STAGE As a Multicenter, Multivendor Protocol for Imaging Parkinson's Disease: a Validation Study on Healthy Controls
He Naying,Wu Bo,Liu Yu,Zhang Chunyan,Cheng Jingliang,Gao Bingbing,Miao Yanwei,Wu Wenjun,Wang Lixia,Sun Rongqing,Sun Wenbo,Xu Haibo,Bai Yan,Wang Meiyun,Chai Chao,Xia Shuang,Zheng Qiao,Li Yongmei,Qin Yan,Liao Weihua,Chen Yongsheng,Jokar Mojtaba,Wang Ying,Yan Fuhua,Haacke E. Mark
DOI: https://doi.org/10.1007/s42058-022-00089-3
2022-01-01
Chinese Journal of Academic Radiology
Abstract:It is becoming increasingly important to develop imaging biomarkers to study Parkinson’s disease (PD) and movement disorders. A number of key new biomarkers in magnetic resonance imaging (MRI) include measuring proton spin density (PSD), T1, iron content and volumes of the deep gray matter and neuromelanin. A new rapid, multi-contrast, multi-parametric quantitative MRI method referred to as strategically acquired gradient echo (STAGE) imaging offers the potential to extract all these biomarkers and provide a reproducible standardization for neuroimaging. In this work, we validate a comprehensive 3T MRI protocol dedicated to imaging PD and demonstrate that STAGE can be used across sites and across platforms to obtain the same quantitative data. More specifically, STAGE collects two different multi-echo 3D gradient-echo full brain data sets using two different flip angles. To test the robustness and reproducibility of the method the following scans were performed using the STAGE protocol: first, a single individual served as a traveling human brain and was imaged on 10 scanners at 9 sites; and second, 11 individuals were scanned 5 times each at the 9 different sites on 10 scanners to evaluate intrasite test–retest repeatability. The tissue properties PSD, T1, R2*, susceptibility and neuromelanin volume were quantified for each site. For the subject scanned at 9 different sites the errors were on the order of 4–6% for PSD, 2–6% for T1, 3–8% for R2* and 1–12 ppb for susceptibility depending on the size of the objects being measured. For the average across all individuals scanned five times, the error of the means were on the order of 2–3% for PSD, 2–3% for T1, 3–5% for R2* and 1–10 ppb for susceptibility values depending on the size of the objects being measured. For those scanned five times at the same sitting, the errors were often only half the maximum values stated above. The mean/standard deviation of all subjects’ neuromelanin volume was 273/17 mm3 and 263/35 mm3 for the left and right sides, while the contrast was 26/7% and 26/6% for the left and right sides, respectively. In conclusion, the ability to collect data across manufacturers at different hospitals using STAGE provides an impetus for broader movement disorder studies using MRI. This standardized approach offers the potential to increase the number of patients studied for a given disease dramatically and to better test imaging biomarker hypotheses.