IRIC - Institute for Research in Immunology and Cancer

Published on February 8, 2018

It will become easier to discover imaging biomarkers and identify digital surrogates for modern end-points in personalized therapies, thanks to a partnership between Sylvain Meloche – Principal Investigator within the signaling and cell growth research unit at the Institute for Research in Immunology and Cancer (IRIC) along with his team, and Imagia, an AI-driven personalized healthcare company.

Sylvain Meloche’s team will try to accurately distinguish different types of cancers using artificial intelligence solutions applied to tumor imaging. “What we need to understand is that each tumour is unique and is characterized by a specific genetic signature, that is, it is generated by the hyperactivation or deficiency of one or more genes,” Meloche says. The goal is to apply Imagia’s Deep Radiomic algorithms to process tumor images from in vivo models and identify imaging biomarkers specific to each type of cancer.

Imagia’s platform allows researchers to take part in an in silico discovery process to investigate hypotheses that would otherwise require extensive benchwork. “Applying Deep Radiomics to tumour images allows us to dissect the oncologic processes such as signal pathway alterations and tumor immune phenotypes from routine medical images,” says Dr. Kam Kafi, Imagia’s director of oncology. According to Sylvain Meloche, it would be a first to distinguish tumors with different molecular causes using image analysis alone.

This collaboration will advance toward some of the promises of the emerging field of radiomics, to predict patient outcome from routine clinical imaging, and enable a cost-effective assessment of disease progression and treatment response. “Our Deep Radiomics platform offers scalable radiomics discovery pipelines, leveraging the unique ability of deep learning to discover actionable imaging biomarkers associated to global tumor status,” says Sébastien Giguère, research scientist (ML) at Imagia.

Easier diagnosis

Liver cancer is diagnosed by imaging techniques and conventional histological analysis after a biopsy. However, neither of these tests can identify the molecular cause of cancer. The only means are to undertake a complex histopathological analysis or sequence the entire tumor genome to know its genetic signature. Imaging biomarkers offer a simpler and less invasive alternative. The sample to be decrypted no longer comes from a biopsy, but is a simple radiological image.

The advantage of being able to accurately identify a tumor type is that it is possible to offer suitable and personalized treatments. The partnership agreement between Imagia and Sylvain Meloche’s laboratory, supported by IRICoR (Institute for Research in Immunology and Cancer – Commercialization of Research), is excellent news that could pave the way for more personalized treatment solutions for cancer.