{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,7]],"date-time":"2026-08-07T17:23:17Z","timestamp":1786123397171,"version":"build-2736575974"},"reference-count":30,"publisher":"Ovid Technologies (Wolters Kluwer Health)","issue":"1","content-domain":{"domain":["lww.com","ovid.com"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2025,1]]},"abstract":"<jats:sec>\n                    <jats:title>BACKGROUND:<\/jats:title>\n                    <jats:p>Due to their invasiveness, arterial lines are not typically used in routine monitoring, despite their superior responsiveness in hemodynamic monitoring and detecting intraoperative hypotension. To address this issue, noninvasive, continuous arterial pressure monitoring is necessary. We developed a deep-learning model that reconstructs continuous mean arterial pressure (MAP) using the photoplethysmograhy (PPG) signal and compared it to the arterial line gold standard.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>METHODS:<\/jats:title>\n                    <jats:p>We analyzed high-frequency PPG signals from 117 patients in neuroradiology and digestive surgery with a median of 2201 (interquartile range [IQR], 788\u20134775) measurements per patient. We compared models with different combinations of convolutional and recurrent layers using as inputs for our neural network high-frequency PPG and derived features including dicrotic notch relative amplitude, perfusion index, and heart rate. Mean absolute error (MAE) was used as performance metrics. Explainability of the deep-learning model was reconstructed with Grad-CAM, a visualization technique using saliency maps to highlight the parts of an input that are significant for a deep-learning model decision-making process.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>RESULTS:<\/jats:title>\n                    <jats:p>An MAP baseline model, which consisted only of standard cuff measures, reached an MAE of 6.1 (\u00b1 14.5) mm Hg. In contrast, the deep-learning model achieved an MAE of 3.5 (\u00b1 4.4) mm Hg on the external test set (a 42.6% improvement). This model also achieved the narrowest confidence intervals and met international standards used within the community (grade A). The saliency map revealed that the deep-learning model primarily extracts information near the dicrotic notch region.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>CONCLUSIONS:<\/jats:title>\n                    <jats:p>Our deep-learning model noninvasively estimates arterial pressure with high accuracy. This model may show potential as a decision-support tool in operating-room settings, particularly in scenarios where invasive blood pressure monitoring is unavailable.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1213\/ane.0000000000007295","type":"journal-article","created":{"date-parts":[[2024,12,16]],"date-time":"2024-12-16T18:00:24Z","timestamp":1734372024000},"page":"119-128","update-policy":"https:\/\/doi.org\/10.1097\/lww.0000000000001000","source":"Crossref","is-referenced-by-count":6,"title":["Blood Pressure Estimation Using Explainable Deep-Learning Models Based on Photoplethysmography"],"prefix":"10.1213","volume":"140","author":[{"given":"Jade","family":"Perdereau","sequence":"first","affiliation":[{"name":"Universit\u00e9 Paris Cit\u00e9, INSERM UMRS 942 (MASCOT), Paris, France"},{"name":"Entrep\u00f4t de donn\u00e9es de sant\u00e9, Assistance Publique H\u00f4pitaux de Paris, Paris, France"},{"name":"Department of Anesthesia and Critical Care, Lariboisi\u00e8re Hospital, APHP, Paris, France"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Thibaut","family":"Chamoux","sequence":"additional","affiliation":[{"name":"Entrep\u00f4t de donn\u00e9es de sant\u00e9, Assistance Publique H\u00f4pitaux de Paris, Paris, France"},{"name":"Department of Anesthesia and Critical Care, Lariboisi\u00e8re Hospital, APHP, Paris, France"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Etienne","family":"Gayat","sequence":"additional","affiliation":[{"name":"Universit\u00e9 Paris Cit\u00e9, INSERM UMRS 942 (MASCOT), Paris, France"},{"name":"Department of Anesthesia and Critical Care, Lariboisi\u00e8re Hospital, APHP, Paris, France"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Arthur","family":"Le Gall","sequence":"additional","affiliation":[{"name":"Department of Anesthesia and Critical Care, Lariboisi\u00e8re Hospital, APHP, Paris, France"},{"name":"Inria, France"},{"name":"Department of Anaesthesia, Critical Care and Peri-operative Medicine, Rennes University Hospital, Rennes, France."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Fabrice","family":"Vall\u00e9e","sequence":"additional","affiliation":[{"name":"Universit\u00e9 Paris Cit\u00e9, INSERM UMRS 942 (MASCOT), Paris, France"},{"name":"Department of Anesthesia and Critical Care, Lariboisi\u00e8re Hospital, APHP, Paris, France"},{"name":"Inria, France"},{"name":"LMS, CNRS, Ecole Polytechnique, Institut Polytechnique de Paris, Paris, France"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"J\u00e9r\u00f4me","family":"Cartailler","sequence":"additional","affiliation":[{"name":"Universit\u00e9 Paris Cit\u00e9, INSERM UMRS 942 (MASCOT), Paris, France"},{"name":"Department of Anesthesia and Critical Care, Lariboisi\u00e8re Hospital, APHP, Paris, France"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jona","family":"Joachim","sequence":"additional","affiliation":[{"name":"Universit\u00e9 Paris Cit\u00e9, INSERM UMRS 942 (MASCOT), Paris, France"},{"name":"Department of Anesthesia and Critical Care, Lariboisi\u00e8re Hospital, APHP, Paris, France"},{"name":"Inria, France"},{"name":"LMS, CNRS, Ecole Polytechnique, Institut Polytechnique de Paris, Paris, France"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"276","published-online":{"date-parts":[[2024,12,16]]},"reference":[{"key":"R1-20250224","doi-asserted-by":"crossref","first-page":"34","DOI":"10.1097\/CCM.0b013e318265ea46","article-title":"Methods of blood pressure measurement in the ICU.","volume":"41","author":"Lehman","year":"2013","journal-title":"Crit Care Med"},{"key":"R2-20250224","doi-asserted-by":"crossref","first-page":"127","DOI":"10.1016\/j.bpa.2019.05.001","article-title":"Perioperative blood pressure monitoring.","volume":"33","author":"Roach","year":"2019","journal-title":"Best Pract Res Clin Anaesthesiol"},{"key":"R3-20250224","first-page":"85","article-title":"Complications of invasive monitoring.","volume":"15","author":"Murray","year":"1981","journal-title":"Med Instrum"},{"key":"R4-20250224","doi-asserted-by":"crossref","first-page":"75","DOI":"10.1016\/0011-3840(88)90009-3","article-title":"Complications of invasive hemodynamic monitoring in the intensive care unit.","volume":"25","author":"Sladen","year":"1988","journal-title":"Curr Probl Surg"},{"key":"R5-20250224","doi-asserted-by":"crossref","first-page":"637","DOI":"10.1007\/s00101-019-0614-y","article-title":"[Principles and pitfalls of arterial blood pressure measurement].","volume":"68","author":"Meidert","year":"2019","journal-title":"Anaesthesist"},{"key":"R6-20250224","doi-asserted-by":"crossref","first-page":"1195","DOI":"10.1097\/ALN.0b013e31825683dc","article-title":"Hospital stay and mortality are increased in patients having a \u201ctriple low\u201d of low blood pressure, low bispectral index, and low minimum alveolar concentration of volatile anesthesia.","volume":"116","author":"Sessler","year":"2012","journal-title":"Anesthesiology"},{"key":"R7-20250224","doi-asserted-by":"crossref","first-page":"213","DOI":"10.1097\/01.anes.0000270724.40897.8e","article-title":"Incidence of intraoperative hypotension as a function of the chosen definition: literature definitions applied to a retrospective cohort using automated data collection.","volume":"107","author":"Bijker","year":"2007","journal-title":"Anesthesiology"},{"key":"R8-20250224","doi-asserted-by":"crossref","first-page":"507","DOI":"10.1097\/ALN.0b013e3182a10e26","article-title":"Relationship between intraoperative mean arterial pressure and clinical outcomes after noncardiac surgery: toward an empirical definition of hypotension.","volume":"119","author":"Walsh","year":"2013","journal-title":"Anesthesiology"},{"key":"R9-20250224","doi-asserted-by":"crossref","first-page":"706","DOI":"10.1016\/j.bja.2018.04.036","article-title":"Intraoperative hypotension and the risk of postoperative adverse outcomes: a systematic review.","volume":"121","author":"Wesselink","year":"2018","journal-title":"Br J Anaesth"},{"key":"R10-20250224","doi-asserted-by":"crossref","first-page":"1346","DOI":"10.1001\/jama.2017.14172","article-title":"Effect of individualized vs Standard blood pressure management strategies on postoperative organ dysfunction among high-risk patients undergoing major surgery: a randomized clinical trial.","volume":"318","author":"Futier","year":"2017","journal-title":"JAMA"},{"key":"R11-20250224","doi-asserted-by":"crossref","first-page":"R1","DOI":"10.1088\/0967-3334\/28\/3\/R01","article-title":"Photoplethysmography and its application in clinical physiological measurement.","volume":"28","author":"Allen","year":"2007","journal-title":"Physiol Meas"},{"key":"R12-20250224","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1016\/j.cmpb.2018.01.012","article-title":"Intradialytic hypotension related episodes identification based on the most effective features of photoplethysmography signal.","volume":"157","author":"Nafisi","year":"2018","journal-title":"Comput Methods Programs Biomed"},{"key":"R13-20250224","doi-asserted-by":"crossref","first-page":"605","DOI":"10.1016\/j.bja.2019.01.037","article-title":"Non-invasive continuous detection of arterial hypotension during induction of anaesthesia using a photoplethysmographic signal: proof of concept.","volume":"122","author":"Coutrot","year":"2019","journal-title":"Br J Anaesth"},{"key":"R14-20250224","doi-asserted-by":"crossref","first-page":"395","DOI":"10.1007\/s10877-020-00486-y","article-title":"Real-time estimation of mean arterial blood pressure based on photoplethysmography dicrotic notch and perfusion index. A pilot study.","volume":"35","author":"Joachim","year":"2020","journal-title":"J Clin Monit Comput"},{"key":"R15-20250224","doi-asserted-by":"crossref","first-page":"808","DOI":"10.1016\/j.bja.2020.12.035","article-title":"Deep learning models for the prediction of intraoperative hypotension.","volume":"126","author":"Lee","year":"2021","journal-title":"Br J Anaesth"},{"key":"R16-20250224","doi-asserted-by":"crossref","first-page":"e3533","DOI":"10.1002\/cnm.3533","article-title":"A method of parameter estimation for cardiovascular hemodynamics based on deep learning and its application to personalize a reduced-order model.","volume":"38","author":"Zhou","year":"2022","journal-title":"Int J Num Methods Biomed Eng"},{"key":"R17-20250224","doi-asserted-by":"crossref","first-page":"688","DOI":"10.1016\/j.bja.2019.07.025","article-title":"Deep-learning model for predicting 30-day postoperative mortality.","volume":"123","author":"Fritz","year":"2019","journal-title":"Br J Anaesth"},{"key":"R18-20250224","doi-asserted-by":"crossref","first-page":"30","DOI":"10.1109\/TASL.2011.2134090","article-title":"Context-dependent pre-trained deep neural networks for large-vocabulary speech recognition.","volume":"20","author":"Dahl","year":"2012","journal-title":"IEEE Trans Audio Speech Lang Process"},{"key":"R19-20250224","doi-asserted-by":"crossref","first-page":"1211","DOI":"10.1213\/ANE.0000000000004651","article-title":"Predicting deep hypnotic state from sleep brain rhythms using deep learning: a data-repurposing approach.","volume":"130","author":"Belur Nagaraj","year":"2020","journal-title":"Anesth Analg"},{"key":"R22-20250224","doi-asserted-by":"crossref","first-page":"634","DOI":"10.1136\/bmj.328.7440.634","article-title":"British Hypertension Society guidelines for hypertension management 2004 (BHS-IV): summary.","volume":"328","author":"Williams","year":"2004","journal-title":"BMJ"},{"key":"R25-20250224","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1109\/TVCG.2017.2744878","article-title":"Visualizing dataflow graphs of deep learning models in TensorFlow.","volume":"24","author":"Wongsuphasawat","year":"2018","journal-title":"IEEE Trans Vis Comput Graph"},{"key":"R26-20250224","doi-asserted-by":"crossref","DOI":"10.1038\/s41598-023-35492-y","article-title":"Continuous cuffless blood pressure monitoring using photoplethysmography-based PPG2BP-net for high intrasubject blood pressure variations.","volume":"13","author":"Joung","year":"2023","journal-title":"Sci Rep"},{"key":"R27-20250224","doi-asserted-by":"crossref","first-page":"7931","DOI":"10.3390\/s23187931","article-title":"Nocturnal blood pressure estimation from sleep plethysmography using machine learning.","volume":"23","author":"Yilmaz","year":"2023","journal-title":"Sensors (Basel, Switzerland)"},{"key":"R28-20250224","doi-asserted-by":"crossref","first-page":"170","DOI":"10.1093\/ehjdh\/ztae001","article-title":"Explainable machine learning using echocardiography to improve risk prediction in patients with chronic coronary syndrome.","volume":"5","author":"Molenaar","year":"2024","journal-title":"Eur. Heart J Digital Health"},{"key":"R29-20250224","doi-asserted-by":"crossref","first-page":"279","DOI":"10.1038\/s41597-022-01411-5","article-title":"VitalDB, a high-fidelity multi-parameter vital signs database in surgical patients.","volume":"9","author":"Lee","year":"2022","journal-title":"Sci Data"},{"key":"R30-20250224","doi-asserted-by":"crossref","first-page":"160035","DOI":"10.1038\/sdata.2016.35","article-title":"MIMIC-III, a freely accessible critical care database.","volume":"3","author":"Johnson","year":"2016","journal-title":"Sci Data"},{"key":"R31-20250224","doi-asserted-by":"crossref","first-page":"103404","DOI":"10.1016\/j.bspc.2021.103404","article-title":"MLP-BP: a novel framework for cuffless blood pressure measurement with PPG and ECG signals based on MLP-mixer neural networks.","volume":"73","author":"Huang","year":"2022","journal-title":"Biomed Signal Proc Control"},{"key":"R32-20250224","doi-asserted-by":"crossref","first-page":"94","DOI":"10.1213\/ANE.0b013e318270a6d9","article-title":"Beat-to-beat tracking of systolic blood pressure using noninvasive pulse transit time during anesthesia induction in hypertensive patients.","volume":"116","author":"Kim","year":"2013","journal-title":"Anesth Analg"},{"key":"R34-20250224","first-page":"3981","article-title":"Bayesian model averaging for improving the accuracy of cuffless blood pressure estimation.","volume":"2022","author":"Shen","year":"2022","journal-title":"Annu Inter Conf IEEE Eng Med Biol Soc"},{"key":"R35-20250224","doi-asserted-by":"crossref","first-page":"206","DOI":"10.1213\/ANE.0000000000000027","article-title":"Anesthesia information management system-based near real-time decision support to manage intraoperative hypotension and hypertension.","volume":"118","author":"Nair","year":"2014","journal-title":"Anesth Analg"}],"container-title":["Anesthesia &amp; Analgesia"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.ovid.com\/10.1213\/ANE.0000000000007295","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,7]],"date-time":"2026-08-07T14:36:43Z","timestamp":1786113403000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.ovid.com\/10.1213\/ANE.0000000000007295"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,12,16]]},"references-count":30,"journal-issue":{"issue":"1","published-print":{"date-parts":[[2025]]}},"URL":"https:\/\/doi.org\/10.1213\/ane.0000000000007295","relation":{},"ISSN":["0003-2999","1526-7598"],"issn-type":[{"value":"0003-2999","type":"print"},{"value":"1526-7598","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024,12,16]]}}}