Unlocking the full value potential in digital pathology

DoMore Diagnostics, one of our NLS Stars 2022, aims to change the one-size-fits-all approach in cancer therapy.

Cancer is a growing global public health problem, and the cost of cancer care is expected to double the next 20 years. It is a leading cause of death worldwide and accounted for nearly 10 million deaths in 2020, according to the WHO. While cancer is a leading cause of death, a major challenge is also over-treatment as it is difficult to identify patients with low risk who has no benefit from toxic treatments. The course of a cancer disease varies widely among patients as the tumors are very heterogeneous.

“Deep-learning image analysis has the potential to decode the complexity in the tumors and personalize medicine.”

“Each patient and each tumor is unique, so when the cancer is detected it is critical for patients and healthcare systems to determine the individual patient’s need for treatment. Current practice has a one-size-fits-all approach for most patients, leading to either over-treatment or under-treatment of many patients. Deep-learning image analysis has the potential to decode the complexity in the tumors and personalize medicine,” says Torbjørn Furuseth, CEO and co-founder of the Norwegian company DoMore Diagnostics.

On the verge of a breakthrough

DoMore Diagnostics, founded in 2020, is a spin-out from a large research program, the DoMore Project, led by Oslo University Hospital in collaboration with University of Oxford and University College of London. The company uses artificial intelligence (AI) to predict the survival outcome of each patient based on standard tumor histology images.

The project was initiated by Professor Håvard Danielsen at the Institute for Cancer Genetics and Informatics (ICGI) at Oslo University Hospital. In May this year Danielsen was awarded the King Olav Vs Cancer Research Award of 2022 as a trailblazer for connecting IT technologies in health. Together with his research team at ICGI he has trained computers to recognize cancer tumors, their characteristics and the patient’s prognosis, by feeding the computer with millions of image tiles of samples from tumors of cancer patients.

“The method was first published in the Lancet, and a new article was published in the Lancet Oncology, where the method is combined with tumor and lymph node status, further strengthening the diagnostic power,” describes Torbjørn. “Our tests have a short turnaround time, are cost-effective, and are integrated into current digital workflows, which makes further development and implementation fast and scalable.”

AI in digital pathology is on the verge of a breakthrough, and outcome prediction in cancer patients can unlock the full value potential in digital pathology, emphasizes Torbjørn. The results can guide patient treatment and make drug development more effective.

“I also believe that our algorithms can be combined with other cancer diagnostics tools like gene profiling and ctDNA to strengthen the diagnostic accuracy.”

“Digital pathology is experiencing an accelerated adoption generally in pathology labs, with several AI tools becoming available. DoMore is well positioned ahead of the curve. I also believe that our algorithms can be combined with other cancer diagnostics tools like gene profiling and ctDNA to strengthen the diagnostic accuracy,” he says.

Torbjørn Furuseth, CEO and co-founder, DoMore Diagnostics and Håvard Danielsen, Professor, the Institute for Cancer Genetics and Informatics (ICGI), Oslo University Hospital. Photo: ICGI

Commercialization and development

DoMore Diagnostics recently achieved a significant milestone by receiving CE-IVD certification for the Histotype Px Colorectal, its AI algorithm that predicts patient outcomes in colorectal cancer patients by analyzing standard histology images.

“We were quite proud of achieving this so fast. The CE-IVD mark allows us to market the product in Europe for clinical use, and is a proof that the company has implemented a robust quality system and a thorough documentation of the product,” says Torbjørn.

Going forward Torbjørn and his team will continue to develop and document their deep-learning algorithms and facilitate implementation into standard practice as an important marker to optimize patient treatment.

“At a later stage, I envision that we will start to develop our own algorithms in collaborations with universities or other cancer diagnostic companies.”

“In the first phase we will focus on commercializing the products we currently have in the pipeline from the license agreement with Oslo University Hospital. At a later stage, I envision that we will start to develop our own algorithms in collaborations with universities or other cancer diagnostic companies,” he says.

There is a large space of opportunities, so it is important to stay focused, he adds. “We are working in more cancer indications, such as prostate cancer, and have very interesting results. Furthermore, we want to expand the clinical documentation of our colorectal algorithm in more settings and geographies.”

Time and resources

There are several advantages to being a Norwegian life science company, according to Torbjørn Furuseth.

“The obvious one for us is that our invention comes from Oslo University Hospital. Generally speaking, the soft funding opportunities are good, and there is growing interest and capacity from Norwegian investors to fund early stage life science companies,” he says.

“Our main challenge is to enter a conservative and slow-moving market. It takes time and resources to attract attention and interest for your product, and we compete against many other companies and products.”

One challenge is to find employees and consultants with the right competence, he adds. “There is not a lot of people who has worked with commercializing AI image analysis and diagnostics. However, I am glad to see that we have been able to establish a very strong team. A great team is critical for every startup. Our main challenge is to enter a conservative and slow-moving market. It takes time and resources to attract attention and interest for your product, and we compete against many other companies and products.”

Featured image of Torbjørn Furuseth