Automatic quantitative measurement of left atrial pressure using mitral regurgitation spectrum：Clinical study on comparison with floating cathete

doi:10.21203/rs.3.rs-1996717/v1

Download PDF

Research Article

Automatic quantitative measurement of left atrial pressure using mitral regurgitation spectrum：Clinical study on comparison with floating cathete

https://doi.org/10.21203/rs.3.rs-1996717/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

To explore how to measure LAP_Ep accurately and quantitatively, that is, the left atrial pressure (LAP) measured and calculated by equation method using mitral regurgitation spectrum.

Methods

The mitral regurgitation spectrum, pulmonary arteriolar wedge pressure (PAWP) and invasive arterial systolic pressure of radial artery of 28 patients were collected simultaneously, including 3 patients with rheumatic heart disease, 15 patients with mitral valve prolapse and 10 patients with coronary artery bypass grafting, patients with moderate or above aortic stenosis were excluded. LAP_Bp(Doppler sphygmomanometer method), LAP_Eq (Equation method) and LAP_C (Catheter method) were measured synchronously, and the measurement results of the three methods were compared and analyzed. We designed a special intelligent Doppler spectrum analysis software to accurately measure LAP_Eq.

Results

Found that there was no significant statistical difference between the measurement results of LAP_C and LAP_Eq (t = 0.954, P = 0.348),and significant correlation between the two methods [r = 0.908(0.844,0.964), P < 0.001]. Although the measurement results of LAP_C and LAP_BP are consistent in some patients, there are significant differences in the overall case and weak correlation between the two methods [r = 0.210, (−0.101, 0.510), P = 0.090]. In MVP patients with P1 or P3 prolapse, the peak pressure difference of MR was underestimated due to the serious eccentricity of MR, which affected the accuracy of LAP_BP measurement. This study had been approved by the ethics committee of the northern theater general hospital ( K-2019-17), applied prospectively for clinical trial (No. Chictr 190023812) and invention patent application (No. 202210711862.X).

Conclusions

This study shows that there is a good correlation between LAP_Eq and LAP_C, which verifies that the noninvasive and direct quantitative measurement of left atrial pressure based on mitral regurgitation spectrum is feasible and has a good application prospect.

Left atrial pressure

Pulmonary arteriole wedge pressure

Mitral regurgitation spectrum

Deep learning model of big data training

Left atrial pressure (LAP) is a reliable data reflecting the left ventricular preload, which can correctly reflect the change of blood volume and sensitively reflect the left ventricular end diastolic pressure. It is an important hemodynamic parameter for adjusting the treatment plan of critically ill patients[1]. However, it is almost impossible to directly measure left atrial pressure in clinical practice. At present, pulmonary arteriole wedge pressure (PAWP) measured by floating catheter is used to replace LAP_C in clinic. However, the measurement of PAWP is also invasive, and its clinical application is greatly limited[2.3].

Different degrees of mitral regurgitation (MR) often occurred in normal subjects and patients [4]. At present, there are two methods to quantitatively measure LAP according to mitral regurgitation spectrum. The first method is the "Doppler + sphygmomanometer" method [5.6], that is, the LAP_BP is the left ventricular systolic pressure (P) minus the mitral regurgitation pressure difference, LAP_BP = P-ΔP. If there is no aortic valve and/or left ventricular outflow tract stenosis, the arterial systolic pressure can replace the left ventricular systolic pressure. The second method is proposed by our team. According to Weiss exponential equation (P=e-t/τ+B), Simplified Bernoulli equation (ΔP=4V²) and P=ΔP+LAP, Bai [7-10] used mathematical methods to derive “binary linear equations”, which have 2 variables, namely the left ventricular relaxation time constant (τ)and LAP, and LAP_Ep is measured and calculated by using the decline curve of mitral regurgitation spectrum. The calculation ofτhas been verified by animal experiments [11]. In animal experiments, our team has confirmed that there is a good correlation LAP_Ep and LAP_C (Catheter measurement) [12].

If a non-invasive, convenient and accurate LAP quantitative measurement method can be developed, the diagnosis and treatment of heart related diseases will enter an accurate stage, which has important clinical value. We evaluated the accuracy of two methods of quantitative measurement of left atrial pressure using mitral regurgitation spectrum, and discussed how to measure LAP noninvasively, accurately and quantitatively.

Study design and patients

Patients who needed cardiac surgery accompanied with mitral regurgitation were prospectively selected as the research objects, from June 2020 to Oct 2020. The mitral regurgitation spectrum, pulmonary arteriolar wedge pressure (PAWP) and invasive arterial systolic pressure of radial artery of 28 patients were collected simultaneously, 13 males and 15 females, including 3 patients with rheumatic heart disease, 15 patients with mitral valve prolapse and 10 patients with coronary artery bypass grafting, patients with moderate or above aortic stenosis were excluded, aged 47-78 years (62.55±7.28 years) and PAWP 7-29mmHg (15.3±4.9mmHg).

Floating catheter

With the assistance of intravenous anesthesia, endotracheal intubation and ventilator, the floating catheter was placed through the jugular vein and the arterial systolic pressure was measured by radial artery puncture. The position of the floating catheter was determined by transthoracic ultrasound. PAWP measured by floating catheter was used to replace LAP_C. The a wave after the P wave of ECG is generated by the active contraction of left atrium, and the C wave is generated by the closure of mitral valve. The V wave after ECG T wave is generated by left ventricular relaxation and left atrial passive filling during mitral valve opening (the pressure generated by this wave cannot be used as left atrial pressure). Therefore, we take the pressure measured on the PAWP pressure curve at the end of ECG P wave as LAP_C. The equipment used include: Fl-005 GE anesthesia monitor (GE Healthcare Finland), Edwards 131 F7 floating catheter (Irvine, USA), PTC-6F pressure monitoring catheter (Jingzhou Yihai Technology Co., Ltd.), etc.

Echocardiography

The patient was in supine position because of perioperative period. All ultrasound examinations were performed by the same echocardiographic doctors with a Philips ultrasound system (Philips iE33 ultrasound machine; Philips Healthcare, Andover Mass). The mitral regurgitation spectrum was collected under CW, and the angle between the sampling line and the mitral regurgitation beam should be less than 15°. Select different recording speeds of 100mm / s or 150mm / s according to the speed of heart rate to obtain a dull, smooth and complete spectrum.

Formulas of LAP_Ep

According to Weiss exponential equation and simplified Bernoulli equation, the left ventricular relaxation time constant (τ) can be obtained,τ=P/(-dP/dt), Where P is the pressure in the left ventricle, and t is the time from -dp/dtmax, as shown in Figure 1. Bring in the intervals between different speeds to obtain the following formula,τ=(t1-t2)/ln((LAP+16)/(LAP+4)) andτ=(t1-t3)/ln((LAP+36)/(LAP+4)). Theoretically, LAP can be calculated by measuring the intervals between any two speeds. In order to facilitate calculation and measurement of , we selected the time t1, t2 and t3 when the descending branch velocity of mitral regurgitation spectrum was 1m/s, 2m/s and 3m/s respectively.

Measurement methods of LAP_Ep

In order to accurately measure the intervals, t1-t2 and t1-t3, it is necessary to detect the spectrum edge firstly, so we propose an intelligent method based on deep learning to complete this task. The method consists of two parts, a basic network for coarse detection and a post-processing module for refining. We adopt BCD-Unet deep learning model, which was proposed at the ICCV conference in 2019 [13], for edge detection firstly, but there are dislocation and fracture in the detection results. So we design a post-processing module to deal with these problems. The post-processing module mainly uses the polynomial fitting method to refine the edges detected by BCD-Unet, making them clearer and smoother. The overall structure of the proposed method is shown as Figure 2.

We use the data collected by the hospital to train the model Iteratively. The trained deep learning model can automatically detect the edge of mitral regurgitation spectrum. Then we encapsulated the model and built the system based on it. The system takes the mitral regurgitation spectrum as input and outputs the edge curve and LAP(directly a number), which is shown as Figure 3. The software can only measure and calculate when the descending branch of the mitral regurgitation spectrum curve is complete, and the curve between at least 1m / s and 3m / S is good. When the peak value of the curve is less than 3m / s, it will not be calculated.

Statistical analysis

The statistical analysis and data visualization were conducted by SPSS 26.0 statistical software. Paired t-test was used to compare and analyze the measurement results of LAP_BP and LAP_Eq with LAP_C method. Meanwhile, correlation analysis was performed on the measurement results of LAP_BP and LAP_Eq with LAP_C method. The difference between LAP_BP and LAP_C was less than 10%, which was defined as the consistency between LAP_BPand LAP_C, otherwise it was inconsistent. 28 patients were divided into two groups, 17 in the consistent group and 11 in the inconsistent group. The causes of inaccurate LAP_BP measurement were analyzed by single factor analysis.The signiﬁcant level is 0.05.

1. Paired t-test of the measurement results of LAP_Eq and LAP_BPwith LAP_Cmethod

A total of 95 mitral regurgitation spectra were obtained in 28 patients. LAP_C, LAP_Eq and LAP_BPmeasured synchronously in 28 patients were visualized, as shown in Figure 4. The average difference between LAP_C and LAP_Eq was 0.353, and the 95% confidence interval was (- 1.112, 0.406). Paired t-test found no significant statistical difference between the measurement results of LAP_C and LAP_Eq (t = 0.954, P = 0.348). The average difference between LAP_C and LAP_BP was 3.332, and the 95% confidence interval was (- 5.577, -1.087). Paired t-test found no significant statistical difference between the measurement results of LAP_C and LAP_BP (t = 3.045, P = 0.005). Although the measurement results of LAP_C and LAP_BP are consistent in some patients, there are significant differences between the two methods in the overall case. As shown in Figure 4.

2. Correlation analysis of the measurement results of LAP_BP and LAP_Eq with LAP_C method

The correlation analysis of LAP_C and LAP_Eq shows that the results measured by the two methods are highly correlated and have significant statistical significance [r=0.908(0.844,0.964), P < 0.001]. The correlation analysis of LAP_C and LAP_BP shows that the results measured by the two methods show a weak correlation, but they do not have significant statistical significance[r=0.210, (-0.101, 0.510), P=0.090]. As shown in Figure 5.

3. Analysis of the difference between LAP_BP and LAP_C measurement

Whether atrial fibrillation or not had no significant effect on the measurement results of the two methods, P > 0.05; The analysis of disease composition found that the consistency rate of the two methods in patients with coronary heart disease and rheumatic heart disease without aortic stenosis was significantly higher than that in patients with mitral valve prolapse, P=0.002. As shown in Table 1. In 11 MVP patients with P1 or P3 prolapse, the MR peak differential pressure was underestimated due to the MR severe eccentric direction. Seven MVP patients with A2 or P2 prolapse could accurately obtain the MR peak pressure difference and quantitatively measure LAP_BP.

Table 1 Analysis of inaccurate LAP_BP measurement

testvar		Consistent (n=17)	Inconsistent (n=11)	statistic	p
Disease composition	CABG & RHD	10(58.82)	0(0)	7.381	0.002
	MVP	7(41.18)	11(100)
Radial artery systolic pressure.mmHg.		111(101,119)	102(96.5,107.5)	1.248	0.212
MR Peak velocity. m/s.		4.9(4.6,5)	4.5(4.25,4.72)	1.677	0.094
Sinus rhythm		11(64.71%)	5(45.45%)	0.452	0.441
Atrial fibrillation		6(35.29%)	6(54.55%)

This study is a methodological study of LAP quantitative measurement. The advantages and disadvantages of two methods of LAP quantitative measurement using mitral regurgitation spectrum are analyzed. The measurement results of LAP_Ep method proposed by our team and LAP_C method have good correlation. LAP_BP method only has good correlation with LAP_Cmethod in some types of patients.

Accurate assessment of LAP is an important determinant in the formulation of clinical treatment plan, so floating catheter must be used in many critically ill patients to replace LAP with PAWP. However, in most cases, the assessment of LAP adopts qualitative or semi quantitative non-invasive methods. The combined application of tissue Doppler at atrioventricular valve annulus, mitral anterior flow spectrum and pulmonary vein spectrum can be used to qualitatively or semi quantitatively evaluate LAP [14.15], and can also be used to quantitatively estimate LAP [16], but the correlation with LAP_Cis only moderate [17.18].

LAP_BP measured by "Doppler + sphygmomanometer" method based on mitral regurgitation spectrum is a practical quantitative measurement method of left atrial pressure. However, it is easily affected by the hemodynamics of aortic valve and mitral valve, with poor repeatability and complex operation [5].We carefully analyzed each patient and found that LAP_BP method has a good correlation with LAP_C method in some types of patients. The LAP_BP method has clinical value in patients with central mitral regurgitation.

We deduce the calculated LAP according to Weiss exponential equation and simplified Bernoulli equation. In theory, it is a method of real-time quantitative measurement of LAP. To verify the correlation between LAP_Eq and LAP_C, the mitral regurgitation spectrum (TTE) and PAWP were collected simultaneously when the floating catheter had been placed in cardiac surgery patients before thoracotomy. The LAP_C of our subjects is between 8-29mmhg. Paired t- test showed that there was no significant statistical difference between LAP_Eqand LAP_C(t = 0.954, P = 0.348). There was a high correlation between the results of the two measurement methods (r=0.908, P＜0.001). It shows that LAP_Eq can quantitatively measure LAP_Eq, that is, two different intervals of the descending branch of the mitral regurgitation spectrum can be brought into the equation calculation, and the specific value of left atrial pressure can be obtained. This method of measuring left atrial pressure is accurate.

Animal experiments and clinical trials have proved that the quantitative measurement of left atrial pressure can be carried out by using mathematical equation. Our theoretical derivation is reasonable. However, there are two preconditions: One, a certain amount of mitral regurgitation is needed to measure left atrial pressure by this method; Second, it is necessary to accurately outline the edge of mitral regurgitation spectrum, which is the key to improve the repeatability of LAP_Eq measurement.

We measured LAP_Eq and LAP_BP for the same mitral regurgitation spectrum. Compared with LAP_C method, it was found that LAP_BP method was easily affected by the eccentricity of mitral regurgitation and underestimated the MR peak pressure difference. LAP_Eqmethod measures the decrease rate of left ventricular pressure, whether the mitral regurgitation spectrum is eccentric or not has little effect on the accuracy of this method. LAP_Eqmethod can accurately measure left atrial pressure.

In this study, the deep learning model of big data training is adopted, and the software with the ability of automatic tracking envelope and automatic calculation is established to improve the repeatability of LAP_Eq measurement. For cases with relatively large mitral regurgitation, the software shows good repeatability; For the cases with relatively few mitral regurgitation, we first sketch the edge manually, then draw the curve of mitral regurgitation spectrum with software, and measure and calculate the LAP, so as to improve the repeatability of this measurement method.

The method we used was the quantitative measurement of left atrial pressure directly, while the floating catheter method measured is PAWP. In some cases, PAWP is not equal to left atrial pressure, such as mechanical ventilation with Peep, endotoxin shock[19], pulmonary embolism[20], ARDS, etc. Therefore, the LAP_Ep method based on the descending branch of mitral regurgitation spectrum is a direct quantitative measurement method of left atrial pressure with good repeatability, which is worthy of further promotion.

Limitations

The application software used in LAP_Ep measurement in this study can only identify the Doppler spectrum of transthoracic echocardiography, but can not identify the Doppler spectrum of transesophageal echocardiography.

This study had been approved by the ethics committee of the northern theater general hospital ( K-2019-17), applied prospectively for clinical trial (No. Chictr 190023812) and invention patent application (No. 202210711862.X). All authors have contributed significantly, and all authors are in agreement with the content of the manuscript. Availability of data and materials had consent for publication. The authors declare that they have no competing interests. .Funding: This work was supported by the National Natural Science Foundation of China (No. 81771833, A study on the method of noninvasive measuring left atrial pressure by using the continuous Doppler spectrum mitral regurgitation). YJ and FY collected and measured the initial data of mitral regurgitation spectrum method and catheter method. YJ was one of major contributors in writing the manuscript. CW was another major contributor in writing the manuscript. HW was the subject designer and the main reviewer. LY and KM had designed proprietary application software to achieve accurate measurement of Doppler spectrum. FX was responsible for statistical analysis of data. All authors read and approved the final manuscript.

Acknowledge

Not applicable

Thomas L, Abhayaratna WP. Left Atrial Reverse Remodeling: Mechanisms, Evaluation, and Clinical Significance. JACC Cardiovasc Imaging 2017;10:65–77.
Liu M, Fang F, Yu C M. Noncardiac comorbidities in heart failure with preserved ejection fraction - commonly ignored fact[J]. Circ J, 2015, 79(5): 954–959.
Nadruz W, Shah A M, Solomon S D. Diastolic Dysfunction and Hypertension[J]. Med Clin North Am, 2017, 101(1): 7–17.
Obase K, Weinert L, Hollatz A, Farooqui F, Roberts JD, Minhaj MM, Tung A, Chaney M, Ota T, Jeevanandam V, Yoshida K, Mor-Avi V, Lang RM. Elongation of chordae tendineae as an adaptive process to reduce mitral regurgitation in functional mitral regurgitation. Eur Heart J Cardiovasc Imaging. 2016 May;17(5):500–9. doi: 10.1093/ehjci/jev314.
Nagueh S F. Noninvasive estimation of LV filling pressures in heart failure and reduced ejection fraction: revisited and verified[J]. JACC Cardiovasc Imaging, 2011, 4(9): 935–937.
Pozzoli M. Non-invasive evaluation of the hemodynamic profile in patients with heart failure: estimation of left atrial pressure. Ital Heart J Suppl. 2000 Oct;1(10):1326–33. Italian. PMID: 11068716.
Bai XF, Ma AX. Symmetry of cardiac function assessment. J Geriatr Cardiol 2016;13:517–20.
Bai X, Wang Q. Time constants of cardiac function and their calculations. Open Cardiovasc Med J 2010;4:168–72.
Bai X. Calculation of left ventricular relaxation time constant-T au in humans by continuous-wave Doppler. Open Cardiovasc Med J 2008;2:9–11.
Bai X, Wen C. Current situation and prospects of left ventricular diastolic time constant (Tau) non-invasive measurement. Cin J Med Imaging Technol 2015; 10:1596–1600.
Wen C, Sun J, Fan C, et al. Calculation of Left V entricular Diastolic Time Constant (TAU) in Dogs with Mitral Regurgitation Using Continuous-Wave Doppler. Ultrasound Med Biol 2018;44:1778–85.
Zheng HN, Jin Y, Fu YW, Wang XQ, Wen CY. Experimental study: using the continuous wave Doppler spectrum technique to detect left atrial pressure. Ann Transl Med 2022;10(2):105.
R. Azad, M. Asadi-Aghbolaghi, M. Fathy and S. Escalera, Bi-Directional ConvLSTM U-Net with Densley Connected Convolutions[J]. arXiv:1909.00166v1 [eess.IV] 31 Aug 2019.
Smallhorn J F, Freedom R M, Olley P M. Pulsed Doppler echocardiographic assessment of extraparenchymal pulmonary vein flow[J]. J Am Coll Cardiol, 1987, 9(3): 573–579.
Geske J B, Sorajja P, Nishimura R A, et al. Evaluation of left ventricular filling pressures by Doppler echocardiography in patients with hypertrophic cardiomyopathy: correlation with direct left atrial pressure measurement at cardiac catheterization[J]. Circulation, 2007, 116(23): 2702–2708.
Cowie B, Kluger R, Rex S, et al. Noninvasive estimation of left atrial pressure with transesophageal echocardiography[J]. Ann Card Anaesth, 2015, 18(3): 312–316.
Figueras-Coll M, Sanchez-De-Toledo J, Gran F, et al. Echocardiography in the Assessment of Left Atrial Pressure After Pediatric Heart Surgery: A Comparison Study With Measurements Obtained From Left Atrial Catheter[J]. World J Pediatr Congenit Heart Surg, 2015, 6(3): 438–442.
Pozzoli M. Non-invasive evaluation of the hemodynamic profile in patients with heart failure: estimation of left atrial pressure. Ital Heart J Suppl. 2000 Oct;1(10):1326–33. Italian. PMID: 11068716.
Krahmer RL, Fang HK, Vitello J, Rypins EB, Law WR. Pulmonary capillary wedge pressure estimates of left ventricular preload are inaccurate in endotoxin shock: contribution of Starling resistor forces to septic pulmonary hypertension. Shock. 1994 Nov;2(5):344 – 50. doi: 10.1097/00024382-199411000-00008. PMID: 7743360.
Entress JJ, Dhamee MS, Olund T, Aggarwal A, Hopwood M, Olinger GN. Pulmonary artery occlusion pressure is not accurate immediately after cardiopulmonary bypass. J Cardiothorac Anesth. 1990 Oct;4(5):558 – 63. doi: 10.1016/0888-6296(90)90404-4. PMID: 2132134.

No competing interests reported.

Download PDF

Editorial decision: Major revision
12 Sep, 2022
Reviews received at journal
12 Sep, 2022
Reviewers agreed at journal
09 Sep, 2022
Reviewers invited by journal
06 Sep, 2022
Editor assigned by journal
06 Sep, 2022
Submission checks completed at journal
05 Sep, 2022
First submitted to journal
25 Aug, 2022

You are reading this latest preprint version

Automatic quantitative measurement of left atrial pressure using mitral regurgitation spectrum：Clinical study on comparison with floating cathete

Status:

Version 1

Abstract

Figures

Introduction

Methods

Results

Discussion

Conclusions

Declarations

References

Additional Declarations

Status:

Version 1