2.7 Machine Learning on Central Hemodynamic Quantities Using Noninvasive Measurements: How Far Can We Go?
Background: Transforming peripheral noninvasive measurements to obtain central hemodynamic quantities, such as cardiac output (CO) and central systolic blood pressure (cSBP), is a highly emerging field [1,2]. However, no holistic investigation has been performed to assess the amount of information c...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2020-02-01
|
| Series: | Artery Research |
| Online Access: | https://www.atlantis-press.com/article/125934400/view |
| _version_ | 1828278317197819904 |
|---|---|
| author | Vasiliki Bikia Stamatia Pagoulatou Nikolaos Stergiopulos |
| author_facet | Vasiliki Bikia Stamatia Pagoulatou Nikolaos Stergiopulos |
| author_sort | Vasiliki Bikia |
| collection | DOAJ |
| description | Background: Transforming peripheral noninvasive measurements to obtain central hemodynamic quantities, such as cardiac output (CO) and central systolic blood pressure (cSBP), is a highly emerging field [1,2]. However, no holistic investigation has been performed to assess the amount of information contained in each peripheral measurement for the prediction of central values. This can be attributed to the inherent difficulty of creating a complete and accurate database; mainly due to the invasive nature of the gold standard techniques [3,4].
Methods: To meet this need, we exploit synthetic data from a previously validated cardiovascular model (CVm) [5]. Our study relies on peripheral quantities including brachial pressure, heart rate (HR), and pulse wave velocity (PWV) simulated by the CVm. A Random Forest model was trained using 2744 synthetic instances and, subsequently, was tested against a subset of 800. Correlations and feature importances of the input parameters were reported (Figure 1).
Results: Our results demonstrated that precise estimates of CO and cSBP were yielded with an RMSE of 0.39 L/min and 1.39 mmHg, respectively (Figures 2 and 3). Low biases were observed, namely 0.03 ± 0.39 L/min for CO and −0.08 ± 1.39 mmHg for cSBP. PWV, HR, and brachial pulse pressure were found to be the most correlated features with CO, whereas brachial SBP was plausibly shown to be the significant determinant of cSBP for our model (Figures 4 and 5).
Conclusion: These findings pave the way for better devising central hemodynamics’ predictions. In the future, our ultimate goal is to examine the sensitivity of cardiac parameters estimation (i.e., elastance) to noninvasive peripheral measurements.
Figure 1Correlation matrix.
Figure 2Scatterplot (A) and Bland-Altman plot (B) between predicted and reference CO values.
Figure 3Scatterplot (A) and Bland-Altman plot (B) between predicted and reference CO values.
Figure 4Feature importances for CO prediction.
Figure 5Feature importances for cSBP. |
| first_indexed | 2024-04-13T07:29:31Z |
| format | Article |
| id | doaj.art-c6f7630073a9437981f2f45d83be33bc |
| institution | Directory Open Access Journal |
| issn | 1876-4401 |
| language | English |
| last_indexed | 2024-04-13T07:29:31Z |
| publishDate | 2020-02-01 |
| publisher | BMC |
| record_format | Article |
| series | Artery Research |
| spelling | doaj.art-c6f7630073a9437981f2f45d83be33bc2022-12-22T02:56:24ZengBMCArtery Research1876-44012020-02-0125110.2991/artres.k.191224.0122.7 Machine Learning on Central Hemodynamic Quantities Using Noninvasive Measurements: How Far Can We Go?Vasiliki BikiaStamatia PagoulatouNikolaos StergiopulosBackground: Transforming peripheral noninvasive measurements to obtain central hemodynamic quantities, such as cardiac output (CO) and central systolic blood pressure (cSBP), is a highly emerging field [1,2]. However, no holistic investigation has been performed to assess the amount of information contained in each peripheral measurement for the prediction of central values. This can be attributed to the inherent difficulty of creating a complete and accurate database; mainly due to the invasive nature of the gold standard techniques [3,4]. Methods: To meet this need, we exploit synthetic data from a previously validated cardiovascular model (CVm) [5]. Our study relies on peripheral quantities including brachial pressure, heart rate (HR), and pulse wave velocity (PWV) simulated by the CVm. A Random Forest model was trained using 2744 synthetic instances and, subsequently, was tested against a subset of 800. Correlations and feature importances of the input parameters were reported (Figure 1). Results: Our results demonstrated that precise estimates of CO and cSBP were yielded with an RMSE of 0.39 L/min and 1.39 mmHg, respectively (Figures 2 and 3). Low biases were observed, namely 0.03 ± 0.39 L/min for CO and −0.08 ± 1.39 mmHg for cSBP. PWV, HR, and brachial pulse pressure were found to be the most correlated features with CO, whereas brachial SBP was plausibly shown to be the significant determinant of cSBP for our model (Figures 4 and 5). Conclusion: These findings pave the way for better devising central hemodynamics’ predictions. In the future, our ultimate goal is to examine the sensitivity of cardiac parameters estimation (i.e., elastance) to noninvasive peripheral measurements. Figure 1Correlation matrix. Figure 2Scatterplot (A) and Bland-Altman plot (B) between predicted and reference CO values. Figure 3Scatterplot (A) and Bland-Altman plot (B) between predicted and reference CO values. Figure 4Feature importances for CO prediction. Figure 5Feature importances for cSBP.https://www.atlantis-press.com/article/125934400/view |
| spellingShingle | Vasiliki Bikia Stamatia Pagoulatou Nikolaos Stergiopulos 2.7 Machine Learning on Central Hemodynamic Quantities Using Noninvasive Measurements: How Far Can We Go? Artery Research |
| title | 2.7 Machine Learning on Central Hemodynamic Quantities Using Noninvasive Measurements: How Far Can We Go? |
| title_full | 2.7 Machine Learning on Central Hemodynamic Quantities Using Noninvasive Measurements: How Far Can We Go? |
| title_fullStr | 2.7 Machine Learning on Central Hemodynamic Quantities Using Noninvasive Measurements: How Far Can We Go? |
| title_full_unstemmed | 2.7 Machine Learning on Central Hemodynamic Quantities Using Noninvasive Measurements: How Far Can We Go? |
| title_short | 2.7 Machine Learning on Central Hemodynamic Quantities Using Noninvasive Measurements: How Far Can We Go? |
| title_sort | 2 7 machine learning on central hemodynamic quantities using noninvasive measurements how far can we go |
| url | https://www.atlantis-press.com/article/125934400/view |
| work_keys_str_mv | AT vasilikibikia 27machinelearningoncentralhemodynamicquantitiesusingnoninvasivemeasurementshowfarcanwego AT stamatiapagoulatou 27machinelearningoncentralhemodynamicquantitiesusingnoninvasivemeasurementshowfarcanwego AT nikolaosstergiopulos 27machinelearningoncentralhemodynamicquantitiesusingnoninvasivemeasurementshowfarcanwego |