The algorithm is the work of Oncode Investigator Jeroen de Ridder in collaboration with Joske Ubels, who, when the project started in 2019 was a PhD student in de Ridder’s lab and a Skyline employee. Oncode then, in collaboration with Utrecht Holdings, put in place the agreement to grant Skyline the option to acquire the IP - which they exercised - and then worked to put the license in place.

"Science is teamwork, but making science actually benefit the patient requires Teamwork with a capital T. It is therefore fantastic to see that with the support of Oncode Institute and Utrecht Holdings, the basic research we do at the UMCU now finds its way to SkylineDx, where it will be used as a foundation for further development of products that improve the diagnosis of cancer patients," says de Ridder.

The algorithm is based on Random Forest, a specific machine learning method, and is able to identify subgroups of patients that will benefit more than the population as a whole from a certain treatment. The performance of this algorithm is demonstrated on a dataset from a phase III colorectal cancer study. This study did not show any benefit for the drug (Cetuximab) under investigation, meaning that patients had the same risk of progressing whether they would have received Cetuximab or not. RAINFOREST, however, was able to identify a subgroup of patients (27.7%) that did significantly benefit and had a lower risk of progression if they would have been treated with the new drug.

RAINFOREST follows the manuscript describing the innovative algorithm GESTURE and both show, although based on different methodology and input information, the importance of subgroup analysis in clinical trials where oncological drugs have been investigated. More than half of the clinical trials fail because the new drug is not able to show enough efficacy in the entire patient population it is being tested on.

RAINFOREST and GESTURE demonstrate that these resources are not wasted if you can find the specific patients that do benefit and develop a biomarker that is able to identify these patients upfront in routine clinical practice. These algorithms can be applied to study results concerning different types of cancer as well as different treatment regimes.

“Oncode is proud to have facilitated the transfer of this novel algorithm developed in the lab of Oncode Investigator Jeroen de Ridder to SkylineDx”, says Ian Bell, Oncode’s Interim Valorization Director. “This agreement not only supports Oncode’s mission of advancing opportunities that impact the treatment and health of cancer patients, but also confirms our support of the Dutch biotech industry, through licensing to homegrown companies. We look forward to seeing the success SkylineDx can achieve with RAINFOREST”.