Sparks Rachel, Vakharia Vejay, Rodionov Roman, Vos Sjoerd B, Diehl Beate, Wehner Tim, Miserocchi Anna, McEvoy Andrew W, Duncan John S, Ourselin Sebastien
Centre for Medical Image Computing, University College London, London, UK.
Department of Clinical and Experimental Epilepsy, University College London Institute of Neurology, London, UK.
Int J Comput Assist Radiol Surg. 2017 Aug;12(8):1245-1255. doi: 10.1007/s11548-017-1628-z. Epub 2017 Jun 15.
Epilepsy is potentially curable with resective surgery if the epileptogenic zone (EZ) can be identified. If non-invasive imaging is unable to elucidate the EZ, intracranial electrodes may be implanted to identify the EZ as well as map cortical function. In current clinical practice, each electrode trajectory is determined by time-consuming manual inspection of preoperative imaging to find a path that avoids blood vessels while traversing appropriate deep and superficial regions of interest (ROIs). We present anatomy-driven multiple trajectory planning (ADMTP) to find safe trajectories from a list of user-defined ROIs within minutes rather than the hours required for manual planning.
Electrode trajectories are automatically computed in three steps: (1) Target Point Selection to identify appropriate target points within each ROI; (2) Trajectory Risk Scoring to quantify the cumulative distance to critical structures (blood vessels) along each trajectory, defined as the skull entry point to target point. (3) Implantation Plan Computation: to determine a feasible combination of low-risk trajectories for all electrodes.
ADMTP was evaluated on 20 patients (190 electrodes). ADMTP lowered the quantitative risk score in 83% of electrodes. Qualitative results show ADMTP found suitable trajectories for 70% of electrodes; a similar portion of manual trajectories were considered suitable. Trajectory suitability for ADMTP was 95% if traversing sulci was not included in the safety criteria. ADMTP is computationally efficient, computing between 7 and 12 trajectories in 54.5 (17.3-191.9) s.
ADMTP efficiently compute safe and surgically feasible electrode trajectories.
如果能够识别出癫痫发作起始区（EZ），癫痫有可能通过切除性手术治愈。如果非侵入性成像无法明确EZ，则可能需要植入颅内电极来识别EZ并绘制皮质功能图。在当前的临床实践中，每条电极轨迹都需要通过对术前成像进行耗时的人工检查来确定，以找到一条在穿过合适的深部和浅部感兴趣区域（ROI）时避开血管的路径。我们提出了一种基于解剖结构的多轨迹规划（ADMTP）方法，可在几分钟内从用户定义的ROI列表中找到安全轨迹，而不是像人工规划那样需要数小时。
电极轨迹通过三个步骤自动计算：（1）目标点选择，以识别每个ROI内合适的目标点；（2）轨迹风险评分，以量化沿每条轨迹到关键结构（血管）的累积距离，定义为从颅骨入口点到目标点的距离。（3）植入计划计算：为所有电极确定低风险轨迹的可行组合。
对20例患者（190根电极）进行了ADMTP评估。ADMTP降低了83%电极的定量风险评分。定性结果显示，ADMTP为70%的电极找到了合适的轨迹；类似比例的人工规划轨迹也被认为是合适的。如果安全标准中不包括穿过脑沟，则ADMTP的轨迹适合率为95%。ADMTP计算效率高，在54.5（17.3 - 191.9）秒内计算7至12条轨迹。
ADMTP能够高效地计算出安全且手术可行的电极轨迹。
