Automating MRI Delta T1 Methods for the Routine Assessment of Brain Tumor Burden

描述(申请人提供)：MRI(磁共振成像)的对比度增强为测量脑部和其他实体肿瘤对治疗的反应提供了目前可用的最佳方法。因此，国际委员会制定了通过测量对比剂增强肿瘤的肿瘤直径来指导肿瘤负担评估的标准(如RECIST、Macdonald和RANO标准)。尽管增强后MRI在日常实践和大多数临床试验中被广泛使用，但这种方法在脑肿瘤中的使用存在一些重要的限制。首先，恶性胶质瘤在组织病理学和放射学表现上具有异质性，边缘不规则，强化程度不同，中央有坏死区或囊性改变，这使得主观和手动识别和测量强化感兴趣区极具挑战性。其次，术后肿瘤体积的评估可能会被血液制品的存在所混淆，这些血液制品在增强后的MRI上也显示出亮点。尽管对比前后T1加权图像的视觉对比通常足以做出这种区分，但当增强很细微时，这可能是非常具有挑战性的。最后，随着具有类固醇效应的抗血管生成药物的使用越来越多，微妙的对比剂增强的病例越来越常见。这些挑战可能解释了为什么观察者之间在评估肿瘤负担方面的巨大差异(超过50%)困扰着大多数临床试验。为了克服这些局限性，我们建议开发Delta T1(Dt1)方法，用于自动检测真实对比剂增强和自动生成增强ROI，并提供生成RECIST/Macdonald/RANO指标的选项。这些工具将与目标2中概述的评估肿瘤负担的标准方法进行比较。[Dt1方法的新颖性源于这样一个事实，即它结合了一项获得专利的“图像强度标准化”技术，使新开发的工具具有相对于现有方法的重要优势，并有可能改变临床实践范式。]具体地说，标准化步骤已经获得专利，并独家授权给成像生物识别有限责任公司，消除了由于正常的MRI系统可变性、成像参数的细微差异等导致的图像对比度的大部分正常可变性。因此，DT1和相关的ROI工具将被纳入成像生物识别公司的低成本产品IB SuiteTM，将在广泛的基础上提供。它将使肿瘤ROI的稳健和可重复性确定成为可能，从而消除或显著减少当前观察者间可变性的问题。因此，生产这一工具有可能导致在临床试验和日常实践中评估肿瘤负担的方式发生范式转变，提高可靠性和工作流程，从而更好地照顾脑瘤患者。 公共卫生相关性：这项第一阶段STTR提案的目标是开发急需的磁共振图像分析工具，用于健壮和自动确定脑瘤负担。结合新的标准化算法，创建差异或“deltaT1”图，以及阈值由生物指标决定的自动ROI方法，使所开发的工具极有可能显著减少困扰当前方法的观察者内和观察者间的高度差异。这些工具的开发和验证将与小型企业成像生物识别有限责任公司合作进行，该公司在将前景看好的实验室医学图像分析软件转化为临床工具方面有着良好的记录。因此，开发的工具将广泛用于肿瘤治疗反应的日常临床评估，以及更大规模的临床试验。这反过来可以带来更有效和更具成本效益的临床试验，以及在个性化的基础上改善对患者的护理。

DESCRIPTION (provided by applicant): Contrast enhancement on MRI (magnetic resonance imaging) provides the best currently available approach for measuring tumor response to therapy in brain and other solid tumors. Accordingly, criteria have been formulated, by international committees, to guide the assessment of tumor burden by measuring tumor diameters of contrast-enhancing tumor (eg RECIST, Macdonald and RANO criterion). Despite the widespread use of post-contrast MRI in daily practice and most clinical trials, there exist several important limitations to this approach for use in brain tumors. First, malignant gliomas are histopathologically and radiographically heterogenous in appearance with geographically irregular margins, variable enhancement, and regions of central necrotic or cystic changes, making the subjective and manual identification and measurement of enhancing ROIs extremely challenging. Second, assessment of postoperative tumor volume can be confounded by the presence of blood products, which also appear bright on post-contrast MRI. Though a visual comparison between pre and post-contrast T1-weighted images is often sufficient to make this distinction, this can be very challenging when the enhancement is subtle. Finally, with the increasing use of anti-angiogenic agents, which have a steroid-like effect, the number of cases with subtle post-contrast enhancement is becoming increasingly common. These challenges may explain the large inter-observer differences (greater than 50%) in assessing tumor burden that plague most clinical trials. To overcome these limitations, we propose to develop the delta T1 (dT1) method for automatic detection of true contrast-agent enhancement and the automatic generation of enhancing ROIs with options to generate RECIST/Macdonald/RANO metrics. These tools will be compared against standard approaches for the assessment of tumor burden as outlined in Aim 2. [The novelty of the dT1 method derives from the fact that it incorporates a patented "image intensity standardization" technology, giving the newly developed tools an important advantage over existing methods and the potential to shift clinical practice paradigms.] Specifically, the standardization step, which has been patented and exclusively licensed to Imaging Biometrics LLC, eliminates much of the normal variability in image contrast due to normal MRI system variability, slight differences in imaging parameters and the like. Thus, the dT1 and associated ROI tools, which will be incorporated into Imaging Biometrics low-cost product, IB SuiteTM, will be made available on a widespread basis. It will enable the robust and reproducible determination of tumor ROIs that eliminate or significantly minimize current issues of inter-observer variability. Accordingly, productizing this tool has the potential to result in a paradigm shift in how tumor burden is assessed in clinical trials and daily practice, improving reliability and workflow resulting in better care for patients with brain tumors. PUBLIC HEALTH RELEVANCE: The goal of this Phase I STTR proposal is the development of much needed MR image analysis tools for the robust and automatic determination of brain tumor burden. The incorporation of novel standardization algorithms, creation of difference or "deltaT1" maps, as well as automatic ROI methods whose thresholds are dictated by biologic indicators, give the developed tools a high likelihood of significantly diminishing the high intra- and inter-observer differences that plague current methods. The development and validation of these tools will be performed in collaboration with Imaging Biometrics LLC, a small business concern, who has a proven track record of translating promising laboratory medical image analysis software into clinical tools. Therefore the developed tools will be made widely available for the daily clinical assessment of tumor response to therapy, as well as for larger scale clinical trials. This in turn can result in the performance of more efficient and cost-effective clinical trials, as well as improved care of patients on an individualized basis.