Preoperative Image Updating for Guidance during Brain Tumor Resection

项目总结 该项目将继续达特茅斯和美敦力之间富有成效的学术-产业合作伙伴关系。 通过它，达特茅斯公司的术中图像更新技术已经与Med- TRONIC最先进的S7导航系统可同时生成和可视化更新的MR(UMR)图像 通过手术。在第二个资助期，我们将全面验证和前瞻性评估开颅手术中的umr。 手术通过与术中MR(IMR)的比较。我们还提出了一个新的方向：生成UMR到 引导微创神经外科手术作为一种可能的低成本、但有效的孔内磁共振的替代方案 导向 (微创脑深部刺激图像更新的新初步结果表明，FEAS- 可信度和建议承诺)。 虽然到目前为止完成的验证研究使用的是被跟踪的触控笔，但这是一种“真实的事实” 令人印象深刻，在手术期间的任何给定时间都相对较少的点可用，并且跟踪和 特征识别/定位误差被嵌入在TRE(目标配准误差)结果中。因此，一个 我们延续计划的基石是通过IMR进行验证，可在达特茅斯中心获得 外科创新(CSI)。由于在切除结束时部署IMR来调查开放手术中的残留疾病 出于对患者安全的考虑，颅脑手术和多个IMR很难被证明是合理的，我们建议 一种新的大型动物胶质瘤模型已经在我们手中成功地开发出来(我们现在可以长得很结实了 位于不同颅内位置的不同大小和形状的肿瘤) 其中IMR将被收购 在切除过程中多次进行UMR验证。这些动物研究也非常适合于 微创病例，因为我们可以确定图像更新要求来指导手术 作为IMR的更有效和更具成本效益的替代方案。CSI可以适应这些实验，并且作为一个 结果，我们处于一个独特的位置，可以在同一空间进行动物和人类的研究 导航/成像设备在开颅和微创条件下应用IMR进行UMR验证 条件。基于到目前为止的进展和这些考虑，我们提出了将 将图像更新应用于微创神经外科手术，通过GPU加速图像更新 处理，并将UMR数据添加到外科医生的平视显示器中；在大型动物胶质瘤中进行验证打开- 颅脑和微创研究，其中IMR采集不受限制，在类似人脑的情况下 限制IMR使用的肿瘤病例；以及对切除结束手术准确性的前瞻性评估 使用术前MR(PMR)导航至也有UMR的手术。到年底的时候 建议的第二个资助期，我们将有一个UMR指导平台，充分验证开放颅脑和 微创手术，并将展示它使神经外科医生能够在多大程度上 更频繁地实现更准确的开颅手术，以及是否足够准确和及时地为 微创神经外科干预的指导相对于孔内磁共振指导。

PROJECT SUMMARY This project will continue a productive academic-industrial partnership between Dartmouth and Medtronic through which Dartmouth’s intraoperative image updating techniques have already been merged with Med- tronic’s state-of-the-art S7 navigation system to produce and visualize updated MR (uMR) images concurrent with surgery. In a second funding period, we will fully validate and prospectively evaluate uMRs in open-cranial surgery via comparisons with intraoperative MR (iMR). We also propose a new direction: generating uMR to guide minimally invasive neurosurgeries as a possible low-cost, yet effective, alternative to in-bore MR- guidance (new preliminary results of image-updating in minimally-invasive deep brain stimulation indicate feas- ibility and suggest promise). While validation studies completed to date with a tracked stylus as “ground truth” have been impressive, relatively few points are available at any given time during surgery, and tracking and feature identification/localization errors are embedded in the TRE (target registration error) results. Thus, a cornerstone of our continuation proposal is validation with iMR, which is available in Dartmouth’s Center for Surgical Innovation (CSI). Since iMR is deployed at end-of-resection to survey for residual disease in open cranial surgery and multiple iMRs are difficult to justify because of patient safety concerns, we have proposed a new large animal glioma model which has been developed successfully in our hands (we can now grow solid tumors of varying size and shape located in different intra-cranial positions) in which iMR will be acquired multiple times during a resection procedure for uMR validation. These animal studies are also ideally-suited to minimally-invasive cases because we can determine the image-updating requirements to guide the procedure as a more efficient and cost-effective alternative to iMR. CSI can accommodate these experiments, and as a result, we are in a unique position to conduct animal and human studies in the same space with the same navigation/imaging equipment for uMR validation with iMR under both open cranial and minimally invasive conditions. Based on progress to date, and these considerations, we propose technical advances that will apply image-updating to minimally-invasive neurosurgical procedures, accelerate image-updating through GPU processing, and add uMR data into the surgeon’s heads-up display; validation in large animal glioma open- cranial and minimally-invasive studies where iMR acquisitions are not limited, and during similar human brain tumor cases where iMR use is restricted; and prospective evaluation of end-of-resection surgical accuracy of procedures navigated with preoperative MR (pMR) to those where uMR is also available. By the end of the proposed 2nd funding period, we will have an uMR guidance platform that is fully validated for open cranial and minimally-invasive procedures, and will have demonstrated the extent to which it enables neurosurgeons to achieve more accurate open-cranial resections more often, and whether it is sufficiently accurate and timely for guidance of minimally-invasive neurosurgical interventions relative to in-bore MR guidance.