Non-invasive Detection of Cerebral Aneurysm Recurrence after Endovascular Treatment Using Automated Image Processing

项目总结 脑动脉瘤破裂出血是一种致死性极高的疾病。对于更多 在美国，每年有30,000名患者被诊断为动脉瘤，治疗包括预防 破裂，并且越来越依赖血管内技术。然而，治疗的持久性是未知的 复发率估计为16%-40%，再次治疗为10%-20%。确保动脉瘤的现行黄金标准 闭塞是基于导管的数字减影血管造影术(DSA)，是一种有意义的侵入性方法 效果。在这里，我们提出了一种替代方案，使用简单的头骨x光和自动图像处理来识别 对复发可能性高的患者，选择他们进行进一步的调查。这方面的发展 技术是医疗创新公司(MIC)和UTHealth合作的结果 和赫尔曼纪念医院(UTH/MHH)。我们计划测试一个假设，即动脉瘤复发可以 使用标准的头骨X光检查。在目标1中，我们将开发一种自动计算机算法 检测弹簧圈后的动脉瘤复发。数据来自于在UTH/MHH接受治疗的患者的既定队列。 对x光的自动计算机分析(在最初的治疗和6个月的随访中)将预测动脉瘤。 使用线圈形态测量法(大小、形状、方位)进行复发。算法将通过将其与 跟进的黄金标准(DSA)。在目标2中，算法性能的初步验证将在 新颖的数据集。使用与训练相同的方案治疗的相似患者的验证数据集(n=150) 数据集将使用自动算法进行处理。将对算法的性能进行评估 使用接收器操作员特性来确定最佳的灵敏度/特异度。如果成功，这样的 该方法可以对患者的风险进行分层，并确定哪些患者应该接受血管造影术。降低利用率 血管造影术的成功将大大减少并发症和医疗成本，对公众有巨大的好处。这 第一阶段STTR拨款将允许在第二阶段应用之前进行算法开发和测试，以及更广泛的 临床试验。MIC和UTH/MHH之间的合作伙伴关系结合了医疗商业化的经验 临床神经外科的软件。 机密-UTHealth

PROJECT SUMMARY Hemorrhage due to cerebral aneurysm rupture is a devastating condition with high mortality. For the more than 30,000 patients in the US who are diagnosed annually with an aneurysm, treatment consists of preventing rupture, and increasingly relies of endovascular techniques. However, treatment durability is unknown with recurrence estimated at 16-40% and the re-treatment of 10-20%. The current gold standard to ensure aneurysm obliteration is catheter-based digital subtraction angiography (DSA), an invasive method with significant side effects. Here, we propose an alternative that uses simple skull x-rays and automated image processing to identify patients who are high likelihood of recurrence and select them for further investigation. Development of this technique is the result of a collaboration between the Medical Innovations Company (MIC) and the UTHealth and Memorial Hermann Hospital (UTH/MHH). We plan to test the hypothesis that aneurysm recurrence can be detected using standard skull x-rays. In Aim 1, we will develop an automated computer algorithm that detects aneurysm recurrence after coiling. Data from an established cohort of patients treated at UTH/MHH. Automated computer analysis of the x-rays (at initial treatment and 6-month follow) will predict aneurysm recurrence using coil morphometry (size, shape, orientation). The algorithm will be trained by comparing it to the gold standard for follow up (DSA). In Aim 2, preliminary validation of algorithm performance will be tested in a novel dataset. A validation dataset (n=150) of similar patients treated with the same protocol as the training dataset will be processed using the automated algorithm. The performance of the algorithm will be assessed using receiver operator characteristics to determine optimal sensitivity/specificity. If successful, such an approach could stratify risk in patients and determine which should undergo angiography. Reducing utilization of angiography will significantly reduce complications and medical cost at an immense benefit to the public. This Phase I STTR grant will allow for algorithm development and testing prior to a Phase II application and broader clinical trials. The partnership between MIC and UTH/MHH combines experience commercializing medical software with clinical neurosurgery. CONFIDENTIAL- UTHEALTH