Biomarkers.

Area of Science:

• Neuroimaging
• Medical Physics
• Radiology

Background:

• Accurate quantification of brain perfusion SPECT scans relies on control databases.
• Establishing local databases is time-consuming, necessitating efficient alternatives.
• Commercial databases offer a potential solution for centers lacking resources for local data collection.

Purpose of the Study:

• To investigate differences in brain region quantification using commercial versus local control databases for SPECT scans.
• To identify specific brain regions most affected by the choice of control database.

Main Methods:

• Processed 35 SPECT scans using MIMNeuro software with both a local (31 controls) and a commercial (91 controls) database.
• Compared z-scores for key neurodegeneration-associated brain regions between the two databases.
• Utilized two-sample t-tests to assess region significance and applied thresholds (1.5 and 2.5 STDs) for disease severity.

Main Results:

• Stronger correlation in quantification was observed on the right side of the brain compared to the left.
• Regions like the Temporals and Precuneus showed significant z-score differences (>20%) between databases, potentially altering disease classification.
• The Precuneus indicated higher disease levels with the commercial database, while the Temporals showed the opposite.

Conclusions:

• The choice of control database (commercial vs. local) can significantly impact brain perfusion SPECT scan quantification and diagnostic interpretation.
• While local databases may offer greater population representation, commercial databases are a practical alternative for resource-limited centers.
• Further studies correlating with disease progression are needed to determine the most accurate quantification method.