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

BMC Medical Imaging

Ullamari Hakulinen¹  Juha Öhman²  Hannu Eskola³  Antti Brander³  Teemu M. Luoto⁴  Tero Ilvesmäki⁴  Mika Helminen⁵ 

[1] Department of Medical Physics, Medical Imaging Center of Pirkanmaa Hospital District;Department of Neurosurgery, Tampere University Hospital and Tampere University;Department of Radiology, Medical Imaging Center of Pirkanmaa Hospital District;Faculty of Medicine and Health Technology, Tampere University;Faculty of Social Sciences, Health Sciences, Tampere University;

关键词: Diffusion tensor imaging;    Reliability;    Repeatability;    Intra-class correlation coefficient;    Fractional anisotropy;    Apparent diffusion coefficient;   

DOI  :  10.1186/s12880-021-00663-8

来源: DOAJ

【 摘 要 】

Abstract Background Diffusion tensor imaging (DTI) is a magnetic resonance imaging (MRI) technique used for evaluating changes in the white matter in brain parenchyma. The reliability of quantitative DTI analysis is influenced by several factors, such as the imaging protocol, pre-processing and post-processing methods, and selected diffusion parameters. The region-of-interest (ROI) method is most widely used of the post-processing methods because it is found in commercial software. The focus of our research was to study the reliability of the freehand ROI method using various intra- and inter-observer analyses. Methods This study included 40 neurologically healthy participants who underwent diffusion MRI of the brain with a 3 T scanner. The measurements were performed at nine different anatomical locations using a freehand ROI method. The data extracted from the ROIs included the regional mean values, intra- and inter-observer variability and reliability. The used DTI parameters were fractional anisotropy (FA), the apparent diffusion coefficient (ADC), and axial (AD) and radial (RD) diffusivity. Results The average intra-class correlation coefficient (ICC) of the intra-observer was found to be 0.9 (excellent). The single ICC results were excellent (> 0.8) or adequate (> 0.69) in eight out of the nine regions in terms of FA and ADC. The most reliable results were found in the frontobasal regions. Significant differences between age groups were also found in the frontobasal regions. Specifically, the FA and AD values were significantly higher and the RD values lower in the youngest age group (18–30 years) compared to the other age groups. Conclusions The quantitative freehand ROI method can be considered highly reliable for the average ICC and mostly adequate for the single ICC. The freehand method is suitable for research work with a well-experienced observer. Measurements should be performed at least twice in the same region to ensure that the results are sufficiently reliable. In our study, reliability was slightly undermined by artifacts in some regions such as the cerebral peduncle and centrum semiovale. From a clinical point of view, the results are most reliable in adults under the age of 30, when age-related changes in brain white matter have not yet occurred.

【 授权许可】

Unknown