By scanning thousands of tissue samples, a deep learning program developed by Owkin was able to help identify new types of patients with mesothelioma and predict which may respond better to certain therapies.
Using whole biopsy slide images taken from nearly 3,000 patients with the aggressive and potentially fatal disease, the artificial intelligence model was able to pick out novel biological features of the tumor and link them to its overall prognosis.
“Patients with mesothelioma exhibit a very high variability in survival, from a few months to a few years, and this makes it challenging for doctors to plan treatment and to care for these patients,” Owkin CEO Thomas Clozel said in a statement.
“Our research helps to explain the biological causes of this variation and will ultimately lead to the development of more targeted drugs and better management of this terrible disease,” Clozel added.
Known as MesoNet, the program was trained using the French national Mesobank database of tissues and cell lines.
Pathologists at the Centre Leon Berard cancer institute in Lyon, France, validated the model’s results, saying it outperformed all existing survival models, according to Owkin. In addition, MesoNet was able to correctly peruse the heterogeneity of the disease when tested against differently stained images provided by The Cancer Genome Atlas. Owkin’s results were published in Nature Medicine.
Additionally, the model was able to visually highlight areas of the image associated with its predictions to help pathologists stratify patients.
“It was a great experience for our lab to work closely with the Owkin scientists to identify new subgroups within our patient population. The collaboration exceeded our expectations,” said Françoise Galateau, a professor of pathology at Centre Leon Berard. “As well as improving our prognostic models, MesoNet was able to identify new biomarkers within the stromal regions of the tumor microenvironment that were predictive of survival.”
Going forward, Owkin said it is working with its partners in the biopharma industry to help guide patient selection and assignment in clinical trials by identifying those most at risk.