Automated Presurgical Language Mapping via Deep Learning for Multimodal Brain Connectivity

通过深度学习进行自动术前语言映射以实现多模式大脑连接

基本信息

批准号：
10415207
负责人：
Tracy Dawn Vannorsdall
金额：
$ 18.09万
依托单位：
JOHNS HOPKINS UNIVERSITY
依托单位国家：
美国
项目类别：
财政年份：
2021
资助国家：
美国
起止时间：
2021-09-01 至 2024-08-31
项目状态：
已结题

来源：
https://reporter.nih.gov/project-details/10415207
关键词：
Address Algorithms Anatomy Anesthesia procedures Aphasia Area Behavioral Paradigm Brain Brain Neoplasms Brain region Classification Clinical Cognitive Collection Complement Complex Computational algorithm Consumption Coupled Data Diagnosis Diffusion Equipment Evaluation Excision Experimental Designs Functional Magnetic Resonance Imaging Goals Gold Grain Graph Hospitals Impairment Individual Joints Language Lead Learning Lesion Link Logistic Regressions Magnetic Resonance Imaging Maps Methods Modality Monitor Morbidity - disease rate Motor Network-based Neurocognitive Neuronal Plasticity Neurosurgeon Operative Surgical Procedures Outcome Pathway interactions Patient-Focused Outcomes Patients Pattern Persons Postoperative Period Primary Brain Neoplasms Property Public Health Quality of Care Quality of life Research Rest Risk Seeds Sleep Supervision Survival Rate System Therapeutic Time Training Treatment outcome United States Work awake base care costs cohort common treatment cortex mapping deep learning deep neural network design experience graph neural network imaging modality improved independent component analysis innovation machine learning algorithm multilayer perceptron multimodality neurosurgery patient population predictive modeling preservation prognostic value success tool tumor white matter

项目摘要

Project Summary/Abstract Approximately 100,000 people in the United States are diagnosed with a primary brain tumor each year. Neu- rosurgery remains the ﬁrst and most common therapeutic option for these patients with outcomes linked to the extent of tumor resection. However, larger resections also increase the risk for postoperative deﬁcits, particularly in the motor and language areas of the eloquent cortex. Task fMRI (t-fMRI) has emerged as a powerful nonin- vasive tool for preoperative mapping, but these acquisitions are lengthy and cognitively demanding for patients. Moreover, t-fMRI is unreliable if the patient cannot perform the tasks while in the scanner. Our long-term goal is to develop an automated platform for reliable eloquent cortex mapping across a broad patient cohort that comple- ments the existing clinical workﬂow. The overall objective of this proposal is to design and validate new machine learning algorithms that leverage the complementary strengths of resting-state fMRI (rs-fMRI) and diffusion MRI (d-MRI), which are both passive modalities and easy to acquire. Our central hypothesis is that the combined structural-functional connectivity information in these modalities will enable us to localize language functionality in patients with brain tumors. Our innovative strategy uses recent advancements in deep learning to capture com- plex interactions in the rs-fMRI and d-MRI data that collectively deﬁne the language areas. We will evaluate our hypothesis via two speciﬁc aims. In Aim 1 we will develop a graph neural network (GNN) that employs specialized convolutional ﬁlters to capture topological properties of the connectivity data across multiple scales. Our GNN will be trained in a supervised fashion and evaluated against t-fMRI activations and intraoperative electrocortical stimulation. In Aim 2 we will conduct an exploratory analysis to retrospectively link our GNN predictions to post- operative changes in language functionality. Namely, we hypothesize that patients for whom the surgical path intersects our GNN predictions will experience greater deﬁcits across ﬁne-grained language subdomains. We will also assess the prognostic value of our GNN predictions, as compared to other clinical factors. We anticipate the proposed research will have a transformative impact on surgical planning by helping neurosurgeons to plan more targeted and safer surgeries, thus improving patient outcomes and overall quality of care.

项目概要/摘要美国每年约有 100,000 人被诊断患有原发性脑肿瘤。对于这些患者来说，呼吸手术仍然是第一个也是最常见的治疗选择，其结果与然而，较大的切除也会增加术后缺陷的风险，特别是任务功能磁共振成像 (t-fMRI) 已成为一种强大的非功能性功能。术前绘图的广泛工具，但这些采集时间较长且对患者的认知要求较高。此外，如果患者在扫描仪中无法执行任务，那么 t-fMRI 就不可靠。我们的长期目标是。开发一个自动化平台，在广泛的患者群体中进行可靠的雄辩皮层映射，该平台完成了该提案的总体目标是设计和验证新机器。利用静息态 fMRI (rs-fMRI) 和扩散 MRI 互补优势的学习算法（d-MRI），这都是被动方式并且易于获取，我们的中心假设是结合起来。这些模式中的结构功能连接信息将使我们能够本地化语言功能我们的创新策略利用深度学习的最新进展来捕捉脑肿瘤患者的症状。我们将评估共同定义语言区域的 rs-fMRI 和 d-MRI 数据中的复杂相互作用。通过两个具体目标进行假设在目标 1 中，我们将开发一个采用专门技术的图神经网络（GNN）。卷积滤波器来捕获跨多个尺度的连接数据的拓扑属性。将以监督方式进行培训，并根据 t-fMRI 激活和术中皮层电图进行评估在目标 2 中，我们将进行探索性分析，以回顾性地将我们的 GNN 预测与后刺激联系起来。也就是说，我们忍受了手术路径的患者。与我们的 GNN 预测相交，我们将在细粒度语言子域中遇到更大的缺陷。与我们预期的其他临床因素相比，还评估了我们的 GNN 预测的预后价值。拟议的研究将帮助神经外科医生制定更多计划，从而对手术计划产生变革性影响有针对性且更安全的手术，从而改善患者的治疗效果和整体护理质量。