An invention of the Finnish doctors Johan and Mikael Lundin provides an effective solution for the analysis and storage of tissue section images.
The volume of research data is increasing enormously year after year, requiring a continuously active approach from software developers. It must be possible to analyse large amounts of data with software that does not jam the workstation. Johan Lundin, Research Director at the Institute for Molecular Medicine Finland (FIMM), studies and develops image-based diagnostics using machine vision solutions. In future, it will become possible to produce individualised disease prognoses and treatments by combining various data sources, genetic data, tissue data and clinical patient data. This has been applied especially in the treatment of breast, prostate and colon cancer.
When he was working at the Helsinki University Hospital in the early 2000s, Lundin became frustrated with how difficult it was to process large tissue section images at the workstations. The size of tissue section images is 1–2 gigabytes, so storing them on your own hard drives does not make sense. Rotating the images is also slow. With his brother Mikael, Johan Lundin started thinking about a functional software solution for the problem.
The brothers developed a fully web-based software program, the essential components of which are an efficient image server and a web interface that works with all browsers. With their compression algorithm, images take up less space and load quickly. A two gigabyte sample image can be compressed to the size of half a gigabyte. The tissue sample is stored in the cloud and large amounts of data can be processed quickly and easily from your own workstation.
The online microscope service can be used with all browsers and tablets, including smartphones. WebMicroscope® is also compatible with the image formats of different microscope manufacturers. WebMicroscope enables the study of very extensive materials and is also ideal for collaborative projects as a joint management and analysis space for digitised images.
“There has been a growing interest in the service. Doctors, researchers and teachers are shifting to digital microscopy. An online cloud-based service is a progressive solution for the users of digital microscopy all over the world”, says Kaisa Helminen, CEO of the service’s provider, Fimmic. Helminen is a trained biochemist and has previously worked for several companies in the bio-industry.
Fimmic was established in 2013 and the commercialisation of the service started the following year. Fimmic’s customers include universities, research institutes, pharmaceutical companies and companies conducting external quality control. External quality control is enhanced when samples can be sent for analysis virtually instead of mailing samples on glass slides to laboratories.
Fimmic uses the cPouta cloud service of CSC’s data centre as a partner in providing these services. It offers the WebMicroscope users their own server, a high-speed bandwidth and a massive amount of storage space. This ensures that the service works as efficiently as possible. WebMicroscope is also suitable for biobanks for tissue sample management. The service can be customised to suit a particular biobank.
Microscope scanners are expensive devices with the price typically varying between 150,000 and 300,000 euros. However, the number of scanners is increasing and, when images are scanned, the most convenient and least expensive solution for many users is to store them directly in the cloud.
“If a customer does not have the opportunity to use a scanner, the samples can be sent to us for scanning. We will store the digitised samples directly in the customer’s WebMicroscope account”, Helminen says.
Through the WebMicroscope portal, users can share their own microscope images with different research groups and partners around the world. This is an important feature because in drug design, for example, the rapid sharing of test results between research groups and pharmaceutical companies is a prerequisite for breakthroughs. Research related to drug development is one of Fimmic’s focuses.
With a traditional microscope, only a small portion of a sample can be examined at one time. A microscope scanner takes a picture of the sample with a large objective, digitising the entire sample in detail. With WebMicroscope, the resulting image can be viewed easily and quickly, regardless of location.
“You can select a section of the tissue sample to be examined, similarly to Google Maps, and only look at a part of it, quickly moving to another spot. The image is not saved on the workstations, but is rather loaded over the network directly from the image server.”
All Finnish universities teaching medicine use WebMicroscope for educational purposes at anatomy and pathology courses. WebMicroscope allows digitised samples to be easily shared with students along with other documents and videos.
You can secure your own pages with a password and the software can also be used to complete exams. The virtual samples can be viewed using a tablet or a smartphone in distance education, or on a large screen in the classroom. The application is ideal for multi-touch screens that utilise multiple points of contact. Massive tissue section images can then be viewed easily and quickly on a large touch screen even with a larger group.
A microscope scanner produces a lot of data. There may be millions of observation points to examine, the processing of which requires computing power and good algorithms. Fimmic plans to further develop the software and introduce quantitative image analysis tools, algorithms. According to Kaisa Helminen, the number of potential research subjects that algorithms can be used for is huge.
“Machine vision algorithms are based on signal processing. With dozens or even hundreds of images, the machine is taught to identify a particular signal from the background, for example stained cells from other tissue. Screening is case-specific and it varies how different samples have been processed. An algorithm is just as good as it has been taught to be.”
All of this requires computing power obtained, for example, from the supercomputers of CSC – IT Center for Science.
“A lot of computing power is required because the images being studied are so-called whole slide images. Smaller sections of these may, of course, be selected for analysis, but a lot of computing power is still required so that the analysis would not take too much time”, Kaisa Helminen notes.
Ari Turunen
Article in PDF
Further information:
CSC – IT Center for Science
CSC – The Finnish IT Center For Science is a non-profit, state-owned company administered by the Ministry of Education and Culture. CSC maintains and develops the state-owned, centralised IT infrastructure.
http://www.csc.fi
https://research.csc.fi/cloud-computing
ELIXIR
ELIXIR builds infrastructure in support of the biological sector. It brings together the leading organisations of 17 European countries and the EMBL European Molecular Biology Laboratory to form a common infrastructure for biological information. CSC – IT Center for Science is the Finnish
centre within this infrastructure.
https://www.elixir-finland.org
http://www.elixir-europe.org
ELIXIR is partly funded by the European Commission