From the Departments of Diagnostic and Interventional Radiology (D.U., A.Y., S.L.W., H. Tatekawa, H. Takita, T.H., A.S., Y.M.), Neurosurgery (T. Ichinose, H.A., Y.W., T.G.), and Medical Statistics (D.K.), Graduate School of Medicine, Osaka City University, 1-4-3 Asahi-machi, Abeno-ku, Osaka 545-8585, Japan; and Department of Radiology, Osaka City University Hospital, 1-5-7 Asahi-machi, Abeno-ku, Osaka, 545-8586, Japan (Y.K., T. Ichida).
Radiology. 2021 Jun;299(3):675-681. doi: 10.1148/radiol.2021203692. Epub 2021 Mar 30.
Background Digital subtraction angiography (DSA) generates an image by subtracting a mask image from a dynamic angiogram. However, patient movement-caused misregistration artifacts can result in unclear DSA images that interrupt procedures. Purpose To train and to validate a deep learning (DL)-based model to produce DSA-like cerebral angiograms directly from dynamic angiograms and then quantitatively and visually evaluate these angiograms for clinical usefulness. Materials and Methods A retrospective model development and validation study was conducted on dynamic and DSA image pairs consecutively collected from January 2019 through April 2019. Angiograms showing misregistration were first separated per patient by two radiologists and sorted into the misregistration test data set. Nonmisregistration angiograms were divided into development and external test data sets at a ratio of 8:1 per patient. The development data set was divided into training and validation data sets at ratio of 3:1 per patient. The DL model was created by using the training data set, tuned with the validation data set, and then evaluated quantitatively with the external test data set and visually with the misregistration test data set. Quantitative evaluations used the peak signal-to-noise ratio (PSNR) and the structural similarity (SSIM) with mixed liner models. Visual evaluation was conducted by using a numerical rating scale. Results The training, validation, nonmisregistration test, and misregistration test data sets included 10 751, 2784, 1346, and 711 paired images collected from 40 patients (mean age, 62 years ± 11 [standard deviation]; 33 women). In the quantitative evaluation, DL-generated angiograms showed a mean PSNR value of 40.2 dB ± 4.05 and a mean SSIM value of 0.97 ± 0.02, indicating high coincidence with the paired DSA images. In the visual evaluation, the median ratings of the DL-generated angiograms were similar to or better than those of the original DSA images for all 24 sequences. Conclusion The deep learning-based model provided clinically useful cerebral angiograms free from clinically significant artifacts directly from dynamic angiograms. Published under a CC BY 4.0 license.
背景 数字减影血管造影（DSA）通过从动态血管造影中减去掩模图像来生成图像。然而，由于患者运动引起的配准伪影可能导致 DSA 图像不清晰，从而中断手术。
目的 训练和验证一种基于深度学习（DL）的模型，该模型可以直接从动态血管造影中生成 DSA 样脑血管造影，并对这些血管造影进行定量和直观评估，以确定其临床应用价值。
材料和方法 本研究回顾性地对 2019 年 1 月至 2019 年 4 月连续采集的动态和 DSA 图像对进行了模型开发和验证研究。两位放射科医生首先根据每位患者的情况将显示配准错误的血管造影图分开，并将其归类为配准测试数据集。将无配准的血管造影图按每位患者 8:1 的比例分为开发和外部测试数据集。DL 模型通过使用训练数据集进行创建，使用验证数据集进行调整，然后使用外部测试数据集进行定量评估，使用配准测试数据集进行视觉评估。定量评估使用混合线性模型的峰值信噪比（PSNR）和结构相似性（SSIM）。视觉评估使用数字评分量表进行。
结果 训练、验证、非配准测试和配准测试数据集分别包含 40 名患者（平均年龄，62 岁±11[标准差]；33 名女性）的 10751 对、2784 对、1346 对和 711 对图像。在定量评估中，DL 生成的血管造影图的平均 PSNR 值为 40.2 dB±4.05，平均 SSIM 值为 0.97±0.02，与配对的 DSA 图像高度吻合。在视觉评估中，对于所有 24 个序列，DL 生成的血管造影图的中位数评分与原始 DSA 图像相似或优于原始 DSA 图像。
结论 该基于深度学习的模型可以直接从动态血管造影中生成无明显临床伪影的临床有用的脑血管造影图。本研究已在知识共享署名 4.0 许可下发布。
