Reliability of the freehand region-of-interest method in quantitative cerebral diffusion tensor imaging

Area of Science:

• Neuroimaging
• Biomedical Engineering
• Radiology

Background:

• Diffusion Tensor Imaging (DTI) is a key MRI technique for assessing white matter integrity in the brain.
• Quantitative DTI analysis reliability depends on imaging protocols, processing methods, and diffusion parameters.
• The Region-of-Interest (ROI) method is a common post-processing technique due to its availability in commercial software.

Purpose of the Study:

• To evaluate the reliability of the freehand ROI method in DTI analysis.
• To conduct intra- and inter-observer analyses to determine the consistency of measurements.
• To assess the impact of age on DTI parameters in specific brain regions.

Main Methods:

• 40 healthy participants underwent 3T brain MRI.
• Freehand ROIs were placed in nine anatomical locations.
• DTI parameters analyzed included fractional anisotropy (FA), apparent diffusion coefficient (ADC), axial diffusivity (AD), and radial diffusivity (RD).
• Intra- and inter-observer variability and reliability were assessed using intraclass correlation coefficients (ICCs).

Main Results:

• Average intra-observer ICC was 0.9 (excellent).
• Single ICCs for FA and ADC were excellent or adequate in 8/9 regions.
• Frontobasal regions demonstrated the highest reliability.
• Younger adults (18-30 years) showed significantly different FA, AD, and RD values compared to older groups.

Conclusions:

• The freehand ROI method offers high average reliability and adequate single-measure reliability for DTI analysis.
• The method is suitable for research with experienced observers, recommending at least two measurements per region.
• Artifacts can impact reliability in certain regions (e.g., cerebral peduncle).
• Results are most clinically relevant in adults under 30 due to minimal age-related white matter changes.