An investigation of quantitative accuracy for deep learning based denoising in oncological PET
An investigation of quantitative accuracy for deep learning based denoising in oncological PET
复制标题
DOI：
10.1088/1361-6560/ab3242
复制
发表时间：
2019-08-01
影响因子：
3.5
通讯作者：
Liu, Chi
中科院分区：
文献类型：
--
作者：
Lu, Wenzhuo;Onofrey, John A.;Liu, Chi
降低辐射剂量对PET成像具有重要意义。然而，减少注射剂量会导致图像噪声增加和信噪比(SNR)降低，从而影响诊断和定量准确性。深度学习方法在降低低剂量PET数据噪声和提高信噪比方面显示出了巨大的潜力。在这项工作中，我们全面研究了基于深度学习的肿瘤PET图像去噪方法，除了视觉图像质量外，还对肺小结节的定量准确性进行了研究。我们应用并优化了一种基于U-Net结构的高级深度学习方法，用于从10%的低剂量PET数据中预测标准剂量PET图像。我们还考察了不同的网络结构、图像维度、标签和输入对深度学习方法的降噪性能和定量精度的影响。用不同感兴趣区的归一化均方误差(NMSE)、信噪比(SNR)和标准摄取值(SUV)偏差进行评估，结果表明U-Net和GaN优于CAE，其SUV均值和SUVmax偏差较小，但SNR较低。与2D和2.5D U网络相比，全3D U网络具有最佳的定量性能，对于所有10名患者而言，SUV平均偏差小于15%。U-Net总体上优于剩余U-Net(r-U-Net)，具有更小的NMSE、更高的SNR和更低的SUVmax偏差。全三维U-Net在图像质量和噪声与偏差的折衷方面优于现有的几种去噪方法，包括高斯滤波、解剖引导的非局部均值(NLM)滤波、基于二次先验和相对差先验的MAP重建。此外，融合对准的CT图像有可能进一步提高多通道U网络的定量精度，我们发现基于深度学习的方法在绝对定量精度和视觉图像质量方面的最佳结构和参数是不同的。我们的定量结果表明，在使用10%的低计数下采样数据生成标准剂量的PET时，全3D U-net可以有效地降低图像噪声和控制偏差，即使对于亚厘米小结节也是如此。
Reducing radiation dose is important for PET imaging. However, reducing injection doses causes increased image noise and low signal-to-noise ratio (SNR), subsequently affecting diagnostic and quantitative accuracies. Deep learning methods have shown a great potential to reduce the noise and improve the SNR in low dose PET data.In this work, we comprehensively investigated the quantitative accuracy of small lung nodules, in addition to visual image quality, using deep learning based denoising methods for oncological PET imaging. We applied and optimized an advanced deep learning method based on the U-net architecture to predict the standard dose PET image from 10% low-dose PET data. We also investigated the effect of different network architectures, image dimensions, labels and inputs for deep learning methods with respect to both noise reduction performance and quantitative accuracy. Normalized mean square error (NMSE), SNR, and standard uptake value (SUV) bias of different nodule regions of interest (ROIs) were used for evaluation.Our results showed that U-net and GAN are superior to CAE with smaller SUVmean and SUVmax bias at the expense of inferior SNR. A fully 3D U-net has optimal quantitative performance compared to 2D and 2.5D U-net with less than 15% SUVmean bias for all the ten patients. U-net outperforms Residual U-net (r-U-net) in general with smaller NMSE, higher SNR and lower SUVmax bias. Fully 3D U-net is superior to several existing denoising methods, including Gaussian filter, anatomical-guided non-local mean (NLM) filter, and MAP reconstruction with Quadratic prior and relative difference prior, in terms of superior image quality and trade-off between noise and bias. Furthermore, incorporating aligned CT images has the potential to further improve the quantitative accuracy in multi-channel U-net.We found the optimal architectures and parameters of deep learning based methods are different for absolute quantitative accuracy and visual image quality. Our quantitative results demonstrated that fully 3D U-net can both effectively reduce image noise and control bias even for sub-centimeter small lung nodules when generating standard dose PET using 10% low count down-sampled data.