Satti Reem, Abid Noor-Ul-Hoda, Bottaro Matteo, De Rui Michele, Garrido Maria, Raoufy Mohammad R, Montagnese Sara, Mani Ali R
UCL Division of Medicine, University College London, London, United Kingdom.
Department of Medicine, University of Padova, Padova, Italy.
Front Physiol. 2019 Feb 19;10:116. doi: 10.3389/fphys.2019.00116. eCollection 2019.
The Poincaré plot is a geometrical technique used to visualize and quantify the correlation between two consecutive data points in a time-series. Since the dynamics of fluctuations in physiological rhythms exhibit long-term correlation and memory, this study aimed to extend the Poincaré plot by calculating the correlation between sequential data points in a time-series, rather than between two consecutive points. By incorporating this so-called lag, we hope to integrate a temporal aspect into quantifying the correlation, to depict whether a physiological system holds prolonged association between events separated by time. In doing so, it attempts to instantaneously characterize the intrinsic behavior of a complex system. We tested this hypothesis on three different physiological time-series: heart rate variability in patients with liver cirrhosis, respiratory rhythm in asthma and body temperature fluctuation in patients with cirrhosis, to evaluate the potential application of the extended Poincaré method in clinical practice. When studying the cardiac inter-beat intervals, the extended Poincaré plot revealed a stronger autocorrelation for patients with decompensated liver cirrhosis compared to less severe cases using Pearson's correlation coefficient. In addition, long-term variability (known as SD2 in the extended Poincaré plot) appeared as an independent prognostic variable. This holds significance by acting as a non-invasive tool to evaluate patients with chronic liver disease and potentially facilitate transplant selection as an adjuvant to traditional criteria. For asthmatics, employing the extended Poincaré plot allowed for a non-invasive tool to differentially diagnose various classifications of respiratory disease. In the respiratory inter-breath interval analysis, the receiver operating characteristic (ROC) curve provided evidence that the extension of the Poincaré plot holds a greater advantage in the classification of asthmatic patients, over the traditional Poincaré plot. Lastly, the analysis of body temperature from patients using the extended Poincaré plot helped identify inpatients from outpatients with cirrhosis. Through these analyses, the extended Poincaré plot provided unique and additional information which could potentially make a difference in clinical practice. Conclusively, the potential use of our work lies in its possible application of predicting mortality for the organ allocation procedure in patients with cirrhosis and non-invasively distinguish between atopic and non-atopic asthma.
庞加莱图是一种几何技术，用于可视化和量化时间序列中两个连续数据点之间的相关性。由于生理节律波动的动力学表现出长期相关性和记忆性，本研究旨在通过计算时间序列中连续数据点之间的相关性，而非两个连续点之间的相关性，来扩展庞加莱图。通过纳入这种所谓的滞后，我们希望将时间因素整合到相关性量化中，以描述生理系统在被时间分隔的事件之间是否存在长期关联。这样做时，它试图即时表征复杂系统的内在行为。我们在三个不同的生理时间序列上验证了这一假设：肝硬化患者的心率变异性、哮喘患者的呼吸节律以及肝硬化患者的体温波动，以评估扩展庞加莱方法在临床实践中的潜在应用。在研究心脏搏动间期时，与病情较轻的病例相比，扩展庞加莱图显示失代偿期肝硬化患者使用皮尔逊相关系数时具有更强的自相关性。此外，长期变异性（在扩展庞加莱图中称为SD2）表现为一个独立的预后变量。这具有重要意义，因为它可作为一种非侵入性工具来评估慢性肝病患者，并有可能作为传统标准的辅助手段，促进移植选择。对于哮喘患者，使用扩展庞加莱图可提供一种非侵入性工具来鉴别诊断各种类型的呼吸系统疾病。在呼吸呼吸间期分析中，受试者操作特征（ROC）曲线提供了证据，表明庞加莱图的扩展在哮喘患者分类方面比传统庞加莱图具有更大优势。最后，使用扩展庞加莱图对患者体温进行分析有助于区分肝硬化住院患者和门诊患者。通过这些分析，扩展庞加莱图提供了独特且额外的信息，这可能会在临床实践中产生影响。总之，我们这项工作的潜在用途在于其可能应用于预测肝硬化患者器官分配程序中的死亡率，以及非侵入性地区分特应性和非特应性哮喘。
