Division of Nuclear Medicine, Department of Radiology, University of Massachusetts Chan Medical School, Worcester, MA, USA.
Electrical and Computer Engineering, Illinois Institute of Technology, Chicago, IL, USA.
J Nucl Cardiol. 2023 Dec;30(6):2427-2437. doi: 10.1007/s12350-023-03295-3. Epub 2023 May 23.
The aim of this research was to asses perfusion-defect detection-accuracy by human observers as a function of reduced-counts for 3D Gaussian post-reconstruction filtering vs deep learning (DL) denoising to determine if there was improved performance with DL.
SPECT projection data of 156 normally interpreted patients were used for these studies. Half were altered to include hybrid perfusion defects with defect presence and location known. Ordered-subset expectation-maximization (OSEM) reconstruction was employed with the optional correction of attenuation (AC) and scatter (SC) in addition to distance-dependent resolution (RC). Count levels varied from full-counts (100%) to 6.25% of full-counts. The denoising strategies were previously optimized for defect detection using total perfusion deficit (TPD). Four medical physicist (PhD) and six physician (MD) observers rated the slices using a graphical user interface. Observer ratings were analyzed using the LABMRMC multi-reader, multi-case receiver-operating-characteristic (ROC) software to calculate and compare statistically the area-under-the-ROC-curves (AUCs).
For the same count-level no statistically significant increase in AUCs for DL over Gaussian denoising was determined when counts were reduced to either the 25% or 12.5% of full-counts. The average AUC for full-count OSEM with solely RC and Gaussian filtering was lower than for the strategies with AC and SC, except for a reduction to 6.25% of full-counts, thus verifying the utility of employing AC and SC with RC.
We did not find any indication that at the dose levels investigated and with the DL network employed, that DL denoising was superior in AUC to optimized 3D post-reconstruction Gaussian filtering.
本研究旨在评估 3D 高斯后重建滤波与深度学习（DL）去噪的减少计数对人类观察者灌注缺陷检测准确性的影响，以确定 DL 是否能提高性能。
这些研究使用了 156 名正常解释的患者的 SPECT 投影数据。一半的数据被改变，包括具有已知存在和位置的混合灌注缺陷。除了距离相关的分辨率（RC）外，还使用有序子集期望最大化（OSEM）重建来进行可选的衰减（AC）和散射（SC）校正。计数水平从全计数（100%）到全计数的 6.25%不等。去噪策略是使用总灌注缺陷（TPD）针对缺陷检测进行了预先优化。四位医学物理学家（博士）和六位医师（MD）观察者使用图形用户界面对切片进行评分。使用 LABMRMC 多读者、多病例接收器操作特征（ROC）软件分析观察者评分，以计算和比较 ROC 曲线下的面积（AUC）的统计数据。
在相同的计数水平下，当计数减少到全计数的 25%或 12.5%时，与高斯去噪相比，DL 没有在 AUC 上表现出统计学上的显著增加。仅使用 RC 和高斯滤波的全计数 OSEM 的平均 AUC 低于使用 AC 和 SC 的策略，除了减少到全计数的 6.25%，这验证了使用 AC 和 SC 与 RC 的效用。
我们没有发现任何迹象表明，在所研究的剂量水平和所使用的 DL 网络下，DL 去噪在 AUC 方面优于经过优化的 3D 后重建高斯滤波。
