| Summary: | 碩士 === 義守大學 === 資訊工程學系碩士在職專班 === 101 === Background: 99mTc Trodat-1 has been widely used to image dopamine system. Many quantitative methods were developed in recent. However, clinical diagnoses are usually through visual interpretation by physician.
Purpose: In this study, the main aim is to find important and significant image features with correlated to clinical diagnosis. Those features can play an important role with future interpretation.
Material and method: Total of 79 cases were retrospectively reviewed in this study in which 59 were enrolled as effective sample. The clinical diagnostic category (Essential tremor, Parkinson’s disease and other diagnosis), gender, age, and 99mTc Trodat-1SPECT were collected for analysis. Six features of image were design as the heist striatal activity (maximum intensity of slice, Max Target), skewness of target slice histogram (Skewness Target), skewness of background brain (Skewness Reference), the ratio of skewness Target and Skewness Reference (Skew Ratio), interaction of Skewness Target and Skewness Reference (Skewness Cross) and product of Skewness Target and Skewness Reference (Skewness Product). The analysis of variance (ANOVA) was applied to test significant difference of image features among categories. The generalized linear model (GLM) is used to examine the correlation of image features with diagnostic groups, patient gender and age. P-value <0.05 is set as attained statistical significance.
Results: The Skewness Cross and Skewness Product were significant difference among clinical diagnosis groups by ANOVA and GLM. The mean and standard variation of Skewness Cross of the three diagnosis groups (Essential tremor, Parkinson's disease, Other diagnosis) are 1.706±-.145, .540±.225, .726±.508 and the mean and standard variation of Skewness Product of the three diagnosis groups are -.395±-1.106, .031±-.085, -.024±-.111. The mean and standard variation of the two image characteristics are not influenced by gender and age.
Conclusion: The Skewness Cross and Skewness Product in 99mTc Trodat-1 image histogram are important and significant features among groups. The higher value of Skewness Cross and the lower value of Skewness Product were shown, the more likely essential tremor was presented. Increasing effective sample size and applied quantitative analysis of the image features will get reliable and useful information in future work.
|
|---|
