Validation of deep-learning-based MRI-to-CT attenuation correction for striatal and extrastriatal [¹²³I]I-FP-CIT...

Area of Science:

• Nuclear medicine
• Radiopharmaceutical imaging
• Neuroimaging

Background:

• Accurate attenuation correction is crucial for quantitative SPECT brain imaging.
• Quantifying dopamine and serotonin transporter binding is vital for diagnosing neurological and psychiatric disorders like Parkinson's disease.

Purpose of the Study:

• To validate deep-learning-based MRI to synthetic CT (DL-MRAC) for attenuation correction in [123I]I-FP-CIT SPECT.
• To quantify striatal and extrastriatal transporter binding in Parkinson's disease patients using DL-MRAC.

Main Methods:

• Generated synthetic CTs from T1-weighted MRIs in 12 Parkinson's disease patients using a 3D residual U-Net.
• Compared DL-MRAC with CT-based attenuation correction (CTAC), uniform correction (UAC), and no correction (NAC).
• Validated results using data from the Parkinson's Progression Markers Initiative (n=18).

Main Results:

• DL-MRAC showed minimal bias compared to CTAC (-0.4% striatal, -0.1% extrastriatal).
• UAC overestimated binding ratios (7.5%–12.4%), while NAC underestimated them (-6.5%–-24.0%).
• DL-MRAC demonstrated equivalence to CTAC for striatal and extrastriatal binding quantification.

Conclusions:

• DL-MRAC is a valid, radiation-free method for attenuation correction in DaTSPECT examinations.
• DL-MRAC accurately quantifies [123I]I-FP-CIT binding to dopamine and serotonin transporters.
• DL-MRAC offers a superior alternative to UAC and NAC, especially for MRI-available patients and SPECT-only cameras.