Abstract #0445

Supervised shallow learning of 129Xe MRI texture features to predict response to Anti-IL-5 biologic therapy in severe asthma

Marrissa McIntosh¹, Rachel Eddy¹, Danielle Knipping², Tamas Lindenmaier², David McCormack³, Christopher Licskai³, Cory Yamashita³, and Grace Parraga⁴

¹Department of Medical Biophysics, Robarts Research Institute, Western University, London, ON, Canada, ²Robarts Research Institute, Western University, London, ON, Canada, ³Division of Respirology, Department of Medicine, Western University, London, ON, Canada, ⁴Department of Medical Biophysics, Division of Respirology, Department of Medicine, Robarts Research Institute, Western University, London, ON, Canada

¹²⁹Xe MRI ventilation images consist of embedded texture features that help explain abnormal ventilation heterogeneity. We postulated that such texture features may help predict severe asthma patient response to anti-IL-5 therapies. Therefore, we employed supervised shallow learning techniques to identify specific ¹²⁹Xe MRI features that help predict anti-IL-5 responders. Texture analysis yielded features that were superior to clinical measurements in identifying severe asthma patients that responded to anti-IL-5 therapy after 28 days. These promising results suggest that texture analysis may help predict asthmatics more likely to respond, before treatment is initiated.

This abstract and the presentation materials are available to 2020 meeting attendees and eLibrary customers only; a login is required.

Join Here