SCG-RHC: Wearable Seismocardiogram Signal and Right Heart Catheter Database

Abstract

We are providing an open access dataset of 73 patients to study better ways to monitor changes in pulmonary arterial pressure (PAP) and pulmonary capillary wedge pressure (PCWP) using a small, non-invasive, chest-worn wearable patch. The ground truth PAP and PCWP were measured using pulmonary artery wedge catheter during the standard Right Heart Catheter (RHC) procedure, and they were modulated by infusing systemic vasodilators. The wearable data including electrocardiogram (ECG) and tri-axial seismocardiogram (SCG) were also measured simultaneously and time synchronized to the ground truth data. Besides waveform records, this dataset also contains demographic data such as sex, age, weight, height, heart failure status, etc.) and other RHC values such as right atrial pressure, right ventricular pressure, stroke volume, and cardiac output.

Background

Monitoring hemodynamics using an implantable hemodynamic congestion monitoring system and subsequent proactive HF management therapies has shown to demonstrated efficacy in reducing heart failure (HF)-related rehospitalization. However, a low-cost alternative that is financially feasible for the large patient population affected with HF in the US has yet to exist. Seismocardiography (SCG), the local mechanical vibration of the chest wall associated with the movement of the heart and blood within the vasculature has demonstrated promise in monitoring cardiovascular health.

Specifically, SCG signals have demonstrated applications in the diagnosis and monitoring of different cardiac conditions, e.g., atrial fibrillation [1], [2], heart valve disease [3], coronary artery disease [4]–[6], and HF [7]–[9]. Importantly, recent studies have demonstrated that SCG can be used to assess the clinical status of patients with decompensated HF [8], [9]. Besides the assessment of clinical status in patients with HF, SCG has exhibited efficacy in tracking instantaneous oxygen uptake during cardiopulmonary exercise tests in patients with HF and daily life activities in healthy individuals [9], [10]. Based on these results in tracking hemodynamics with SCG for both healthy individuals and patients with HF, we hypothesized that the changes in hemodynamic congestion could be tracked with the simultaneously recorded SCG signal by estimating changes in PAP and pulmonary capillary wedge pressure (PCWP).

Methods

Right Heart Catheter (RHC) procedures were conducted on a total of 73 patients (14 inpatients and 59 outpatients) who were referred for hemodynamic evaluation of their heart failure (HF) status. Age range was from 26 to 84, with a mean of 55.53 years. Weight range was from 40.8125 to 149.75 with a mean of 87.83 kg. Height range was from 149.875 to 193.0 with a mean of 172.35 cm. 24 subjects are female. NYHAC range was from 1 to 4 with a mean of 2.69 (1 subjects doesn't have NYHAC score). Some patients underwent more than 1 RHC procedure (83 total RHC procedures).

The RHC procedure was carried out in a quiet, environmentally controlled cardiac catheterization laboratory with an ambient temperature of 25°C. The wearable patch was attached to the mid-sternum. Local anesthesia was administered with 2% lidocaine. Under ultrasound guidance, venous access was obtained, and a 6 French (F) introducer sheath (St. Jude Medical, St. Paul, MN) was placed in the right internal jugular or right brachial vein. After at least 20 min of rest in a supine position, a 6F balloon-tipped pulmonary artery wedge catheter (Teleflex, Morrisville, NC) was advanced under fluoroscopic guidance into the right atrium, right ventricle, pulmonary arterial, and pulmonary capillary wedge positions. At each position, pressures were acquired over 60 seconds, repeated in triplicate and averaged, per standard RHC protocols [11], [12]. Cardiac output was obtained by the Fick principle and thermodilution. After baseline hemodynamics and cardiac output were measured, and after a 10 min rest in a supine position, pharmacological agents were administered at the discretion of the HF physician performing the case. Nitroglycerin was given as sublingual spray (400 or 800 mcg), and nitroprusside was administered as intravenous (IV) infusion starting at 0.3 mcg/kg/min (and titrated by 0.3 mcg/kg/min every 5 min until a hemodynamic effect was achieved). At the peak hemodynamic effect as determined by the HF physician, the hemodynamics were repeated as per baseline protocol. Thereafter, the balloon tipped pulmonary artery wedge catheter and the venous sheath were removed. Patients were observed for adverse events following the RHC procedure and later discharged. This procedure induces statistically significant (Wilcoxon signed rank test, p<0.05) changes in PAMP, PCWP, RAMP, CO (estimated by Fick’s law and by thermodilution) for the first 20 subjects.

The clinical research coordinator (CRC) recorded the approximate times of the RHC events:

1. When the catheter was in the pulmonary artery position during baseline
2. When pulmonary capillary wedge position during baseline
3. When the catheter was in the pulmonary artery position during vasodilator infusion
4. When pulmonary capillary wedge position during vasodilator infusion

The ground truth data were measured by 6F balloon-tipped pulmonary artery wedge catheter (Teleflex, Morrisville, NC). The RHC pressure values (250 Hz) were recorded by cath lab Mac-Lab system (Mac-Lab Hemodynamic Recording System, GE Healthcare, Chicago, IL, USA).

The wearable ECG (1 kHz) and SCG (500 Hz) signals were recorded continuously throughout the RHC procedure.

Data Description

Under the /root directory, you can find a few folders: /processed_data, /raw_data, /meta_information, and /signal_preview. We recommend users use /processed_data & /meta_information for their primary analysis. Users should only use /raw_data if you find the information there useful. /signal_preview folder prepares signals plots to help users preview the dataset. 

Contents in processed data

This folder stores the data for a Right Heart Catheter (RHC) procedure of each recording. The recording ID is defined as follows:

TRMXXX-RHCY

• XXX = patient number (107-282)
• Y = Yth RHC procedure (1-4)

For each of the subfolders, three files can be found:

• TRMXXX-RHCY.dat: timeseries signals
• TRMXXX-RHCY.hea: header file of the timeseries signals
• TRMXXX-RHCY.json: demographic data of the subject and other relevant information of the RHC procedure

The timeseries signals and their header files are stored in the WFDB (WaveForm DataBase) format. The WFDB data files contain the continuous timeseries measured by both Mac-Lab Hemodynamic Recording System and the custom-built patch. The signal units and the start time of the measurements can be found in the header file. The signals have been synchronized and resampled to 500Hz. The MacStTime key in the json file corresponds to the timestamp of the first data point of these signals. 

Signals in the dataset recorded by the wearable patch (stored in the WFDB format: TRMXXX-RHCY.dat)

• Patch_ECG: ECG signal measured by the patch (unit: mV)
• Patch_ACC_lat: ACC signal in the lateral direction measured by the patch (unit: mg)
• Patch_ACC_hf: ACC signal in the head-to-foot direction measured by the patch (unit: mg)
• Patch_ACC_dv: ACC signal in the dorsal-ventral direction measured by the patch (unit: mg)
• Patch_Hum: Humidity signal measured by the patch (unit: %)
• Patch_Pre: Ambient pressure measured by the patch (unit: mbar)
• Patch_Temp: Skin temperature measured by the patch (unit: Celsius)

Note that SCG signals can be extracted using a [1, 40] Hz bandpass filter.

Signals in the dataset recorded by Mac-Lab (also stored in the same WFDB file: TRMXXX-RHCY.dat)

• RHC_pressure: RHC blood pressure recorded by Mac-Lab (unit: mmHg)
• ART: Arterial blood pressure recorded by Mac-Lab using arterial line (unit: mmHg)
• ECG_lead_I: Lead I ECG recorded by Mac-Lab (unit: mV)
• ECG_lead_II: Lead II ECG recorded by Mac-Lab (unit: mV)
• ECG_lead_III: Lead III ECG recorded by Mac-Lab (unit: mV)
• aVR: Augmented Vector Right ECG recorded by Mac-Lab (unit: mV)
• aVL: Augmented Vector Left ECG recorded by Mac-Lab (unit: mV)
• aVF: Augmented Vector Foot ECG recorded by Mac-Lab (unit: mV)
• ECG_lead_V1: V1 pericordial lead ECG recorded by Mac-Lab (unit: mV)
• ECG_lead_V2: V2 pericordial lead ECG recorded by Mac-Lab (unit: mV)
• ECG_lead_V3: V3 pericordial lead ECG recorded by Mac-Lab (unit: mV)
• ECG_lead_V4: V4 pericordial lead ECG recorded by Mac-Lab (unit: mV)
• ECG_lead_V5: V5 pericordial lead ECG recorded by Mac-Lab (unit: mV)
• ECG_lead_V6: V6 pericordial lead ECG recorded by Mac-Lab (unit: mV)
• PLETH: Plethysmogram signal recorded by Mac-Lab using pulse oximeter sensor (unit: a.u.) 
• RESP: Respiratory signal recorded by Mac-Lab (unit: a.u.) 

The demographic data (stored in the json format: TRMXXX-RHCY.json)

• age: age of the patient in the recording (yr)
• height: height of the patient in the recording (cm)
• weight: weight of the patient in the recording (kg)
• gender: gender of the patient in the recording
• sbp: systolic (arterial) blood pressure of the patient in the recording (mmHg)
• dbp: diastolic (arterial) blood pressure of the patient in the recording (mmHg)
• history of patient: medical history of the patient
• CDecomp: Clinical decompensation status (0: compensated; 1=decompensated)
• PDecomp: Physiological decompensation status (0: compensated; 1=decompensated)
• NYHAC: New York Heart Association Functional Classification
• IsChallenge: patient underwent physiological challenge
• DevLoc: location of the wearable patch placed on the patient
• MacStTime: start time of the Mac-Lab recording (TRMXXX-RHCY.dat data start at this time)
• MacEndTime: end time of the Mac-Lab recording
• ChamEvents: timestamps (absolute, date shifted) when nurses determined the catheter has entered a specific cardiac chamber during baseline (prior to physiological challenge). Note that these are not quite accurate but approximate. Please refer to clinic_names_RHC.csv for the abbreviation reference.
• ChamEvents_in_s: timestamps (relative to the begining of the data) when nurses determined the catheter has entered a specific cardiac chamber during baseline (prior to physiological challenge). Note that these are not quite accurate but approximate. Please refer to clinic_names_RHC.csv for the abbreviation reference
• outpatient: outpatient?
• heart failure: diagnosed with heart failure?
• Missing_MaclabRHC: Is Mac-Lab RHC timeseries missing?
• fine_alignment: Whether Mac-Lab signals and wearable patch signals are aligned precisely to msec level (True: Yes; False: No, only aligned coarsely to sec level). Note that for precisely aligned recordings, Mac-Lab and wearable patch signals can still drift slightly over time for a small number of recordings (<0.25s)

Note that PHI has been removed and the dates have been shifted in the json files as well as hea files. We date shifted all dates by replacing them with a fixed date (2000/01/01).

Contents in raw data

• /annotations: stores the raw annotation files for all recordings
• /macLabData: stores the raw Mac-Lab data for all recordings. In other words, this folder stores the raw Mac-Lab data used to generate the Mac-Lab signals in the WFDB file. Note that raw data are exported by Mac-Lab so they are stored in the .txt format. 
• /wearable_patch: stores the raw wearable patch data for all recordings. In other words, this folder stores the raw wearable patch data used to generate the wearable patch signals in the WFDB file. Note that raw data are parsed and exported by custom-built software so they are store in the .mat format. 

Contents in meta information

• acronyms_RHC.csv: provides the description of the abbreviations found in TRMXXX-RHCY.json
• HO_names_RHC.csv: provides the description of the abbreviations of the history of patients key found in TRMXXX-RHCY.json

{ 
  history of patient: "NICM s/p OHTx, CKD3, HTN",
}

• clinic_names_RHC.csv: provides abbreviation reference of the maclabMeas key found in TRMXXX-RHCY.json

{ 
  maclabMeas: {
        RAA Wave :10,
        RAV Wave :10,
        RAM :8,
        RAHR :60,
        ... ,
      }
}

• PAM_PCWP_timestamp_in_TBME.json: provides timestamps used for extracting measurements during baseline (BL) and during vasodilator infusion (VI) [13]. TRM235.RHC1 is the only recording without VI timestamps.

1. PA_BL: When the catheter was in the pulmonary artery position during baseline
2. PCW_BL: When pulmonary capillary wedge position during baseline
3. PA_VI: When the catheter was in the pulmonary artery position during vasodilator infusion
4. PCW_VI: When pulmonary capillary wedge position during vasodilator infusion

• RHC_values.csv: provides the RHC values recorded, clinical status classification, compensated/decompensated, etc.
• list_coarsely_aligned.txt: provides the recording IDs that weren't precisely time aligned. Caution should be used when processing these recordings for estimation tasks.
• list_exported_recs.txt: provides all recording IDs used in this dataset.

Contents in signal preview

In this folder, plots are provided to help users quickly preview the quality of the data are provided. Three different windows of the WFDB signals are plotted for all subjects: 

• /FirstChamber_20s: shows the WFDB signals 5s before and 15s when RHC first entered a chamber.
• /full: shows the full recording of the WFDB signals
• /random_20s: shows a randomly selected 20s window of the WFDB signals

Usage Notes

This dataset could be used to train and test models for estimating PAP, PWCP, changes in PAP, changes in PWCP, using ECG and tri-axial SCG measured from a wearable patch. Further, it could be used to develop or validate models of pulmonary mechanics and/or to develop methods to identify patient-specific parameters which cannot be measured non-invasively. These models and values, particularly if available breath-to-breath in real-time, could assist clinicians in the prescription or optimisation of CPAP therapy, including optimising PEEP settings. Establishing a method of extrapolating patient breathing effort from this data could also help in the prescription of ventilation therapy. It could provide another mode of feedback to clinicians on the efficacy of the CPAP therapy at given PEEP settings.

Finally, this dataset may also be used to assess correlation between wearable patch features and stoke volume, ejection fraction, cardiac output, cardiac index, or cardiac disease severity, etc.

Despite the careful data collection and post-processing, there are still a few known limitations. Note that since Mac-Lab data might not have been exported correctly for some segments within a recording. Identifying these segments should be straightforward since these RHC data are not in physiological plausible range. 

Release Notes

This is the initial release of data version 1.0.0.

Ethics

The study was conducted under a protocol reviewed and approved by the University of California, San Francisco Institutional Review Boards (IRB number: 16-20442 and the date of approval: December 20, 2016). Patients were recruited from the catheterization laboratory at the University of California, San Francisco and all patients provided written consent. 

Acknowledgements

This work was supported in part by National Heart, Lung and Blood Institute under Grant R01HL130619. (Liviu Klein and Omer T. Inan are co-senior authors.)

Conflicts of Interest

O. T. Inan is a co-founder and board member of Cardiosense, Inc., a company focusing on commercializing wearable patch technologies and artificial intelligence algorithms for cardiovascular disease monitoring and treatment.

References

1. T. Hurnanen et al., “Automated detection of atrial fibrillation based on time–frequency analysis of seismocardiograms,” IEEE J. Biomed. Health Inform., vol. 21, no. 5, pp. 1233–1241, Nov. 2016.
2. O. Lahdenoja et al., “Atrial fibrillation detection via accelerometer and gyroscope of a smartphone,” IEEE J. Biomed. Health Inform., vol. 22, no. 1, pp. 108–118, Apr. 2017.
3. C. Yang, et al., “Classification of aortic stenosis before and after transcatheter aortic valve replacement using cardio-mechanical modalities,” in Proc. 42nd Annu. Int. Conf. IEEE Eng. Med. Biol. Soc., Jul. 2020, pp. 2820–2823.
4. Z. Iftikhar et al., “Multiclass classifier based cardiovascular condition detection using smartphone mechanocardiography,” Sci. Rep., vol. 8, no. 1, pp. 1–14, 2018.
5. K. Pandia, et al., “Extracting respiratory information from seismocardiogram signals acquired on the chest using a miniature accelerometer,” Physiol. Meas., vol. 33, no. 10, 2012, Art. no. 1643.
6. D. M. Salerno and J. Zanetti, “Seismocardiography formonitoring changes in left ventricular function during ischemia,” Chest, vol. 100, no. 4, pp. 991–993, 1991.
7. A. Q. Javaid et al., “Quantification of posture induced changes in wearable seismocardiogram signals for heart failure patients,” in Proc. Comput. in Cardiol. Conf., Sep. 2016, pp. 777–780.
8. O. T. Inan et al., “Novel wearable seismocardiography and machine learning algorithms can assess clinical status of heart failure patients,” Circulation: Heart Failure, vol. 11, no. 1, 2018, Art. no. e004313. [Online]. Available: https://www.ahajournals.org/doi/ pdf/10.1161/CIRCHEARTFAILURE.117.004313?download=true
9. M. M. H. Shandhi et al., “Wearable patch based estimation of oxygen uptake and assessment of clinical status during cardiopulmonary exercise testing in patients with heart failure,” J. Cardiac Failure, vol. 26, pp. 948–958, 2020.
10. M. M. H. Shandhi et al., “Estimation of instantaneous oxygen uptake during exercise and daily activities using a wearable cardio-electromechanical and environmental sensor,” IEEE J. Biomed. Health Inform., vol. 25, no. 3, pp. 634–46, Jul. 2020.
11. E. L. Bonno, et al., “Modern right heart catheterization: Beyond simple hemodynamics,” Adv. Pulmonary Hypertension, vol. 19, no. 1, pp. 6–15, 2020.
12. P. Sorajja et al., “SCAI/HFSA clinical expert consensus document on the use of invasive hemodynamics for the diagnosis and management of cardiovascular disease,” Catheterization Cardiovasc. Interv.: Official J. Soc. for Cardiac Angiography Interv., vol. 89, no. 7, pp. E233–E247, 2017.
13. M. M. H. Shandhi, J. Fan, J. A. Heller, M. Etemadi, L. Klein, and O. T. Inan, “Estimation of Changes in Intracardiac Hemodynamics Using Wearable Seismocardiography and Machine Learning in Patients With Heart Failure: A Feasibility Study,” IEEE Transactions on Biomedical Engineering, vol. 69, no. 8, pp. 2443–2455, Aug. 2022, doi: 10.1109/TBME.2022.3147066.