PDF(Pubmed)
Abstract：
Microwave ablation (MWA) of liver tumors presents challenges like under- and over-ablation, potentially leading to inadequate tumor destruction and damage to healthy tissue. This study aims to develop personalized three-dimensional (3D) models to simulate MWA for liver tumors, incorporating patient-specific characteristics. The primary objective is to validate the predicted ablation zones compared to clinical outcomes, offering insights into MWA before therapy to facilitate accurate treatment planning. Contrast-enhanced CT images from three patients were used to create 3D models. The simulations used coupled electromagnetic wave propagation and bioheat transfer to estimate the temperature distribution, predicting tumor destruction and ablation margins. The findings indicate that prolonged ablation does not significantly improve tumor destruction once an adequate margin is achieved, although it increases tissue damage. There was a substantial overlap between the clinical ablation zones and the predicted ablation zones. For patient 1, the Dice score was 0.73, indicating high accuracy, with a sensitivity of 0.72 and a specificity of 0.76. For patient 2, the Dice score was 0.86, with a sensitivity of 0.79 and a specificity of 0.96. For patient 3, the Dice score was 0.8, with a sensitivity of 0.85 and a specificity of 0.74. Patient-specific 3D models demonstrate potential in accurately predicting ablation zones and optimizing MWA treatment strategies.
摘要：
肝肿瘤的微波消融（MWA）面临着诸如消融不足和过度消融的挑战，可能导致肿瘤破坏不足和对健康组织的损害。这项研究旨在开发个性化的三维（3D）模型来模拟肝脏肿瘤的MWA，结合患者特有的特征。主要目标是验证预测的消融区与临床结果的比较，在治疗前提供对MWA的见解，以促进准确的治疗计划。来自三名患者的对比增强CT图像用于创建3D模型。模拟使用耦合电磁波传播和生物传热来估计温度分布，预测肿瘤破坏和消融边缘。研究结果表明，一旦达到足够的边缘，长时间的消融并不能显着改善肿瘤的破坏。虽然它增加了组织损伤。临床消融区和预测消融区之间存在大量重叠。对于患者1，Dice评分为0.73，表明准确性高，灵敏度为0.72，特异性为0.76。对于患者2，Dice评分为0.86，灵敏度为0.79，特异性为0.96。对于患者3，Dice评分为0.8，灵敏度为0.85，特异性为0.74。患者特定的3D模型显示出准确预测消融区域和优化MWA治疗策略的潜力。
