Speakers:
Student Presentation
Amir Jabbarpour, PhD Candidate, Carleton University
Title: AI for Detection of Pulmonary Embolism—VQ-SPRINT: Segmentation, Pseudo-planar Generation, and Registration INtegration Tool
Abstract:
Pulmonary embolism (PE) remains a critical diagnostic challenge in ventilation–perfusion (V/Q) imaging due to limited spatial localization, inter-observer variability, and the continued reliance on planar acquisitions in clinical practice. This presentation introduces VQ-SPRINT, an integrated artificial intelligence (AI) framework for PE detection that unifies Segmentation, Pseudo-planar Generation, and Registration INtegration into a single end-to-end pipeline.
VQ-SPRINT encompasses a series of complementary deep learning models developed across the V/Q imaging workflow, including automated defect segmentation, synthesis of pseudo-planar perfusion images from volumetric SPECT data, and AI-based bronchopulmonary atlas image registration for anatomical alignment and quantitative defect analysis. By harmonizing these components, the framework enables computer-assisted detection and localization of perfusion defects on both real and synthetic planar images, while facilitating communication and comparison of findings across defects, projections, and patients
Collectively, this work demonstrates how integrated AI can enhance the diagnostic utility of V/Q imaging, revitalizes AI in V/Q abandoned field, and provide a scalable foundation for clinically deployable PE decision-support tools.
Member Presentation
Reid Townson, PhD, Research Officer, Metrology, National Research Council of Canada
Title: An update on EGSnrc developments and direction
Abstract:
EGSnrc continues to evolve, and this presentation will highlight recent additions as well as delve into some of the challenges of scientific software development. How does one validate a change so dramatic as to affect the way material properties are defined? This talk also serves as your direct line to an EGSnrc developer - are there missing capabilities that could benefit your research? What do you see as the horizon of Monte Carlo simulation needs?