Impact of a combination of quantitative indices representing uptake intensity, shape, and asymmetry in DAT SPECT using machine learning: comparison of different volume of interest settings
Impact of a combination of quantitative indices representing uptake intensity, shape, and asymmetry in DAT SPECT using machine learning: comparison of different volume of interest settings
复制标题
DOI：
10.1186/s13550-019-0477-x
复制
发表时间：
2019-01-28
期刊：
影响因子：
3.2
通讯作者：
Jinzaki, Masahiro
中科院分区：
文献类型：
--
作者：
Iwabuchi, Yu;Nakahara, Tadaki;Jinzaki, Masahiro
背景：我们试图评估使用多巴胺转运蛋白单光子发射计算机断层扫描（DAT SPECT）获得的三种定量指标的基于机器学习的组合诊断准确性——特异性结合比（SBR）、壳核与尾状核比率（PCR）/分形维数（FD）和不对称指数（Al）——帕金森综合征（PS）。我们还旨在比较市售软件包 DaTQUANT (Q) 和 DaTView (V) 中两种不同类型的感兴趣体积 (VOl) 设置对诊断准确性的影响。 方法：纳入 71 名患有 PS 的患者和 40 名不患有 PS (NPS) 的患者。使用从这些患者获得的 SPECT 图像，分别在两种不同的 VOl 设置下计算三个定量指数。 SBR-Q、PCR-Q和Al-Q是使用来自DaTQUANT的VO1设置导出的，而SBR-V、FD-V和Al-V是使用来自DalView的VOl设置导出的。我们比较了这六项指标对PS的诊断价值。我们采用支持向量机（SVM）分类器来评估三个指标的组合精度（SVM-Q：SBR-Q、PCR-Q 和 Al-Q 的组合 SVM-V：SBR-V、FD-V 和 Al-V 的组合）。采用Mann-Whitney U检验和接受者操作特征（ROC）分析进行统计分析。结果：ROC分析表明，SBR-Q、PCR-Q、Al-Q SBR-V、FD-V和Al-V的曲线下面积（AUC）分别为0.978、0.837、0.802、0.906、0.972和0.829，分别。在比较两种类型的 VO1 设置之间的相应定量指数时，SBR-Q 的表现优于 SBR-V (p = 0.006)，而 FD-V 的表现优于 PCR-Q (p = 0.0003)。 Al-Q 和 Al-V 之间没有观察到显着差异 (p = 0.56)。 SVM-Q 和 SVM-V 的 AUC 分别为 0.988 和 0.994；两种不同的 VOl 设置在诊断准确性方面没有显着差异 (p = 0.48)。结论：使用 SVM 分类器获得的三个指数的组合提高了 PS 的诊断性能；该性能并没有因 VOl 设置和所使用的软件而异。
Background: We sought to assess the machine learning-based combined diagnostic accuracy of three types of quantitative indices obtained using dopamine transporter single-photon emission computed tomography (DAT SPECT) -specific binding ratio (SBR), putamen-to-caudate ratio (PCR)/fractal dimension (FD), and asymmetry index (Al)-for parkinsonian syndrome (PS). We also aimed to compare the effect of two different types of volume of interest (VOl) settings from commercially available software packages DaTQUANT (Q) and DaTView (V) on diagnostic accuracy.Methods: Seventy-one patients with PS and 40 without PS (NPS) were enrolled. Using SPECT images obtained from these patients, three quantitative indices were calculated at two different VOl settings each. SBR-Q, PCR-Q, and Al-Q were derived using the VOl settings from DaTQUANT, whereas SBR-V, FD-V, and Al-V were derived using those from DalView. We compared the diagnostic value of these six indices for PS. We incorporated a support vector machine (SVM) classifier for assessing the combined accuracy of the three indices (SVM-Q: combination of SBR-Q, PCR-Q, and Al-Q SVM-V: combination of SBR-V, FD-V, and Al-V). A Mann-Whitney U test and receiver-operating characteristics (ROC) analysis were used for statistical analyses.Results: ROC analyses demonstrated that the areas under the curve (AUC) for SBR-Q, PCR-Q, Al-Q SBR-V, FD-V, and Al-V were 0.978, 0.837, 0.802, 0.906, 0.972, and 0.829, respectively. On comparing the corresponding quantitative indices between the two types of VOl settings, SBR-Q performed better than SBR-V (p = 0.006), whereas FD-V performed better than PCR-Q (p = 0.0003). No significant difference was observed between Al-Q and Al-V (p = 0.56). The AUCs for SVM-Q and SVM-V were 0.988 and 0.994, respectively; the two different VOl settings displayed no significant differences in terms of diagnostic accuracy (p = 0.48).Conclusion: The combination of the three indices obtained using the SVM classifier improved the diagnostic performance for PS; this performance did not differ based on the VOl settings and software used.