Abstract #3814

Addressing multi-centre image registration of 3T arterial spin labeling images from the GENetic Frontotemporal dementia Initiative (GENFI)

Henk Mutsaerts¹, David Thomas², Jan Petr³, Enrico de Vita², David Cash², Matthias Van Osch⁴, Paul Groot⁵, John Van Swieten⁶, Robert Laforce Jr⁷, Fabrizio Tagliavini⁸, Barbara Borroni⁹, Daniela Galimberti⁸, James Rowe¹⁰, Caroline Graff¹¹, Giovanni Frisoni⁹, Elizabeth Finger¹², Sandro Sorbi¹³, Alexandre Mendonça¹⁴, Martin Rossor², Jonathan Rohrer², Mario Masellis¹, and Bradley MacIntosh¹

¹Sunnybrook Research Institute, Toronto, ON, Canada, ²London, United Kingdom, ³Dresden, Germany, ⁴Leiden, Netherlands, ⁵Amsterdam, Netherlands, ⁶Rotterdam, Netherlands, ⁷Quebec City, QC, Canada, ⁸Milan, Italy, ⁹Brescia, Italy, ¹⁰Cambridge, United Kingdom, ¹¹Stockholm, Sweden, ¹²London, ON, Canada, ¹³Florence, Italy, ¹⁴Lisbon, Portugal

One obstacle in multi-centre arterial spin labeling (ASL) studies is the variability attributed to differences between vendor- or site-specific ASL implementations. This multi-centre study compares spatial registration methods from ASL to 3D-T1, to reduce the between-subject variability of cerebral blood flow (CBF) maps. Our results demonstrate that choices of image registration have profound effects on ASL data collected using different pulse sequences and/or sites. A rigid-body registration of CBF images to segmented gray matter images produced the most robust similarity outcome as a standard approach across the different ASL implementations.

This abstract and the presentation materials are available to members only; a login is required.

Join Here