Approach-specific, multi-GPU, multi-tool, high-realism neurosurgery simulation

描述（由申请人提供）：模拟在神经外科中取得了有限的进展，尽管验证研究表明其优于传统培训。它强调了针头插入，而没有解决未来大多数病例对受训外科医生的要求。现有的模拟器表现出大脑网格，缺乏可分离性的裂缝，是不够稀疏，以解决大位移交互式，并没有考虑到关键组织。这些模拟器也只模拟了简单的仪器。我们的长期目标是建立一个交互式模拟器，可以复制年轻外科医生未来的大多数病例，使面临压缩实习时间表的医院能够加速培训并提高技能，其中患者结果与手术经验相关。这一目标将通过整合多分辨率脑网格来实现，该网格在粗层次上是稀疏的和脑沟可分离的，在细层次上是描述性的，生物力学基于交互式非线性多网格有限元，通过柔性支撑襻实现多工具交互，以及通过手术本体指定的临床要求。我们的假设是，交互式神经外科模拟是相关的，技术上可行的，解剖建模必须充分描述术中运动和组织形态，生物力学必须忠实于组织反应，触觉必须提供多工具的相互作用，系统必须满足临床要求，反映特定的手术方法和病理。 公共卫生相关性：现有的神经外科手术模拟器不能满足受训外科医生的需求，因为合成生物力学组织响应所需的大脑网格模型以及由用户操纵以向计算机输入手势的触觉设备没有足够的描述性，并且因为用于模拟的临床规范不够严格。我们的长期目标是一个互动的模拟器，复制大多数年轻外科医生未来的病例。该目标将通过以下方式实现：1）多分辨率脑网格，其在脑裂处稀疏且可分离，在粗水平上描述相关组织，2）基于交互式非线性有限元的生物力学，3）通过柔性支撑襻实现的多工具交互，以及4）通过基于特定方法和病理学的手术本体指定的临床要求。

DESCRIPTION (provided by applicant): Simulation has made limited inroads in neurosurgery, despite validation studies demonstrating its advantages over traditional training. It has emphasized needle insertion and failed to address requirements of surgeons-in-training on most future caseloads. Existing simulators exhibit brain mesh that lacks separability about fissures, is insufficiently sparse to resolve large displacements interactively, and does not account for critical tissues. These simulators also have only modeled simple instruments. Our long-term objective is an interactive simulator that replicates most future cases of young surgeons, enabling hospitals faced with compressed internship schedules to accelerate training and improve skills where patient outcome correlates with surgical experience. This objective will be met by integrating a multi-resolution brain mesh that is sparse and sulcal-separable at the coarse level and descriptive at the fine level, biomechanics based on interactive nonlinear multi-grid finite elements, multi-tool interaction achieved through flexible haptics, and clinical requirements specified through surgical ontologies. Our hypothesis is that for interactive neurosurgery simulation to be relevant and technically feasible, anatomical modeling must be sufficiently descriptive in intra-surgical motion and tissue morphology, biomechanics must be faithful to tissue response, haptics must afford multi-tool interaction, and the system must meet clinical requirements reflecting the specific surgical approach and pathology. PUBLIC HEALTH RELEVANCE: Existing neurosurgery simulators fail to meet the needs of surgeons-in-training because the brain mesh model, needed for synthesizing biomechanical tissue response, as well as the haptic device, manipulated by the user to enter a gesture to the computer, are not descriptive enough, and because clinical specifications for simulation are insufficiently rigorous. Our long-term objective is an interactive simulator that replicates most future caseloads of young surgeons. This objective will be met by integrating 1) a multi-resolution brain mesh that is sparse and separable at brain fissures at the coarse level and descriptive of relevant tissues at the fine level, 2) biomechanics based on interactive nonlinear finite elements, 3) multi-tool interaction achieved through flexible haptics, and 4) clinical requirements specified through surgical ontologies based on the specific approach and pathology.