AI colorectal cancer diagnostics receives CE-IVD mark

DoMore Diagnostics, a developer of pathology AI algorithms, has achieved CE-IVD mark for the Histotype Px Colorectal, a deep learning algorithm that predict patient outcome based on analysis of digital histology images.

The product uses artificial intelligence to predict patient outcome based on advanced image analysis. Histotype Px Colorectal builds on the research led by Institute for Cancer Genetics and Informatics at Oslo University Hospital, and the deep learning algorithm is trained on and developed with close to 100 million image tiles. The analysis is conducted on high resolution scans from standard H&E stained slides providing shorter turnaround times.

Every year close to 2 million patients worldwide are diagnosed with colorectal cancer, and despite decades of comprehensive research and large investments into gene sequencing, there is large medical need for better prognostic and predictive markers for patients with colorectal cancer. Many of these patients will suffer from unnecessary and harmful overtreatment while others may receive less intensive treatment than they need.

The results of an early version of the algorithm were first published in The Lancet. During the last year, DoMore Diagnostics has improved the algorithm and developed a commercially available product.

"With the Histotype Px Colorectal, clinicians and patients can receive patient risk information immediately after surgery, weeks to months faster than markers available today", Dr Tomas Nordheim Alme, MD said.

DoMore Diagnostics CEO and co-founder Torbjørn Furuseth added: “We are very excited about this CE-mark as it enables clinical use across Europe. The product fits well within the existing hospital workflow and adds real value to the various digital pathology platforms available. This product makes so much sense for patients, healthcare providers and payers. With the simplicity and unmatched turnaround time we can provide physicians with additional valuable information to make the best treatment decisions for patients. We believe prognostic and predictive algorithms will unlock the full value of digital pathology.”