首页> 外文会议>SPIE Conference on Ultrasonic Imaging, Tomography, and Therapy >A two dimensional locally regularized strain estimation technique:preliminary clinical results for the assessment of benign and malignant breast lesions
【24h】

A two dimensional locally regularized strain estimation technique:preliminary clinical results for the assessment of benign and malignant breast lesions

机译:二维局部正则化应变估计技术:对良性和恶性乳腺病变评估的初步临床结果

获取原文

摘要

We previously developed a 2D locally regularized strain estimation technique that was already validated with ex vivo tissues. In this study, our technique is assessed with in vivo data, by examining breast abnormalities in clinical conditions. Method reliability is analyzed as well as tissue strain fields according to the benign or malignant character of the lesion. Ultrasound RF data were acquired in two centers on ten lesions, five being classified as fibroadenomas, the other five being classified as malignant tumors, mainly ductal carcinomas from grades Ito III. The estimation procedure we developed involves maximizing a similarity criterion (the normalized correlation coefficient or NCC) between pre-and post-compression images, the deformation effects being considered. The probability of correct strain estimation is higher if this coefficient is closer to 1. Results demonstrated the ability of our technique to provide good-quality strain images with clinical data. For all lesions, movies of tissue strain during compression were obtained, with strains that can reach 15%. The NCC averaged over each movie was computed, leading for the ten cases to a mean value of 0.93, a minimum value of 0.87 and a maximum value of 0.98. These high NCC values confirm the reliability of the strain estimation. Moreover, lesions were clearly identified for the ten cases investigated. Finally, we have observed with malignant lesions that compared to ultrasound data, strain images can put in relief a more important lesion size, and can help in evaluating the lesion invasive character.
机译:我们以前开发一个已经与离体组织验证二维局部正应变估计技术。在这项研究中,我们的技术与体内的数据评估,通过临床检查条件的乳房异常。方法的可靠性是根据病灶的良性或恶性性质分析以及组织的应变场。超声射频数据在两个中心获得十项病变,五被归类为纤维腺瘤,其他五个被分类为恶性肿瘤,从等级I至III主要导管癌。我们开发的估计过程涉及最大化前和后压缩图像之间的相似性准则(正规化相关系数或NCC),所考虑的变形效​​果。正确应变估计的概率较高,如果该系数接近1。结果显示我们的技术来提供具有临床数据高质量应变图像的能力。对于所有的病变,获得压缩期间组织应变的电影,与菌株,可以达到15%。该NCC平均超过每部电影被计算,领先的十大案件的0.93的平均值,0.87的最小值和0.98的最大值。这些高NCC值确认应变估计的可靠性。此外,病变明确了调查的十大案件。最后,我们有恶性病变的观察,相比于超声数据,应变图像可以缓解放一个更重要的病灶大小,并且可以在评估病变侵入字符帮助。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号