partners from Norway (Prof. Henrik Hjorth-Hansen), Sweden (Adj. Prof. Ulla Olsson-Strömberg, Prof. Sören Lehmann), Finland (Kontro Lab, Paavolainen Lab, Samuli Rounioja, Mustjoki Lab), Japan (Prof. Takashi Kumagai and Prof. Shinya Kimura), and Australia (Prof. David Ross). These studies are still heavily ongoing, but we expect to share something next year.

    Funding

    We have been actively applying for funding. Most applications unfortunately rejected, but an important proportion has returned positive. For all their support during 2020-2024, The Hematoscope Lab wants to sincerely thank the Helsinki University Hospital, the Finnish Medical Foundation (Lääketieteen Säätiö), the Finnish Cancer Foundation (Syöpäsäätiö), the Runar Bäckström Foundation, the Signe & Ane Gyllenberg Foundation, the Minerva Foundation (Ralph Gräsbeck Scholarship), the Swedish-Finnish Medical Association (Finska Läkaresällskapet), K. Albin Johansson Foundation, Paulo Foundation, Pfizer and Gilead Sciences. Without your support, none of this work would have been done and will ensure that we will achieve our research objectives!

    Achievements so far

    Without doubt, the most important achievement is creating an environment, where researchers can develop their careers, learn new skills, and enjoy working together. Also, I feel overwhelmed by the trust and will of our collaborators to contribute their time and effort to our studies.

    Main achievements

    1. The world’s largest dataset of digital high-resolution (63-100x) May-Grünwald-Giemsa (MGG) slides. These are analogical to H&E-stained slides commonly used in histopathology. MGG-stained slides are challenging to digitize as these require immersion oil and ≥63-100x microscope objectives. By the end of the year, we have digized over 40 000 bone marrow slides including all pediatric and almost all adult hematological patients of the Helsinki University Hospital district for 2009-2022. Next in line - cerebrospinal and peripheral blood samples.
    2. Image analysis algorithms. We deeply believe in supervised learning. Our team has created a pipeline to optimize image annotation and annotated thousands of images. We have dedicated this Fall to improve computational speed. The speed up for some algorithms has been around 10-fold, which likely does not say anything but is fundamental if an algorithm needs to be run for >10 000 samples. Most of our models are currently being validated and we look forward to present these in more detail in 2023.
    3. The Hematoscope app. My #1 frustration in cancer research is the lack of translation. In other words, how can we translate the innovative work of researchers to clinical practice? What aspects need to be considered in algorithm infrastructure considering the IT development in the next 10 years? The Hematoscope app will address these questions. Simply put, it connects the bioanalyst, the hematopathologist, the electronic health records, and the deep learning networks together. Official launch expected in 2023.
    4. Publications. We are working on 2 manuscripts in 2023. Stay tuned!

      Wishing you a happy holiday season and prosperous year 2023!
      On the behalf of the Hematoscope Lab,
      Oscar Brück

Previous post
Business Finland Cancer Partnering Days Next post
Medical Device Regulation (MDR) Specialist