Results

In the DoMore! project, we see the need for a lot of new development. We will need new applications as well as further development and new functionality in the existing applications. Many of the tools that will be developed and applied over the course of the DoMore! project are the culmination of work completed by and with our partners over the years, with the foremost being: automatic delineation, pathology imaging tools, segmentation, DNA ploidy in tissue sections, texture analysis in tissue sections, and MicroTracker.

Beginning with automatic delineation (WP2), today, pathologists manually define the tumor area, known as the region of interest (ROI). This is a step we want to automate to save resources and obtain objective results. This is currently a development project with one of our commercial partners. Building on this, and our experience with in silico grading, we aim to develop and establish new objective grading systems for colorectal cancer and lung cancer. If successful, these will be products with potential applications for millions of cancer patients around the world.

The next method is segmentation (WP3). This is a prerequisite for all image analysis, and particularly challenging in tissue sections. We have developed a successful method for prostate cancer10, and intend to do the same for colorectal cancer and lung cancer. On the basis of this, we expect to deliver a general solution for segmentation of nuclei in tissue sections. This will open the field of in silico pathology for other research groups and companies.

We have previously developed methods for DNA ploidy and Nucleotyping and demonstrated their usefulness as clinical prognostic markers. DNA ploidy is also established as a diagnostic service at Oslo University Hospital and productized. This analysis requires special sample preparations We now aim to develop these methods for similar analysis in routine tissue sections , which will make the methods fully automated, more applicable and less resource-demanding.

Currently, we have done a proof of concept with ploidy analysis in prostate cancer, with >80% correlation between monolayers and sections. To further improve throughput, we intend to go from a microscopy platform to a high resolution scanner. These results will provide both a product and a service for general use in pathology with huge potential in cancer prognosis.

The development of a clinical decision support system for prognostic markers in our selected cancer types will have a direct impact on the clinical use of prognostic markers, and hopefully a strong impact on the treatment of these cancers. In addition to the prognostic methods discussed above, we will establish stroma fraction and proliferation index as in silico markers in routine tissue sections. All markers will be tested and validated in our large cohorts of 7,000 patients altogether, and will be published in high impact international journals.

To summarize, we expect a number of different project results; increased efficiency in pathology, methods and markers to aid the clinician to give better and more personalized treatment to cancer patients, patents and publications, products (algorithms, applications, services, data) and spin-off companies.


© DoMore! | Editor-in-chief: Håvard E. Greger Danielsen | Webmaster: Marian Seiergren