French American biotech company Owkin has shared validation data for MSIntuit CRC, its artificial intelligence-powered digital pre-screening solution for optimizing colorectal cancer diagnosis and treatment. Details of the validation process and Owkin’s results were published in Nature Communications in a paper titled, “Validation of MSIntuit as an AI-based pre-screening tool for MSI detection from colorectal cancer histology slides.”
Owkin’s solution screens patients for microsatellite instability (MSI), a key genomic biomarker in colorectal cancer found in 15% of cases. Evidence from some recent clinical trials indicate that the MSI phenotype has prognostic and therapeutic importance. In fact, patients whose cancers show MSI are more likely to respond to immune checkpoint inhibitor (ICI) therapies such as pembrolizumab. These therapies are not recommended for colorectal cancer patients that are microsatellite stable (MSS).
Given MSI’s importance as a colorectal cancer biomarker, organizations like the National Comprehensive Cancer Network recommend screening newly diagnosed colorectal cancer cases for evidence of MSI. Pre-screening helps minimize the time spent exploring various treatment options and ensures that patients get the most effective treatments for their tumors the first time around. Screening for MSI can be done using methods like immunohistochemistry, PCR, or next-generation sequencing but each method has limitations including cost and turnaround time.
Results from Owkin’s validation study suggest that AI-based solutions could be taught to reliably screen patients in the clinic. For the study, Owkin trained its AI model on samples from the Cancer Genome Atlas. They then performed a blind validation of the model on a separate dataset of 600 colorectal cancer cases that were diagnosed across nine pathology labs. According to the paper, MSIntuit CRC accurately rules out nearly 50% of MSS patients, while correctly classifying more than 96% of MSI patients, which is consistent with the current gold standard. Its results were consistent across two scanners and slides containing samples from different tumor regions. And oncologists could get results much sooner—a matter of hours rather than days.
Proponents of the solution believe in its potential to change how colorectal cancer cases are screened and improve patients’ outcomes. Specifically it will “have a direct impact on oncologist decision making and help bring the best treatment to patients sooner,” said Magali Svrcek, PhD, professor at Saint Antoine Hospital, Sorbonne Université, and a co-author of the publication. “It could also optimize costs and organization of MSI testing in pathology labs, especially for countries applying universal MSI screening.”
Meriem Sefta, PhD, Owkin’s chief diagnostic officer, added, “Our solution represents the first step towards the development of an AI diagnostic that can identify actionable biomarkers from a single H&E slide used in clinical routine.”
Owkin appears to be ramping up its activities in the drug discovery space through partnerships with pharma and other companies. For example, last month the company announced a partnership with Servier, an independent pharmaceutical group that is focused on using AI to advance improved therapies across multiple disease areas.
The two companies have said that they will focus initially on two major challenges in translational medicine and digital pathology. Under the terms of their agreement, Owkin will apply its machine learning capabilities to Servier’s clinical data with an eye toward extracting novel insights about disease biology and identifying patient populations most likely to benefit from Servier’s new therapies.
Their first project will focus on identifying tumor types, tumor microenvironments, and patient subgroups that will best respond to an unnamed Servier asset. The partners will analyze data on tumor evolution to identify drug combinations that may act on additional immune checkpoints or tumor cell-intrinsic mechanisms. Their second project will explore the implementation of digital pathology and its potential to speed up screening and enable broader analysis of tissue-based biomarkers.
Owkin also has an agreement with Genmab, a biotech company that specializes in developing antibody-based therapeutics for cancer and other diseases. The agreement allows Owkin to use its platform to evaluate the impact of Genmab’s clinical assets on certain cancer types.
