1
|
Montomoli J, Bitondo MM, Cascella M, Rezoagli E, Romeo L, Bellini V, Semeraro F, Gamberini E, Frontoni E, Agnoletti V, Altini M, Benanti P, Bignami EG. Algor-ethics: charting the ethical path for AI in critical care. J Clin Monit Comput 2024:10.1007/s10877-024-01157-y. [PMID: 38573370 DOI: 10.1007/s10877-024-01157-y] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/18/2023] [Accepted: 03/22/2024] [Indexed: 04/05/2024]
Abstract
The integration of Clinical Decision Support Systems (CDSS) based on artificial intelligence (AI) in healthcare is groundbreaking evolution with enormous potential, but its development and ethical implementation, presents unique challenges, particularly in critical care, where physicians often deal with life-threating conditions requiring rapid actions and patients unable to participate in the decisional process. Moreover, development of AI-based CDSS is complex and should address different sources of bias, including data acquisition, health disparities, domain shifts during clinical use, and cognitive biases in decision-making. In this scenario algor-ethics is mandatory and emphasizes the integration of 'Human-in-the-Loop' and 'Algorithmic Stewardship' principles, and the benefits of advanced data engineering. The establishment of Clinical AI Departments (CAID) is necessary to lead AI innovation in healthcare, ensuring ethical integrity and human-centered development in this rapidly evolving field.
Collapse
Affiliation(s)
- Jonathan Montomoli
- Department of Anesthesia and Intensive Care, Infermi Hospital, Romagna Local Health Authority, Viale Settembrini 2, Rimini, 47923, Italy.
- Health Services Research, Evaluation and Policy Unit, Romagna Local Health Authority, Viale Settembrini 2, Rimini, 47923, Italy.
| | - Maria Maddalena Bitondo
- Department of Anesthesia and Intensive Care, Infermi Hospital, Romagna Local Health Authority, Viale Settembrini 2, Rimini, 47923, Italy
| | - Marco Cascella
- Unit of Anesthesia and Pain Medicine, Department of Medicine, Surgery and Dentistry "Scuola Medica Salernitana, " University of Salerno, Baronissi, Salerno, Italy
| | - Emanuele Rezoagli
- School of Medicine and Surgery, University of Milano-Bicocca, Via Cadore, 48, Monza, 20900, Italy
- Dipartimento di Emergenza e Urgenza, Terapia intensiva e Semintensiva adulti e pediatrica, Fondazione IRCCS San Gerardo dei Tintori, Via Pergolesi, 33, Monza, 20900, Italy
| | - Luca Romeo
- Department of Economics and Law, University of Macerata, Macerata, 62100, Italy
| | - Valentina Bellini
- Anesthesiology, Critical Care and Pain Medicine Division, Department of Medicine and Surgery, University of Parma, Via Gramsci 14, Parma, 43125, Italy
| | - Federico Semeraro
- Department of Anesthesia, Intensive Care and Prehospital Emergency, Ospedale Maggiore Carlo Alberto Pizzardi, Largo Bartolo Nigrisoli, 2, Bologna, 40133, Italy
| | - Emiliano Gamberini
- Department of Anesthesia and Intensive Care, Infermi Hospital, Romagna Local Health Authority, Viale Settembrini 2, Rimini, 47923, Italy
| | - Emanuele Frontoni
- Department of Political Sciences, Communication and International Relations, University of Macerata, Macerata, 62100, Italy
| | - Vanni Agnoletti
- Department of Surgery and Trauma, Anesthesia and Intensive Care Unit, Maurizio Bufalini Hospital, Romagna Local Health Authority, Viale Giovanni Ghirotti, 286, Cesena, 47521, Italy
| | - Mattia Altini
- Hospital Care Sector, Emilia-Romagna Region, Via Aldo Moro, 21, Bologna, 40127, Italy
| | - Paolo Benanti
- Pontifical Gregorian University, Piazza della Pilotta 4, Roma, 00187, Italy
| | - Elena Giovanna Bignami
- Anesthesiology, Critical Care and Pain Medicine Division, Department of Medicine and Surgery, University of Parma, Via Gramsci 14, Parma, 43125, Italy
| |
Collapse
|
2
|
Khamooshi M, Wickramarachchi A, Byrne T, Seman M, Fletcher DF, Burrell A, Gregory SD. Blood flow and emboli transport patterns during venoarterial extracorporeal membrane oxygenation: A computational fluid dynamics study. Comput Biol Med 2024; 172:108263. [PMID: 38489988 DOI: 10.1016/j.compbiomed.2024.108263] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/11/2023] [Revised: 02/15/2024] [Accepted: 03/06/2024] [Indexed: 03/17/2024]
Abstract
PROBLEM Despite advances in Venoarterial Extracorporeal Membrane Oxygenation (VA-ECMO), a significant mortality rate persists due to complications. The non-physiological blood flow dynamics of VA-ECMO may lead to neurological complications and organ ischemia. Continuous retrograde high-flow oxygenated blood enters through a return cannula placed in the femoral artery which opposes the pulsatile deoxygenated blood ejected by the left ventricle (LV), which impacts upper body oxygenation and subsequent hyperoxemia. The complications underscore the critical need to comprehend the impact of VA-ECMO support level and return cannula size, as mortality remains a significant concern. AIM The aim of this study is to predict and provide insights into the complications associated with VA-ECMO using computational fluid dynamics (CFD) simulations. These complications will be assessed by characterising blood flow and emboli transport patterns through a comprehensive analysis of the influence of VA-ECMO support levels and arterial return cannula sizes. METHODS Patient-specific 3D aortic and major branch models, derived from a male patient's CT scan during VA-ECMO undergoing respiratory dysfunction, were analyzed using CFD. The investigation employed species transport and discrete particle tracking models to study ECMO blood (oxygenated) mixing with LV blood (deoxygenated) and to trace emboli transport patterns from potential sources (circuit, LV, and aorta wall). Two cannula sizes (15 Fr and 19 Fr) were tested alongside varying ECMO pump flow rates (50%, 70%, and 90% of the total cardiac output). RESULTS Cannula size did not significantly affect oxygen transport. At 90% VA-ECMO support, all arteries distal to the aortic arch achieved 100% oxygen saturation. As support level decreased, oxygen transport to the upper body also decreased to a minimum saturation of 73%. Emboli transport varied substantially between emboli origin and VAECMO support level, with the highest risk of cerebral emboli coming from the LV with a 15 Fr cannula at 90% support. CONCLUSION Arterial return cannula sizing minimally impacted blood oxygen distribution; however, it did influence the distribution of emboli released from the circuit and aortic wall. Notably, it was the support level alone that significantly affected the mixing zone of VA-ECMO and cardiac blood, subsequently influencing the risk of embolization of the cardiogenic source and oxygenation levels across various arterial branches.
Collapse
Affiliation(s)
- Mehrdad Khamooshi
- Cardio-Respiratory Engineering and Technology Laboratory (CREATElab), Department of Mechanical and Aerospace Engineering, Monash University, Wellington Road, Clayton, 3800, Victoria, Australia.
| | - Avishka Wickramarachchi
- Cardio-Respiratory Engineering and Technology Laboratory (CREATElab), Department of Mechanical and Aerospace Engineering, Monash University, Wellington Road, Clayton, 3800, Victoria, Australia.
| | - Tim Byrne
- Intensive Care Unit, Alfred Hospital, 89 Commercial Road, Melbourne, 3004, Victoria, Australia.
| | - Michael Seman
- Cardio-Respiratory Engineering and Technology Laboratory (CREATElab), Department of Mechanical and Aerospace Engineering, Monash University, Wellington Road, Clayton, 3800, Victoria, Australia.
| | - David F Fletcher
- School of Chemical and Biomolecular Engineering, The University of Sydney, Darlington, 2006, New South Wales, Australia.
| | - Aidan Burrell
- Intensive Care Unit, Alfred Hospital, 89 Commercial Road, Melbourne, 3004, Victoria, Australia.
| | - Shaun D Gregory
- Cardio-Respiratory Engineering and Technology Laboratory (CREATElab), Department of Mechanical and Aerospace Engineering, Monash University, Wellington Road, Clayton, 3800, Victoria, Australia.
| |
Collapse
|
3
|
De Somer F, Swol J. Extracorporeal membrane oxygenation from last resort to indispensable tool in the treatment of respiratory and/or circulatory failure. Perfusion 2024; 39:3S-4S. [PMID: 38651579 DOI: 10.1177/02676591231226292] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 04/25/2024]
Affiliation(s)
| | - Justyna Swol
- Department of Respiratory Medicine, Paracelsus Medical University, Nuremberg, Germany
| |
Collapse
|
4
|
Kalra A, Bachina P, Shou BL, Hwang J, Barshay M, Kulkarni S, Sears I, Eickhoff C, Bermudez CA, Brodie D, Ventetuolo CE, Kim BS, Whitman GJR, Abbasi A, Cho SM. Predicting Acute Brain Injury in Venoarterial Extracorporeal Membrane Oxygenation Patients with Tree-Based Machine Learning: Analysis of the Extracorporeal Life Support Organization Registry. RESEARCH SQUARE 2024:rs.3.rs-3848514. [PMID: 38260374 PMCID: PMC10802703 DOI: 10.21203/rs.3.rs-3848514/v1] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 01/24/2024]
Abstract
Objective To determine if machine learning (ML) can predict acute brain injury (ABI) and identify modifiable risk factors for ABI in venoarterial extracorporeal membrane oxygenation (VA-ECMO) patients. Design Retrospective cohort study of the Extracorporeal Life Support Organization (ELSO) Registry (2009-2021). Setting International, multicenter registry study of 676 ECMO centers. Patients Adults (≥18 years) supported with VA-ECMO or extracorporeal cardiopulmonary resuscitation (ECPR). Interventions None. Measurements and Main Results Our primary outcome was ABI: central nervous system (CNS) ischemia, intracranial hemorrhage (ICH), brain death, and seizures. We utilized Random Forest, CatBoost, LightGBM and XGBoost ML algorithms (10-fold leave-one-out cross-validation) to predict and identify features most important for ABI. We extracted 65 total features: demographics, pre-ECMO/on-ECMO laboratory values, and pre-ECMO/on-ECMO settings.Of 35,855 VA-ECMO (non-ECPR) patients (median age=57.8 years, 66% male), 7.7% (n=2,769) experienced ABI. In VA-ECMO (non-ECPR), the area under the receiver-operator characteristics curves (AUC-ROC) to predict ABI, CNS ischemia, and ICH was 0.67, 0.67, and 0.62, respectively. The true positive, true negative, false positive, false negative, positive, and negative predictive values were 33%, 88%, 12%, 67%, 18%, and 94%, respectively for ABI. Longer ECMO duration, higher 24h ECMO pump flow, and higher on-ECMO PaO2 were associated with ABI.Of 10,775 ECPR patients (median age=57.1 years, 68% male), 16.5% (n=1,787) experienced ABI. The AUC-ROC for ABI, CNS ischemia, and ICH was 0.72, 0.73, and 0.69, respectively. The true positive, true negative, false positive, false negative, positive, and negative predictive values were 61%, 70%, 30%, 39%, 29% and 90%, respectively, for ABI. Longer ECMO duration, younger age, and higher 24h ECMO pump flow were associated with ABI. Conclusions This is the largest study predicting neurological complications on sufficiently powered international ECMO cohorts. Longer ECMO duration and higher 24h pump flow were associated with ABI in both non-ECPR and ECPR VA-ECMO.
Collapse
Affiliation(s)
| | | | | | | | | | | | - Isaac Sears
- Warren Alpert Medical School of Brown University
| | | | | | | | | | - Bo Soo Kim
- Johns Hopkins University School of Medicine
| | | | - Adeel Abbasi
- Warren Alpert Medical School of Brown University
| | | |
Collapse
|
5
|
Kalra A, Bachina P, Shou BL, Hwang J, Barshay M, Kulkarni S, Sears I, Eickhoff C, Bermudez CA, Brodie D, Ventetuolo CE, Whitman GJR, Abbasi A, Cho SM. Utilizing Machine Learning to Predict Neurological Injury in Venovenous Extracorporeal Membrane Oxygenation Patients: An Extracorporeal Life Support Organization Registry Analysis. RESEARCH SQUARE 2023:rs.3.rs-3779429. [PMID: 38196631 PMCID: PMC10775358 DOI: 10.21203/rs.3.rs-3779429/v1] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 01/11/2024]
Abstract
Background Venovenous extracorporeal membrane oxygenation (VV-ECMO) is associated with acute brain injury (ABI), including central nervous system (CNS) ischemia (defined as ischemic stroke or hypoxic-ischemic brain injury) and intracranial hemorrhage (ICH). There is limited data on prediction models for ABI and neurological outcomes in VV-ECMO. Research Question Can machine learning (ML) accurately predict ABI and identify modifiable factors of ABI in VV-ECMO? Study Design and Methods We analyzed adult (≥18 years) VV-ECMO patients in the Extracorporeal Life Support Organization Registry (2009-2021) from 676 centers. ABI was defined as CNS ischemia, ICH, brain death, and seizures. Overall, 65 total variables were extracted including clinical characteristics and pre-ECMO and on-ECMO variables. Random Forest, CatBoost, LightGBM, and XGBoost ML algorithms (10-fold leave-one-out cross-validation) were used to predict ABI. Feature Importance Scores were used to pinpoint variables most important for predicting ABI. Results Of 37,473 VV-ECMO patients (median age=48.1 years, 63% male), 2,644 (7.1%) experienced ABI: 610 (2%) and 1,591 (4%) experienced CNS ischemia and ICH, respectively. The median ECMO duration was 10 days (interquartile range=5-20 days). The area under the receiver-operating characteristics curves to predict ABI, CNS ischemia, and ICH were 0.67, 0.63, and 0.70, respectively. The accuracy, positive predictive, and negative predictive values for ABI were 79%, 15%, and 95%, respectively. ML identified pre-ECMO cardiac arrest as the most important risk factor for ABI while ECMO duration and bridge to transplantation as an indication for ECMO were associated with lower risk of ABI. Interpretation This is the first study to use machine learning to predict ABI in a large cohort of VV-ECMO patients. Performance was sub-optimal due to the low reported prevalence of ABI with lack of standardization of neuromonitoring/imaging protocols and data granularity in the ELSO Registry. Standardized neurological monitoring and imaging protocols may improve machine learning performance to predict ABI.
Collapse
Affiliation(s)
| | | | | | | | | | | | - Isaac Sears
- Warren Alpert Medical School of Brown University
| | | | | | | | | | | | - Adeel Abbasi
- Warren Alpert Medical School of Brown University
| | | |
Collapse
|
6
|
Riera J, Bělohlávek J, Jung C. De senectute and the art of medicine: how old is too old for ECMO in cardiogenic shock? Intensive Care Med 2023; 49:1511-1513. [PMID: 37922009 DOI: 10.1007/s00134-023-07251-0] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/29/2023] [Accepted: 10/09/2023] [Indexed: 11/05/2023]
Affiliation(s)
- Jordi Riera
- Intensive Care Department, Vall d'Hebron University Hospital, Barcelona, Spain.
- SODIR, Vall d'Hebron Research Institute, Barcelona, Spain.
| | - Jan Bělohlávek
- Faculty of Medicine, Charles University in Prague, Prague, Czech Republic
- Department of Internal Medicine, Cardiovascular Medicine, General University Hospital, Prague, Czech Republic
| | - Christian Jung
- Division of Cardiology, Pulmonology and Vascular Medicine, University Hospital Dusseldorf, Heinrich-Heine-University, Dusseldorf, Germany
- Cardiovascular Research Institute Düsseldorf (CARID), Medical Faculty, Heinrich-Heine University, Dusseldorf, Germany
| |
Collapse
|
7
|
Stephens AF, Šeman M, Hodgson CL, Gregory SD. SHAP Model Explainability in ECMO - PAL mortality prediction: A Critical Analysis. Author's reply. Intensive Care Med 2023; 49:1560-1562. [PMID: 37906257 DOI: 10.1007/s00134-023-07237-y] [Citation(s) in RCA: 1] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Accepted: 09/19/2023] [Indexed: 11/02/2023]
Affiliation(s)
- Andrew F Stephens
- Cardio-Respiratory Engineering and Technology Laboratory, Department of Mechanical and Aerospace Engineering, Monash University, Melbourne, AU, USA.
| | - Michael Šeman
- Cardio-Respiratory Engineering and Technology Laboratory, Department of Mechanical and Aerospace Engineering, Monash University, Melbourne, AU, USA
| | - Carol L Hodgson
- School of Public Health and Preventive Medicine, Monash University, Melbourne, AU, USA
| | - Shaun D Gregory
- Cardio-Respiratory Engineering and Technology Laboratory, Department of Mechanical and Aerospace Engineering, Monash University, Melbourne, AU, USA
| |
Collapse
|
8
|
Valiente Fernández M, Lesmes González de Aledo A, Delgado Moya FDP, Martín Badía I. SHAP model explainability in ECMO-PAL mortality prediction: a critical analysis. Intensive Care Med 2023; 49:1559. [PMID: 37906260 DOI: 10.1007/s00134-023-07252-z] [Citation(s) in RCA: 1] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Accepted: 10/10/2023] [Indexed: 11/02/2023]
Affiliation(s)
- Marcos Valiente Fernández
- Intensive Care Medicine Department, Hospital Universitario 12 de Octubre, Avda. de Córdoba, s/n, 28041, Madrid, Spain.
| | | | | | - Isaías Martín Badía
- Intensive Care Medicine Department, Hospital Universitario 12 de Octubre, Avda. de Córdoba, s/n, 28041, Madrid, Spain
| |
Collapse
|