STTR Phase I: A Clinical Decision Support Tool for Brain Magnetic Resonance Imaging (MRI) in Children

这项小型企业技术转移（STTR）第一阶段项目的更广泛影响/商业潜力是使医生能够最大限度地从昂贵的儿童脑部磁共振成像（MRI）扫描中提取诊断信息。每年大约进行350万次脑部MRI扫描，平均价格从500美元到2000美元不等。尽管获取和解读扫描结果的成本很高，但医生通常只通过视觉检查图像来诊断异常。与核磁共振成像的总成本相比，拟议中的软件相对便宜，有望让医生对大脑的关键结构进行准确测量。这一点很重要，因为发育性疾病和其他疾病会导致人眼不易看到的小体积和表面积变化。通过早期发现这些变化，医生应该能够在早期阶段治疗疾病，从而有可能提高生活质量并节省资金。通过测量药物或治疗对大脑特定结构的影响，该工具还有望使医生更好地跟踪治疗结果。最终，预计拟议的技术将提高护理质量并降低医疗保健成本。该项目利用当前先进的计算技术，如机器学习和计算机视觉，自动测量儿童大脑关键部位的体积、表面积和形状。然后将这些测量结果与0-12岁儿童的大型数据库进行比较，以确定他们是否偏离了正常水平。尽管由于缺乏辐射，核磁共振成像已经越来越多地被用作儿童脑部疾病的诊断工具，但目前还没有一种工具可以让医生通过核磁共振扫描准确地测量儿童大脑的变化。因为，在儿童中，随着她/他的成长，大脑正在迅速变化，很难从视觉上确定这些变化是由于正常发育还是疾病。通过精确测量大脑，并将其与该团队已经收集的大型数据库进行比较，预计儿科医生将能够确定儿童的大脑是否偏离正常。它还可以让医生更好地选择治疗方法，并监测患者对特定治疗的反应。

The broader impact/commercial potential of this Small Business Technology Transfer (STTR) Phase I project is to enable doctors to maximize the diagnostic information extracted from costly Pediatric Magnetic Resonance Imaging (MRI) scans of the brain. Annually, approximately 3.5 million brain MRI scans are performed with average prices with interpretation from $500 - $2,000. Even though the scans are expensive to acquire and interpret, doctors generally only use visual inspection of the images to diagnose abnormalities. The proposed software will be relatively cheap compared to the overall cost of an MRI and is expected to let doctors make accurate measurements of key structures in the brain. This is important because developmental and other diseases can cause small volume and surface area changes that cannot be easily seen by the human eye. By detecting these changes early, doctors should be able to treat the disease at early stages, potentially leading to better quality of life and financial savings. The proposed tool is also expected to enable doctors to better track treatment outcomes, by measuring the effect a drug or treatment has on particular structures in the brain. Ultimately, it is expected that the proposed technology will improve quality of care as well as reduce healthcare costs.The proposed project leverages current advances in computational technology such as machine learning and computer vision to automatically measure volume, surface area and shapes in critical parts of the brain in children. These measurements are then compared to a large database of children over the ages of 0-12 years to determine if they have deviated from normal. Although MRIs have been increasing adopted as the diagnostic tool of choice for childhood brain disorders due to the lack of radiation, there has yet to be a tool that allows doctors to accurately measure changes in a child's brain from MRI scans. Because, in children, the brain is rapidly changing as she/he grows, it is difficult to determine visually whether the changes are due to normal development or disease. By measuring the brain accurately and comparing it to a large database already collected by the proposing team, it is expected that a pediatric doctor will be able to determine whether a child's brain has deviated from normal. It could also allow physicians to better select treatments and monitor the patient response to a specific therapy.