Computer aided diagnosis of neurological damage to improve care for infants born prematurely

早产是儿童神经精神障碍的主要原因，并导致重大的长期临床、教育和社会问题。在工业化国家，早产和低出生体重的发生率在过去十年中有所增加，在失业者和受教育程度较低的人群中，早产的发生率更高。损伤的负担是相当大的：在33周之前出生的所有婴儿中，约有10%发展为脑性瘫痪；超过30%的婴儿有神经认知问题；在25周或更少出生的所有存活婴儿中，有一半在30个月大时出现神经发育障碍。这些问题会持续到晚年，可能会给个人及其家庭带来毁灭性的后果。与早产儿及其家人一起工作的临床医生面临的一个主要问题是确定哪些婴儿最有可能随后发生神经发育障碍，以及哪些婴儿可能从早期干预服务中受益。改善对日后残疾的预测有可能立即改善对早产儿及其家人的护理。与此同时，改进的诊断也将有助于越来越多的人寻找特定的治疗方法来减少脑损伤。几种有希望的方法正在积极研究中，所有这些方法都依赖于或将得到不良结果诊断的改进。目前，对早产儿脑发育和预后的早期评估在很大程度上依赖于对临床和低分辨率成像数据的主观评估。该项目的目标是创建能够基于高分辨率磁共振成像(MRI)信息检测和诊断大脑发育异常的工具和算法。通过在循证统计框架内解释这些图像，将有可能作出更全面、更客观、更循证的解释。该项目将结合计算机和成像科学中的两个新兴范式，以应对识别大脑发育异常和预测结果的挑战：机器学习技术和计算解剖学。结合起来，这些方法有可能提供有用的和描述性的潜在解剖学模型，可以用于跨对象和随时间的比较。这为了解正常和异常大脑发育的模式以及预测未来大脑发育的模式提供了可能性。这项研究的结果将是显著提高预测晚年神经发育结果的能力。预测结果的能力改善了父母的咨询和婴儿的选择，以便采取旨在预防或改善脑损伤的早期治疗策略。

Preterm birth is a major cause of neuropsychiatric impairment in childhood and leads to significant long-term clinical, educational and social problems. The incidence of preterm birth and low birth weight has increased over the last decade in industrialised countries, and preterm delivery has a higher prevalence among the unemployed and poorly educated. The burden of impairment is considerable: about 10% of all infants born before 33 weeks of age develop cerebral palsy; over 30% have neurocognitive problems; and half of all surviving infants born at 25 weeks or less show neurodevelopmental impairment at 30 months of age. These problems persist into later life which can have devastating consequences for the individuals and their families. A major issue confronting clinicians who work with preterm infants and their families is the identification of infants who are most at risk for subsequent neurodevelopmental disability and who may benefit from early intervention services. Improved prediction of later handicaps has the potential immediately to improve the delivery of care for preterm infants and their families. At the same time, the improved diagnosis will also aid the growing search for specific treatments to reduce brain injury. Several promising approaches are under active investigation, all of which rely or would be aided by improved diagnosis of adverse outcomes. Currently, the early assessment of brain development in preterm infants and prognosis of outcome is heavily dependent on a subjective assessment of clinical and low resolution imaging data. The aim of this project is the creation of tools and algorithms that enable the detection and diagnosis of abnormal brain development based on high-resolution magnetic resonance imaging (MRI) information. By interpreting these images within an evidence-based statistical framework, a more complete and objective, evidence-based interpretation will be possible. The project will combine two emerging paradigms in computer and imaging science to address the challenge of identifying abnormal brain development and predicting outcome: Machine learning techniques and computational anatomy. In combination these approaches have the potential to provide useful and descriptive models of the underlying anatomy that can be used for comparisons across subjects and over time. This offers the possibility to learn patterns of normal and abnormal brain development and to predict the pattern of future brain development. The result of the research will be a significantly improved ability to predict neurodevelopmental outcome in later life. The ability to predict outcome improves parental counseling and selection of infants for early therapeutic strategies aiming at preventing or ameliorating cerebral injury.