Assessment of organ-specific biological age based on whole-body MR data from the German National Cohort Study. Phase II: Correlation with epidemiological, functional and clinical data

年龄是许多不同医学疾病的重要风险因素，在临床决策中起着重要作用。生物年龄的概念（与实足年龄相反）旨在捕捉衰老过程中的个体间差异，并为个人与年龄有关的变化提供一种衡量标准。然而，由于缺乏明确的定义和生物年龄概念的重叠，生物年龄很难量化。例如，与年龄相关的变化可以是器官特异性的，因此生物学年龄可以指特定的器官系统。医学成像提供的形态学数据有助于更好地了解这一衰老过程，也有助于了解特定器官系统。NAKO研究提供了一个独特的数据集，包括约30，000次全身成像检查以及参与者的流行病学，功能和临床信息。在该项目正在进行的第一阶段，我们的团队能够建立和验证基于NAKO研究的前11，000个成像数据集的衰老过程的基于图像的器官特异性分析的机器学习方法。我们在该项目扩展提案中的目标是将这些既定方法扩展到现有的30，000个NAKO数据集，将分析扩展到进一步的器官系统，并对成像和非成像数据进行联合分析，旨在获得表型衰老过程的全面表征并了解其决定因素。此外，我们的目标是实现基于图像的分析过程的自动化质量分析，以确保结果的有效性，并实现方法学的稳健临床翻译。

Age is an important risk factor for a number of different medical disorders and plays an important role in clinical decision-making. The concept of biological age (in contrast to chronological age) aims to capture interindividual differences in the ageing process and provides a measure for age-related changes in individuals. However, biological age is hard to quantify due to lack of a clear definition and due to overlapping concepts of biological age. For example, age-related changes can be organ-specific so that biological age can refer to specific organ systems. Morphological data provided by medical imaging can contribute to a better understanding of this ageing process, also with respect to specific organ systems. The NAKO study offers a unique dataset of about 30,000 whole-body imaging examinations together with epidemiological, functional and clinical information of participants. In the ongoing Phase I of this project our group was able to establish and validate machine learning methods for image-based organ-specific analysis of the ageing process based on the first 11,000 imaging data sets of the NAKO study. Our goals in this project extension proposal are the extension of these established methods to the now available 30,000 NAKO datasets, the extension of the analysis to further organ systems and the joint analysis of imaging and non-imaging data aiming to derive a comprehensive characterization of the phenotypic ageing process and to understand its determinant factors. In addition, we aim to implement automated quality analysis of the image-based analysis process in order to ensure validity of results and enable robust clinical translation of the methodology.