Collaboration will use Deep Learning techniques to establish medical imaging biomarkers for more accurate diagnosis

Median Technologies (Paris: ALMDT), the industry-leading Imaging Phenomics Company® and the Nice University Hospital (CHU de Nice) have announced a collaborative agreement that uses Artificial Intelligence to identify medical imaging biomarkers for lung cancer screening. These efforts will enable more accurate diagnosis and provide physicians with new therapeutic decision-making tools, based on medical imaging.

As part of the collaboration, medical imaging data from the AIR study - a French, multicenter cohort study, led by the Nice Hospital that has enrolled, to-date, more than 600 high-risk patients (smokers or former smokers with Chronic Obstructive Pulmonary Disease [COPD]) screened for lung cancer - will be analyzed to identify and characterize pulmonary nodules visible in thoracic CT scans. By using Deep Learning methods, a discipline of Artificial Intelligence, Median will develop new algorithms to identify imaging biomarkers that indicate pulmonary nodule malignity.

While current CT scan performance enables more pulmonary abnormalities to be identified, post-treatment image applications do not allow for an automatic, accurate characterization of the malignity or benignity of these pulmonary abnormalities. Lung nodule biopsies, which are invasive, are needed to confirm a diagnosis - potentially leading to complications for patients. By using medical imaging biomarkers, clinicians can reduce unnecessary biopsies and more accurately diagnose patients.

"Early detection of lung cancer is of paramount importance if we want to lessen mortality of this disease", says Professor Charles Marquette, coordinator of clinical teams in the AIR study. "The rationale for screening is based on the tight relationship between outcome and extent of the disease at time of diagnosis. However, large scale screening of unselected population with chest computed tomography (CT) is expensive and has a high harm to benefit ratio, which explains why many health agencies are reluctant to implement screening of lung cancer with chest CT alone. We are developing a multimodal approach to lung cancer screening, including refinement of screening criteria (e.g. focus on COPD), non-invasive biomarkers and use of Artificial Intelligence to better characterize chest CT findings. We place a great deal of hope in our partnership with Median Technologies" added Professor Marquette.

"We are thrilled to collaborate with the Nice University Hospital", says Peter Bannister, Chief Technology Officer at Median Technologies. "This collaboration illustrates the fundamental contribution of Artificial Intelligence in the field of health, and more particularly in medical imaging. Today, many pulmonary biopsies are performed unnecessarily; Artificial Intelligence is going to make imaging, which represents non-invasive and less expensive procedures, an improved therapeutic decision-making tool. With Artificial Intelligence, imaging will help to identify patients who really need a biopsy and will contribute to advance clinical practice," added Bannister.