Data analyzed for the article of <b>Evaluating photoplethysmography-based pulsewave parameters and composite scores for assessment of cardiac function: A comparison with echocardiography</b>

The study involved healthy volunteers (n=37, 21 female, 16 male, mean age 37.0, standard deviation: 11.4)), who claimed themselves healthy by filling out a detailed questionnaire, had normal BMI (18 - 25 kg/m2), had no history of smoking, and denied drinking alcohol regularly. Those who had been diagnosed with or had received treatment for diabetes or any CV disease were excluded from the study. Further exclusion criteria involved: pregnancy, previous cancerous disease, wearing false nails; SARS-CoV-2 infection in the last 6 months before the exam. The measurements were conducted at Semmelweis University's Városmajor Heart and Vascular Clinic. The study was approved by the Regional and Institutional Committee of Science and Research Ethics at Semmelweis University (Budapest, Hungary) (approval number 120/2018-3). Protocol In this study, we employed simultaneous recordings to capture and compare data from photoplethysmography (PPG) and echocardiography, ensuring synchronized measurement across both methodologies. PPG parameters were calculated by averaging measurements from each heartbeat during continuous recordings over a two-minute period. Concurrently, echocardiographic parameters were derived by averaging the data from 1-3 heartbeats, allowing for a direct comparison of cardiac function assessments between the two techniques, by the following. Echocardiography Blood pressure (BP) was measured thrice using an automatic sphygmomanometer before conducting a cardiac ultrasound scan. During the scan, the participant lay on the examination bed with the upper body undressed, positioned on the left side. 2D echocardiography examinations were performed with a GE Vivid E95 system with a 4Vc-D phased-array transducer (GE Vingmed Ultrasound, Horten, Norway). LV focused, ECG-gated datasets were obtained from parasternal long and short axis, apical four-chamber, apical three-chamber and apical two-chamber views at a minimum rate of 50 frames per second. Offline analyses of these datasets were performed after selecting the optimal heart cycle using commercially available software (Autostrain LV, TOMTEC Imaging Systems GmbH, Unterschleissheim, Germany). The algorithm automatically generated the endocardial contours of the cavities, which were manually corrected throughout the entire cardiac cycle. Speckle tracking technique was used for the deformation analysis. The assessed parameters can be found in Error! Reference source not found. PPG measurements During the cardiac ultrasound, a pulse waveform was recorded for 140 seconds using a special pulse oximeter on the patient's right index finger, with a 200 Hz sampling frequency (Shanghai Berry Electronic Tech Co., Ltd., Shanghai, China). The patient lay on their side, staying still. The oximeter, wirelessly connected to the SCN4ALL mobile app (E-Med4All Europe Ltd, Budapest, Hungary), sent the anonymized data in real time to a secure online database. The SCN4ALL software analyzed the signals, its proprietary algorithm identifies points of interest on the pulse wave from which it calculates over 30 morphological and pulse rate variability parameters online. Previous studies have provided insights into the system's reliability, repeatability, and detailed descriptions of its architecture and signal processing. [11], [12] (The SCN4ALL parameters assessed in this study are found in – List of echocardiographic and PPG parameters, with abbreviations and definitions.Table 1) Besides “conventional” PPG parameters, already known from the literature, composite parameters, called “Scores” were also analyzed. The different “Scores” are constructed using different combinations of parameters, each of which is assigned a value based on specific cutoff values along a monotonous or U-shaped Likert scale. The scores, with a maximum of 100, indicate health levels for evaluated aspects. Their actual reliability and validity in clinical practice are evaluated based on the current and upcoming studies. The exact constituents of the Scores are a proprietary secret, kept confidential at the manufacturer’s discretion.本研究纳入37名健康志愿者（女性21名，男性16名，平均年龄37.0岁，标准差11.4），志愿者通过填写详细问卷自述健康，体重指数（body mass index, BMI）处于18~25 kg/m²区间，无吸烟史，且否认规律饮酒。本研究排除曾被诊断为糖尿病或任何心血管（cardiovascular, CV）疾病并接受过治疗的受试者，进一步排除标准还包括妊娠、既往恶性肿瘤病史、佩戴假指甲，以及检查前6个月内曾感染严重急性呼吸综合征冠状病毒2型（SARS-CoV-2）。所有测量均在塞麦尔维斯大学瓦罗斯马约尔心脏与血管诊所完成。本研究已通过塞麦尔维斯大学区域与机构科学研究伦理委员会（匈牙利布达佩斯）审批（审批号：120/2018-3）。 实验方案：本研究采用同步记录方案，采集并对比光电容积描记法（photoplethysmography, PPG）与超声心动图的数据，确保两种检测方法的测量同步。PPG参数通过对2分钟连续记录中每次心跳的测量值取平均计算得到。与此同时，超声心动图参数通过对1~3次心跳的数据取平均推导得出，以此实现两种心脏功能评估技术的直接对比，具体如下。 超声心动图：在进行心脏超声扫描前，使用自动血压计对血压（blood pressure, BP）进行三次测量。扫描期间，受试者脱去上衣，左侧卧位躺于检查床。采用GE Vivid E95超声系统搭配4Vc-D相控阵探头（GE Vingmed Ultrasound，挪威霍滕）完成二维超声心动图检查。从胸骨旁长轴、短轴、心尖四腔、心尖三腔及心尖二腔切面获取左心室（left ventricle, LV）聚焦、心电图（electrocardiogram, ECG）门控数据集，采集帧率最低为50帧/秒。采用商用软件（Autostrain LV，TOMTEC Imaging Systems GmbH，德国翁特施莱伊斯海姆）选取最优心动周期后，对数据集进行离线分析。该算法可自动生成心腔的心内膜轮廓，随后在整个心动周期中进行手动校正。采用斑点追踪技术进行形变分析。本研究评估的参数详见表1。 PPG测量：在心脏超声扫描期间，使用专用脉搏血氧仪在受试者右手食指处记录脉搏波形140秒，采样频率为200 Hz（上海贝瑞电子科技有限公司，中国上海）。受试者侧卧并保持静止。该血氧仪通过无线连接至SCN4ALL移动应用（E-Med4All欧洲有限公司，匈牙利布达佩斯），将匿名数据实时发送至安全在线数据库。SCN4ALL软件对信号进行分析，其专有算法可识别脉搏波上的兴趣点，据此在线计算超过30项形态学参数与脉搏率变异性参数。既往研究已对该系统的可靠性、可重复性，以及其架构与信号处理流程进行了详细阐述[11,12]。（本研究中评估的SCN4ALL参数详见表1：超声心动图与PPG参数列表，含缩写与定义。） 除文献中已报道的“常规”PPG参数外，本研究同时分析了名为“评分”的复合参数。不同“评分”由不同参数组合构建而成，每个评分基于单调或U型李克特量表的特定截断值赋予分值，总分最高为100分，用于反映评估维度的健康水平。该评分在临床实践中的实际可靠性与有效性，正通过本研究及后续研究进行评估。“评分”的具体构成属于制造商的专有商业机密，由制造商自行保密。