Although the stroke volume (SV) estimation by arterial blood pressure has been widely used in clinical practice, its accuracy is questionable, especially during periods of hemodynamic instability. We aimed to create novel SV estimating model based on deep-learning (DL) method. A convolutional neural network was applied to estimate SV from arterial blood pressure waveform data recorded from liver transplantation (LT) surgeries. The model was trained using a gold standard referential SV measured via pulmonary artery thermodilution method. Merging a gold standard SV and corresponding 10.24 seconds of arterial blood pressure waveform as an input/output data set with 2-senconds of sliding overlap, 484,384 data sets from 34 LT surgeries were used for training and validation of DL model. The performance of DL model was evaluated by correlation and concordance analyses in another 491,353 data sets from 31 LT surgeries. We also evaluated the performance of pre-existing commercialized model (EV1000), and the performance results of DL model and EV1000 were compared. The DL model provided an acceptable performance throughout the surgery (r = 0.813, concordance rate = 74.15%). During the reperfusion phase, where the most severe hemodynamic instability occurred, DL model showed superior correlation (0.861; 95% Confidence Interval, (CI), 0.855–0.866 vs. 0.570; 95% CI, 0.556–0.584, P < 0.001) and higher concordance rate (90.6% vs. 75.8%) over EV1000. In conclusion, the DL-based model was superior for estimating intraoperative SV and thus might guide physicians to precise intraoperative hemodynamic management. Moreover, the DL model seems to be particularly promising because it outperformed EV1000 in circumstance of rapid hemodynamic changes where physicians need most help.

Background: We aimed to create a novel model using a deep learning method to estimate stroke volume variation (SVV), a widely used predictor of fluid responsiveness, from arterial blood pressure waveform (ABPW). Methods: In total, 557 patients and 8,512,564 SVV datasets were collected and were divided into three groups: training, validation, and test. Data was composed of 10 s of ABPW and corresponding SVV data recorded every 2 s. We built a convolutional neural network (CNN) model to estimate SVV from the ABPW with pre-existing commercialized model (EV1000) as a reference. We applied pre-processing, multichannel, and dimension reduction to improve the CNN model with diversified inputs. Results: Our CNN model showed an acceptable performance with sample data (r = 0.91, MSE = 6.92). Diversification of inputs, such as normalization, frequency, and slope of ABPW significantly improved the model correlation (r = 0.95), lowered mean squared error (MSE = 2.13), and resulted in a high concordance rate (96.26%) with the SVV from the commercialized model. Conclusions: We developed a new CNN deep-learning model to estimate SVV. Our CNN model seems to be a viable alternative when the necessary medical device is not available, thereby allowing a wider range of application and resulting in optimal patient management.

Although visualization of heart sounds, known as phonocardiography, provides valuable information on cardiovascular hemodynamics, its use has not been widely encouraged due to the scarcity of information on its interpretation. In the present study, using the intraoperative phonocardiogram recorded by an esophageal stethoscope, we quantitatively evaluated the time and frequency domains of modulation of the heart sounds components and their association with left ventricular contractility and systemic vascular resistance under the effects of various cardiovascular drugs. We analyzed 29 pairs of intraoperative digitalized phonocardiographic signals and their corresponding hemodynamic data before and after cardiovascular drug administration (ephedrine, esmolol, phenylephrine, and/or nicardipine) in 17 patients who underwent liver transplantation. The S1 and S2 components of the heart sounds (the first and second heart sounds, respectively) were identified and their modulation in time and frequency domains was analyzed. As an index of cardiovascular function, systolic tissue Doppler wave velocity (TDI S’), maximal dP/dt from the arterial waveform, and systemic vascular resistance were simultaneously evaluated. Ephedrine/esmolol and phenylephrine/nicardipine primarily affected the S1 and S2 components of the heart sounds, respectively. This result implies that the intraoperative phonocardiogram may have the potential to be useful in detecting the changes in contractility and afterload that commonly occur in patients receiving anesthesia. In an era of constant need for noninvasive hemodynamic assessment, phonocardiography has the potential for use as a novel and informative tool for monitoring of hemodynamic function.

Purpose: To investigate the possibility of esophageal phonocardiography as a monitor for invasively measured pulse pressure (PP) and its respiratory variation (PPV) in patients undergoing liver transplantation. Methods: In 24 liver transplantation recipients, all hemodynamic parameters, including PP and PPV, were measured during five predetermined surgical phases. Simultaneously, signals of esophageal heart sounds (S1, S2) were identified, and S1–S2 interval (phonocardiographic systolic time, PST) and its respiratory variation (PSV) within a 20-s window were calculated. Beat-to-beat correlation between PP and its corresponding PST was assessed during each time window, according to the surgical phases. To compare PPV and PSV along with 5 phases (a total of 120 data pairs), Pearson correlation was conducted. Results: Beat-to-beat PST values were closely correlated with their corresponding 3360 pairs of PP values (median r = 0.568 [IQR 0.246–0.803]). Compared with the initial phase of surgery, correlation coefficients were significantly lower during the reperfusion period (median r = 0.717 [IQR 0.532–0.886] vs. median r = 0.346 [IQR 0.037–0.677]; p = 0.002). The correlation between PSV and PPV showed similar variation according to the surgical phases (r = 0.576 to 0.689, p < 0.05, for pre-reperfusion; 0.290 to 0.429 for the post-reperfusion period). Conclusions: Continuous monitoring of intraoperative PST with an esophageal stethoscope has the potential to act as an indirect estimator of beat-to-beat arterial PP. Moreover, PSV appears to exhibit a trend similar to that of PPV with moderate accuracy. However, variation according to the surgical phase limits the merit of the current results, thereby necessitating cautious interpretation.

Although intravenous administration of contrast media may trigger a variety of adverse reactions, sedated patients undergoing computed tomography (CT) scanning usually are not able to report their symptoms, which may delay detection of adverse reactions. Furthermore, changes in vital signs cannot be typically measured during mobile CT scanning, which worsens the situation. We aimed to characterize contrast-related hemodynamic changes that occur during mobile CT scanning and predict sudden hypotension based on subtle but robust changes in the electrocardiogram (ECG). We analyzed the digitized hemodynamic data of 20 consecutive patients who underwent clipping of a cerebral artery aneurysm and contrast-enhanced CT scanning following the surgical procedure. Hemodynamic variables, including ECG findings, invasive blood pressure (BP), pulse oximetry results, capnography findings, cardiac output, and systemic vascular resistance, were monitored simultaneously. We measured morphological changes in ECG-derived parameters, including the R–R interval, ST height, and QRS R-amplitude, on a beat-to-beat basis, and evaluated the correlation between those parameters and hemodynamic changes. After the radiocontrast injection, systolic BP decreased by a median 53 mmHg from baseline and spontaneously recovered after 63 ± 19 s. An increase in QRS R-amplitude (median 0.43 mV) occurred 25 ± 10 s before hypotension developed. The receiver operating characteristic curve showed that a 16% increase in QRS R-amplitude can predict a decrease in systolic BP of >25% (area under the curve 0.852). Increased cardiac output (median delta 2.7 L/min from baseline) and decreased systemic vascular resistance (median delta 857 dyn·s/cm⁵ from baseline) were also observed during hypotension. During mobile CT scanning, profound but transient hypotension can be observed, associated with decreased vascular resistance. Augmentation of QRS R-amplitude from an ECG represents a sensitive surrogate for onset of a hypotensive episode after contrast injection, thereby serving as a simple and continuous noninvasive hemodynamic monitoring tool.