Vitale Santiago, Orlando José Ignacio, Iarussi Emmanuel, Díaz Alejandro, Larrabide Ignacio
National Scientific and Technical Research Council (CONICET), Buenos Aires, Argentina.
Pladema Institute, UNICEN, Tandil, Buenos Aires, Argentina.
Med Phys. 2025 Jun;52(6):4540-4556. doi: 10.1002/mp.17801. Epub 2025 Mar 30.
Ultrasound (US) simulation helps train physicians and medical students in image acquisition and interpretation, enabling safe practice of transducer manipulation and organ identification. Current simulators generate realistic images from reference scans. Although physics-based simulators provide real-time images, they lack sufficient realism, while recent deep learning-based models based on unpaired image-to-image translation improve realism but introduce anatomical inconsistencies.
We propose a novel framework to reduce hallucinations from generative adversarial networks (GANs) used on physics-based simulations, enhancing anatomical accuracy and realism in abdominal US simulation. Our method aims to produce anatomically consistent images free from artifacts within and outside the field of view (FoV).
We introduce a segmentation-guided loss to enforce anatomical consistency by using a pre-trained Unet model that segments abdominal organs from physics-based simulated scans. Penalizing segmentation discrepancies before and after the translation cycle helps prevent unrealistic artifacts. Additionally, we propose training GANs on images in polar coordinates to limit the field of view to non-blank regions. We evaluated our approach on unpaired datasets comprising 617 real abdominal US images from a SonoSite-M turbo v1.3 scanner and 971 artificial scans from a ray-casting simulator. Data was partitioned at the patient level into training (70%), validation (10%), and testing (20%). Performance was quantitatively assessed with Frechet and Kernel Inception Distances (FID and KID), and organ-specific histogram distances, reporting 95% confidence intervals. We compared our model against generative methods such as CUT, UVCGANv2, and UNSB, performing statistical analyses using Wilcoxon tests (FID and KID with Bonferroni-corrected , with ). A perceptual realism study involving expert radiologists was also conducted.
Our method significantly reduced FID and KID by 66% and 89%, respectively, compared to CycleGAN, and by 34% and 59% compared to the leading alternative UVCGANv2 ( ). No significant differences ( ) in echogenicity distributions were found between real and simulated images within liver and gallbladder regions. The user study indicated our simulated scans fooled radiologists in 36.2% of cases, outperforming other methods.
Our segmentation-guided, polar-coordinates-trained CycleGAN framework significantly reduces hallucinations, ensuring anatomical consistency, and realism in simulated abdominal US images, surpassing existing methods.
超声（US）模拟有助于培训医生和医学生进行图像采集和解读，使他们能够安全地进行换能器操作和器官识别练习。当前的模拟器通过参考扫描生成逼真的图像。尽管基于物理的模拟器能够提供实时图像，但它们缺乏足够的真实感，而最近基于深度学习的基于无配对图像到图像转换的模型提高了真实感，但却引入了解剖结构不一致的问题。
我们提出了一种新颖的框架，以减少基于物理模拟的生成对抗网络（GAN）产生的幻觉，提高腹部超声模拟中的解剖准确性和真实感。我们的方法旨在生成在视野（FoV）内外均无伪影的解剖结构一致的图像。
我们引入了一种分割引导损失，通过使用预训练的Unet模型来强制解剖结构的一致性，该模型可从基于物理的模拟扫描中分割出腹部器官。对转换循环前后的分割差异进行惩罚有助于防止出现不真实的伪影。此外，我们建议在极坐标图像上训练GAN，将视野限制在非空白区域。我们在由来自SonoSite-M turbo v1.3扫描仪的617张真实腹部超声图像和来自光线投射模拟器的971张人工扫描图像组成的无配对数据集上评估了我们的方法。数据在患者层面被划分为训练集（70%）、验证集（10%）和测试集（20%）。使用弗雷歇距离和核内插距离（FID和KID）以及特定器官的直方图距离对性能进行定量评估，并报告95%置信区间。我们将我们的模型与诸如CUT、UVCGANv2和UNSB等生成方法进行比较，使用威尔科克森检验进行统计分析（FID和KID采用邦费罗尼校正， ， ）。还进行了一项涉及专家放射科医生的感知真实感研究。
与CycleGAN相比，我们的方法分别显著降低了FID和KID的66%和89%，与领先的替代方法UVCGANv2相比分别降低了34%和59%（ ）。在肝脏和胆囊区域的真实图像和模拟图像之间，未发现回声性分布存在显著差异（ ）。用户研究表明，我们的模拟扫描在36.2%的情况下能够骗过放射科医生，优于其他方法。
我们的分割引导、极坐标训练的CycleGAN框架显著减少了幻觉，确保了模拟腹部超声图像中的解剖一致性和真实感，超越了现有方法。
