3D Texture Analyses Within the Substantia Nigra of Parkinson's Disease Patients on Quantitative Susceptibility Maps and R2∗ Maps.

Gaiying Li,Guoqiang Zhai,Xinxin Zhao,Hedi An,Pascal Spincemaille,Kelly M. Gillen,Yixuan Ku,Yi Wang,Dongya Huang,Jianqi Li
DOI: https://doi.org/10.1016/j.neuroimage.2018.12.041
IF: 5.7
2018-01-01
NeuroImage
Abstract:Iron accumulation in the substantia nigra (SN) is spatially heterogeneous, yet no study has quantitatively evaluated how the texture of quantitative susceptibility maps (QSM) and R2∗ might evolve with Parkinson's disease (PD) and healthy controls (HC). The aim of this study was to discriminate between patients with PD and HC using texture analysis in the SN from QSM and R2∗ maps. QSM and R2∗ maps were obtained from 28 PD patients and 28 HC on a clinical 3T MR imaging scanner using 3D multi-echo gradient-echo sequence. The first- and second- order texture features of the QSM and R2∗ images were obtained to evaluate group differences using two-tailed t-test. After correction for multiple comparisons, for the first-order analysis, the susceptibility of SN from patients with PD was significantly greater (p = 0.017) compared with the SN from HC. For the second-order texture analysis, angular second moment, entropy, and sum of entropy showed significant differences in QSM (p < 0.001) and R2∗ maps (p < 0.01). In addition, correlation, contrast, sum of variance and difference of variance, significantly separated the subject groups in QSM maps (p < 0.05) but not in R2∗ images. Receiver operating characteristic analysis showed that entropy and sum of entropy of the QSM maps in the SN yielded the highest performance for differentiating PD patients from HC (area under the curve = 0.89). In conclusion, most first- and second- order QSM texture features successfully distinguished PD patients from HC and significantly outperformed R2∗ texture analysis. The second-order texture features were more accurate and sensitive than first-order texture features for classifying PD patients.
What problem does this paper attempt to address?