Within-study repeated measurements to increase sensitivity for positron emission tomography activation studies

Area of Science:

• Neuroimaging
• Nuclear Medicine

Background:

• Positron emission tomography (PET) activation studies require robust data collection and analysis strategies.
• Accurate statistical analysis is crucial for detecting subtle physiological changes in neuroimaging.

Purpose of the Study:

• To evaluate a novel data collection strategy for PET activation studies.
• To improve the statistical power and feasibility of single-subject PET analyses.

Main Methods:

• A new data collection scheme using H2(15)O injection and multiple 10-second frames was proposed.
• Blocked analysis of variance (ANOVA) with blocking across frames was employed.
• The method was validated using phantom data and applied to human subjects in a visceral sensation study.

Main Results:

• The novel strategy increased statistical significance by obtaining residual error estimates from more independent measurements.
• Phantom data demonstrated higher significance with increased frame division.
• Single-subject analysis successfully detected changes as small as 10 mL/(min x 100 mL) in visceral sensation studies, aligning with group analysis.
• Statistical assumptions of independence, normality, and homoscedasticity were not violated.

Conclusions:

• The proposed data collection and analysis method enhances statistical significance in PET activation studies.
• This approach facilitates reliable single-subject analysis, crucial for personalized medicine and detailed physiological investigations.
• Further optimization may be needed to improve specificity in real-world experimental conditions.