QSM vs dopamine transporter PET: Mapping Shows Stronger Longitudinal Associations with Nigral Degeneration than Conventional QSM in Parkinson’s Disease

BACKGROUND:

Iron accumulation in the substantia nigra (SN) is a hallmark of Parkinson’s disease (PD). Quantitative susceptibility mapping (QSM) including source separation can generate maps of total susceptibility χ, paramagnetic or positive susceptibility ${\chi }^{+}$ likely dominated by tissue iron, and diamagnetic or negative susceptibility ${\chi }^{-}$ reflecting myelin and other sources.

PURPOSE:

To compare χ, ${\chi }^{+}$, and ${\chi }^{-}$ for tracking SN longitudinal changes in PD and for correlating with dopaminergic PET measures and cognitive decline.

MATERIALS AND METHODS:

This longitudinal study included 32 patients with PD (mean age, 65.88 ± 8.47 years) who underwent 3T MRI with 3D multi-echo gradient-echo (mGRE) and dynamic ¹¹C-N-(3-iodoprop-2E-enyl)-2β-carbomethoxy-3β-(4-methylphenyl) nortropane (PE2I) PET between August 2018 and April 2024. χ, ${\chi }^{+}$, and ${\chi }^{-}$ maps were reconstructed from mGRE data using the MEDI algorithm with susceptibility source separation. Dopamine transporter binding potential (BP) images were estimated from dynamic PET using a deep learning based kinetic model. Paired t tests, linear regression, and Pearson correlation analyses with Bonferroni correction were performed, with P < .05 considered significant.

RESULTS:

Both χ and ${\chi }^{+}$ from QSM revealed significant longitudinal susceptibility increases in the rostral and caudal SN (Bonferroni-corrected P < .05), with ${\chi }^{+}$ showing larger effect sizes (t = 9.99 vs 2.65 for the caudal SN). The left putaminal BP from PET significantly decreased at follow-up (P < .05). The longitudinal increase in ${\chi }^{+}$ within the rostral SN strongly correlated with the reduction in putaminal BP (r = – 0.60, P = .005, Bonferroni-corrected), whereas χ showed no significant association. Higher baseline ${\chi }^{+}$ values in the rostral and caudal SN predicted greater longitudinal cognitive decline (P = .01–.02, Bonferroni-corrected). Subregional analyses confirmed significant ${\chi }^{+}$ increases predominantly in the rostral-anterior SN. No significant change or correlation was observed with ${\chi }^{-}$ mapping at the SN.

CONCLUSION:

Positive susceptibility ${\chi }^{+}$ mapping showed larger longitudinal effect sizes for SN changes and stronger associations with dopaminergic PET decline and cognitive deterioration than to total susceptibility χ. These findings support ${\chi }^{+}$ mapping as a biologically informative MRI measure for the longitudinal assessment of SN changes in PD.