RI-MUHC study uses AI to understand liver cancer  – Health e-News

A new publication describes the use of ALiX, a unique AI model, to analyze data from liquid biopsy

In work recently published in the journal Informatics in Medicine Unlocked, researchers from the Research Institute of the McGill University Health Centre (RI-MUHC) together with Modal Technology Corporation are using a unique AI system known as ALiX to seek more personalised and less invasive treatments for patients with liver cancers.

A liquid biopsy is a blood test that seeks signs of cancer, including tumour cells and DNA. Research is ongoing to determine how liquid biopsy may effectively replace surgical biopsies, which are more costly, painful and sometimes high risk. However, liquid biopsy is still an emerging science.

“Accurate, low-cost liquid biopsy can transform how we diagnose and manage cancer,” says Peter Metrakos, MD, a Senior Scientist in the Cancer Research Program at the RI-MUHC and a surgeon at the McGill University Health Centre. “We are actively seeking effective biomarkers as well as new ways to manage the vast amounts of data being generated in this research.”

Dr. Metrakos, together with Anthoula Lazaris, PhD, a Scientist at the RI-MUHC, have built an important tissue and liquid database through the MUHC Liver Disease Biobank. For this study, the team isolated extracellular vesicles from the blood of patients with colorectal cancer liver metastases, both before resection (with disease) and after resection (without disease). The proteins found in these vesicles are biological molecules, also known as biomarkers that can help identify cancer and disease progression.

Recognizing the wealth of invaluable information they had generated, Metrakos and Lazaris decided to explore the promise of artificial intelligence (AI). They forged a partnership with Modal Technology Corporation, a Minnesota-based company renowned for its expertise in AI tailored for the life sciences.

“Modal’s software, known as ALiX, can provide deep insights from data like ours with an unprecedented level of confidence,” says Anthoula Lazaris. “One of the key reasons for that is because ALiX is the world’s first AI platform not based on statistics but set theory.”

The researchers applied the ALiX software to their data in an effort to predict the response of a patient to their treatment for colorectal cancer liver metastases. From the patient’s blood, they were able to identify a unique blood profile with specific biomarkers. The ALiX software ranked the biomarkers according to their relative significance, and allowed the team to identify those which were key to predicting response to treatment.

“To our knowledge, there is no other AI system or method that can provide a ranked list of proteins like this,” says Anthoula Lazaris. “Of course, the ranked biomarker solution ALiX has provided now leads us to new questions and further hypotheses. We need to develop new strategies to understand the wealth of information that can be found in blood.”

Key to this research was the human samples obtained from patients who consented to participate in the Liver Disease Biobank. Without their support, this work would not have been possible. The research team is immensely grateful.

About the study:

Read the publication: Predictive biomarker discovery in cancer using a unique AI model based on set theory in the journal Informatics in Medicine Unlocked. First published: 20 March 2024 https://doi.org/10.1016/j.imu.2024.101481

The authors gratefully acknowledge funding from the McGill University Health Centre Foundation.

The authors thank the patients who have contributed to the MUHC Liver Disease Biobank. Without their support, this initiative would not have been possible.

About Modal Technology Corporation

Modal Technology Corporation is replacing the current generation of statistical artificial intelligence (AI) with a new platform based on a branch of applied mathematics known as modal interval arithmetic. This new platform, called ALiX™, is the competitive advantage and differentiating factor to faster, better and simplified AI. Modal licenses this technology to partners and helps them design custom AI solutions that capture the added value of using a non-statistical approach in their AI applications. Modal Technology Corporation is a privately held Minnesota corporation headquartered in Minneapolis, Minnesota, USA.

Related news

Liquid biopsy research: a generous boost to help children with cancer
Exosomes, vesicles and particles are no longer obscure