Pseudo low-energy monochromatic imaging of head and neck cancers: Deep learning image reconstruction with dual-energy CT
Pseudo low-energy monochromatic imaging of head and neck cancers: Deep learning image reconstruction with dual-energy CT
复制标题
DOI：
10.1007/s11548-022-02627-x
复制
发表时间：
2022-04-12
影响因子：
3
通讯作者：
Ogawa, Kazuhiko
中科院分区：
文献类型：
--
作者：
Koike, Yuhei;Ohira, Shingo;Ogawa, Kazuhiko
目的低能量虚拟单色图像(VMI)来自双能量计算机断层扫描(DECT)系统，与单能量CT(SECT)相比，提高了头颈部肿瘤的病变清晰度。然而，DECT系统安装在数量有限的设施中；因此，只有少数设施受益于VMIS。在这项工作中，我们提出了一种深度学习(DL)体系结构，适用于为使用切片成像的设施生成头颈部癌症的伪低能量VMI。方法回顾分析115例头颈部肿瘤患者行增强扫描的临床资料。分别使用70keV和50keV的VMI作为输入和基础真实(GT)。我们将它们分成两个数据集：DL(104例患者)和SECT推断(11例患者)。我们比较了四种DL架构：U-Net、基于DenseNet和两种基于ResNet的模型。通过Hounsfield单位(HU)值的平均绝对误差(MAE)、峰值信噪比(PSNR)和结构相似性(SSIM)，比较了50keV(pVMI(50keV))下的伪VMI和GT。评估肿瘤、血管、腮腺、肌肉、脂肪和骨骼的HU值。PVMI(50keV)是从实际切片图像中产生的，并评估了HU值。结果U-net的MAE最低(13.32+/-2.20HU)，PSNR最高(47.03+/-2.33dB)，SSIM(0.9965+/-0.0009)，差异有统计学意义(P<；0.001)。HU评估表明GT和U-Net之间有很好的一致性。U-Net对肿瘤的绝对HU差值最小，为5.0HU。结论物理参数的定量比较表明，与已有的DL结构相比，所提出的U-net能够在更短的时间内产生高精度的pVMI(50keV)。虽然还需要进一步评估诊断的准确性，但我们的方法可以帮助从没有DECT系统的断层图像中获得低能量的VMI。
Purpose Low-energy virtual monochromatic images (VMIs) derived from dual-energy computed tomography (DECT) systems improve lesion conspicuity of head and neck cancer over single-energy CT (SECT). However, DECT systems are installed in a limited number of facilities; thus, only a few facilities benefit from VMIs. In this work, we present a deep learning (DL) architecture suitable for generating pseudo low-energy VMIs of head and neck cancers for facilities that employ SECT imaging. Methods We retrospectively analyzed 115 patients with head and neck cancers who underwent contrast enhanced DECT. VMIs at 70 and 50 keV were used as the input and ground truth (GT), respectively. We divided them into two datasets: for DL (104 patients) and for inference with SECT (11 patients). We compared four DL architectures: U-Net, DenseNet-based, and two ResNet-based models. Pseudo VMIs at 50 keV (pVMI(50keV)) were compared with the GT in terms of the mean absolute error (MAE) of Hounsfield unit (HU) values, peak signal-to-noise ratio (PSNR), and structural similarity (SSIM). The HU values for tumors, vessels, parotid glands, muscle, fat, and bone were evaluated. pVMI(50keV) were generated from actual SECT images and the HU values were evaluated. Results U-Net produced the lowest MAE (13.32 +/- 2.20 HU) and highest PSNR (47.03 +/- 2.33 dB) and SSIM (0.9965 +/- 0.0009), with statistically significant differences (P < 0.001). The HU evaluation showed good agreement between the GT and U-Net. U-Net produced the smallest absolute HU difference for the tumor, at < 5.0 HU. Conclusion Quantitative comparisons of physical parameters demonstrated that the proposed U-Net could generate high accuracy pVMI(50keV) in a shorter time compared with the established DL architectures. Although further evaluation on diagnostic accuracy is required, our method can help obtain low-energy VMI from SECT images without DECT systems.