MoTrPAC Supplemental Funding for Miller

研究策略： 这项工作的主要目的将是开展协调所需的前期工作 使用MoTrPAC的不同系统收集的双能X射线吸收(DXA)数据 学习网站。制造扫描仪(例如，制造商和型号)和软件版本为 不同地点的情况有所不同。由于使用了专有技术和算法，计算出的 不同机器的测量结果可能会有所不同。标准化和交叉校准很重要 了解测量之间的差异是否超过最小显著变化 或最小的可检测到的差异。为了将来使用这些数据，关键是要了解 站点之间的差异是否表示评估之间的实际差异，或者 评估之间的差异是否在预期的测量误差范围内。至 获得一致的DXA扫描结果，在此之前需要执行质量控制(QC) 数据可以进一步用于分析。 一些研究已经建立了DXA仪器之间的换算方程。然而，这些 研究没有使用MoTrPAC中使用的相同工具和/或方程 基于小型数据集生成。我们建议通过分析来做更多的验证工作 几个外部数据集，用于评估是否可以在我们自己的系统中执行校准 数据集。除了体内(即队列比较)的方法外，我们还将使用体外(即， 模体比较)检测仪器间差异的方法。仪器间 将对3个临床站点进行校准。

Research Strategy: The overarching purpose of this work will be to perform preliminary work required to harmonize dual-energy X-ray absorptiometry (DXA) data collected using diverse systems across MoTrPAC study sites. The manufacture scanners (e.g., makes and models) and software versions are different across sites. Because of proprietary techniques and algorithms used, the calculated measurements can vary across machines. Standardization and cross calibration are important to understand whether differences between measurements exceed the least significant change or smallest detectable difference. For future use of these data, it is critical to understand whether differences between sites indicate an actual difference between assessments, or whether differences between assessments fall within the expected measurement error. To obtain consistent DXA scan results, quality control (QC) needs to be performed before these data can be further used for analysis. A few studies have developed conversion equations between DXA instruments. However, these studies did not use the same instruments used in MoTrPAC and/or the equations were generated based on small datasets. We propose to do more validation work by analyzing several external datasets to evaluate whether the calibration can be performed in our own dataset. Besides the in vivo (i.e., comparison of cohorts) approach, we will also use in vitro (i.e., comparison of phantoms) approach to detect the inter-instrument difference. Inter-instrument calibration will be carried out for 3 clinical sites.