Buchert Ralph, Lange Catharina, Spehl Timo S, Apostolova Ivayla, Frings Lars, Jonsson Cathrine, Meyer Philipp T, Hellwig Sabine
Department for Diagnostic and Interventional Radiology and Nuclear Medicine, University Hospital Hamburg-Eppendorf, Martinistr. 52, 20246, Hamburg, Germany.
Department of Nuclear Medicine, Charité - Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität zu Berlin, and Berlin Institute of Health, Augustenburger Platz 1, 13353, Berlin, Germany.
EJNMMI Res. 2019 May 7;9(1):37. doi: 10.1186/s13550-019-0506-9.
The specific uptake size index (SUSI) of striatal FP-CIT uptake is independent of spatial resolution in the SPECT image, in contrast to the specific binding ratio (SBR). This suggests that the SUSI is particularly appropriate for multi-site/multi-camera settings in which camera-specific effects increase inter-subject variability of spatial resolution. However, the SUSI is sensitive to inter-subject variability of striatum size. Furthermore, it might be more sensitive to errors of the estimate of non-displaceable FP-CIT binding. This study compared SUSI and SBR in the multi-site/multi-camera (MULTI) setting of a prospective multi-center study and in a mono-site/mono-camera (MONO) setting representative of clinical routine.
The MULTI setting included patients with Parkinson's disease (PD, n = 438) and healthy controls (n = 207) from the Parkinson Progression Marker Initiative. The MONO setting included 122 patients from routine clinical patient care in whom FP-CIT SPECT had been performed with the same double-head SPECT system according to the same acquisition and reconstruction protocol. Patients were categorized as "neurodegenerative" (n = 84) or "non-neurodegenerative" (n = 38) based on follow-up data. FP-CIT SPECTs were stereotactically normalized to MNI space. SUSI and SBR were computed for caudate, putamen, and whole striatum using unilateral ROIs predefined in MNI space. SUSI analysis was repeated in native patient space in the MONO setting. The area (AUC) under the ROC curve for identification of PD/"neurodegenerative" cases was used as performance measure.
In both settings, the highest AUC was achieved by the putamen (minimum over both hemispheres), independent of the semi-quantitative method (SUSI or SBR). The putaminal SUSI provided slightly better performance with ROI analysis in MNI space compared to patient space (AUC = 0.969 vs. 0.961, p = 0.129). The SUSI (computed in MNI space) performed slightly better than the SBR in the MULTI setting (AUC = 0.993 vs. 0.991, p = 0.207) and slightly worse in the MONO setting (AUC = 0.969 vs. AUC = 0.976, p = 0.259). There was a trend toward larger AUC difference between SUSI and SBR in the MULTI setting compared to the MONO setting (p = 0.073). Variability of voxel intensity in the reference region was larger in misclassified cases compared to correctly classified cases for both SUSI and SBR (MULTI setting: p = 0.007 and p = 0.012, respectively).
The SUSI is particularly useful in MULTI settings. SPECT images should be stereotactically normalized prior to SUSI analysis. The putaminal SUSI provides better diagnostic performance than the SUSI of the whole striatum. Errors of the estimate of non-displaceable count density in the reference region can cause misclassification by both SUSI and SBR, particularly in borderline cases. These cases might be identified by visual checking FP-CIT uptake in the reference region for particularly high variability.
与特异性结合率（SBR）不同，纹状体FP-CIT摄取的特异性摄取大小指数（SUSI）与SPECT图像中的空间分辨率无关。这表明SUSI特别适用于多部位/多相机设置，在这种设置中，相机特定效应会增加空间分辨率的受试者间变异性。然而，SUSI对纹状体大小的受试者间变异性敏感。此外，它可能对不可置换的FP-CIT结合估计误差更敏感。本研究在一项前瞻性多中心研究的多部位/多相机（MULTI）设置以及代表临床常规的单部位/单相机（MONO）设置中比较了SUSI和SBR。
MULTI设置包括帕金森病进展标志物倡议中的帕金森病（PD，n = 438）患者和健康对照（n = 207）。MONO设置包括122例来自常规临床患者护理的患者，他们根据相同的采集和重建方案，使用同一双头SPECT系统进行了FP-CIT SPECT检查。根据随访数据，将患者分为“神经退行性”（n = 84）或“非神经退行性”（n = 38）。FP-CIT SPECT通过立体定向归一化到MNI空间。使用MNI空间中预先定义的单侧感兴趣区（ROI）计算尾状核、壳核和整个纹状体的SUSI和SBR。在MONO设置中，在患者原始空间中重复进行SUSI分析。用于识别PD/“神经退行性”病例的ROC曲线下面积（AUC）用作性能指标。
在两种设置中，壳核的AUC最高（两个半球中的最小值），与半定量方法（SUSI或SBR）无关。与患者空间相比，在MNI空间中进行ROI分析时，壳核SUSI的性能略好（AUC = 0.969对0.961，p = 0.129）。在MULTI设置中，SUSI（在MNI空间中计算）的性能略优于SBR（AUC = 0.993对0.991，p = 0.207），而在MONO设置中略差（AUC = 0.969对AUC = 0.976，p = 0.259）。与MONO设置相比，MULTI设置中SUSI和SBR之间的AUC差异有增大趋势（p = 0.073）。对于SUSI和SBR，误分类病例中参考区域内体素强度的变异性均大于正确分类病例（MULTI设置：分别为p = 0.007和p = 0.012）。
SUSI在MULTI设置中特别有用。在进行SUSI分析之前，SPECT图像应进行立体定向归一化。壳核SUSI比整个纹状体的SUSI具有更好的诊断性能。参考区域中不可置换计数密度估计误差可导致SUSI和SBR出现误分类，尤其是在临界病例中。这些病例可通过目视检查参考区域中FP-CIT摄取的变异性是否特别高来识别。
