Hippocampal Magnetization Transfer Ratio at 3T: Validation of Automated Postprocessing and Comparison of Quantification Metrics

• Published in: Journal of neuroimaging 2013-07, Vol.23 (3), p.484-490
• Main Authors: Sidharthan, Shawn, Hutten, Ryan, Glielmi, Christopher, Du, Hongyan, Malone, Fiona, Ragin, Ann B., Edelman, Robert R., Wu, Ying
• Format: Article
• Language: English
• Subjects: Adolescent; Adult; Aged; Algorithms; Automation; Female; Hippocampus - anatomy & histology; histogram analysis; Humans; Image Enhancement - methods; Image Interpretation, Computer-Assisted - methods; Magnetic Resonance Imaging - methods; Magnetization transfer imaging; magnetization transfer ratio; Male; Middle Aged; Neuroimaging; NMR; normal-control analysis; Nuclear magnetic resonance; Pattern Recognition, Automated - methods; reproducibility; Reproducibility of Results; Sensitivity and Specificity; Studies; Young Adult
• ISSN: 1051-2284; 1552-6569
• DOI: 10.1111/j.1552-6569.2011.00697.x

ABSTRACT BACKGROUND To investigate the reliability of a novel magnetization transfer ratio (MTR) postprocessing technique for the hippocampus using histogram analysis, and compare results to more established volumetric measurements. This study is conducted in healthy volunteers as a precursor to future applications regarding progressive neurologic diseases, such as Alzheimer's disease. METHODS Eight healthy subjects were scanned twice with interval of 1 week using quantitative magnetic resonance imaging (MRI). Automated pixel‐wise analysis was performed for the hippocampal regions of each patient. Reliability was assessed using intraclass correlation coefficients (ICCs), coefficients of variation (COVs), and instrumental standard deviation (ISD). RESULTS Reliable metrics were 25th percentile, median, 75th percentile, peak location, and mean approach (ranges: ICC = .93‐.96, COV = 2.71‐3.88%, ISD .78‐1.01). Histogram peak height had ICC below .7, and a COV above 10%. Volumetric measurements had (ICC = .95‐.97, COV = 1.43‐3.39%). CONCLUSION Excellent scan‐rescan reproducibility (ICC > .9, COV < 10%) was observed for specific MTR histogram metrics and the mean MTR approach. These results are comparable to the volumetric approach. Future studies can examine the possibility that MTR changes precede morphological changes as this study suggests that both MTR and volumetric measurements of the hippocampus can be used as reliable imaging tools.