1
|
Barison A, Ricci F, Pavon AG, Muscogiuri G, Bisaccia G, Camastra G, De Lazzari M, Lanzillo C, Raguso M, Monti L, Vargiu S, Pedrotti P, Piacenti M, Todiere G, Pontone G, Indolfi C, Dellegrottaglie S, Lombardi M, Schwitter J, Aquaro GD. Cardiovascular Magnetic Resonance in Patients with Cardiac Electronic Devices: Evidence from a Multicenter Study. J Clin Med 2023; 12:6673. [PMID: 37892813 PMCID: PMC10607654 DOI: 10.3390/jcm12206673] [Citation(s) in RCA: 2] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/13/2023] [Revised: 10/11/2023] [Accepted: 10/17/2023] [Indexed: 10/29/2023] Open
Abstract
BACKGROUND Most recent cardiac implantable electronic devices (CIEDs) can safely undergo a cardiovascular magnetic resonance (CMR) scan under certain conditions, but metal artifacts may degrade image quality. The aim of this study was to assess the overall diagnostic yield of CMR and the extent of metal artifacts in a multicenter, multivendor study on CIED patients referred for CMR. METHODS We analyzed 309 CMR scans from 292 patients (age 57 ± 16 years, 219 male) with an MR-conditional pacemaker (n = 122), defibrillator (n = 149), or loop recorder (n = 38); CMR scans were performed in 10 centers from 2012 to 2020; MR-unsafe implants were excluded. Clinical and device parameters were recorded before and after the CMR scan. A visual analysis of metal artifacts was performed for each sequence on a segmental basis, based on a 5-point artifact score. RESULTS The vast majority of CMR scans (n = 255, 83%) were completely performed, while only 32 (10%) were interrupted soon after the first sequences and 22 (7%) were only partly acquired; CMR quality was non-diagnostic in 34 (11%) scans, poor (<1/3 sequences were diagnostic) in 25 (8%), or acceptable (1/3 to 2/3 sequences were diagnostic) in 40 (13%), while most scans (n = 201, 68%) were of overall good quality. No adverse event or device malfunctioning occurred, and only nonsignificant changes in device parameters were recorded. The most affected sequences were SSFP (median score 0.32 [interquartile range 0.07-0.91]), followed by GRE (0.18 [0.02-0.59]) and LGE (0.14 [0.02-0.55]). ICDs induced more artifacts (median score in SSFP images 0.87 [0.50-1.46]) than PMs (0.11 [0.03-0.28]) or ILRs (0.11 [0.00-0.56]). Moreover, most artifacts were located in the anterior, anteroseptal, anterolateral, and apical segments of the LV and in the outflow tract of the RV. CONCLUSIONS CMR is a versatile imaging technique, with a high safety profile and overall good image quality even in patients with MR-conditional CIEDs. Several strategies are now available to optimize image quality, substantially enhancing overall diagnostic yield.
Collapse
Affiliation(s)
- Andrea Barison
- Fondazione Toscana Gabriele Monasterio, 56127 Pisa, Italy
- Life Science Institute, Scuola Superiore Sant’Anna, 56127 Pisa, Italy
| | - Fabrizio Ricci
- Department of Neuroscience, Imaging and Clinical Sciences, G. d’Annunzio University of Chieti-Pescara, 66100 Chieti, Italy
| | - Anna Giulia Pavon
- Cardiocentro Ticino Institute, Ente Ospedaliero Cantonale, 6900 Lugano, Switzerland
| | - Giuseppe Muscogiuri
- Department of Perioperative Cardiology and Cardiovascular Imaging, Centro Cardiologico Monzino IRCCS, 20138 Milan, Italy
| | - Giandomenico Bisaccia
- Department of Neuroscience, Imaging and Clinical Sciences, G. d’Annunzio University of Chieti-Pescara, 66100 Chieti, Italy
| | | | - Manuel De Lazzari
- Department of Cardiac Thoracic and Vascular Sciences and Public Health, University of Padua, 35122 Padova, Italy
| | | | - Mario Raguso
- Ospedale Policlinico Casilino, 00169 Roma, Italy
| | - Lorenzo Monti
- IRCCS Humanitas Research Hospital, 20089 Rozzano, Italy
| | - Sara Vargiu
- Cardiologia 3, ASST Grande Ospedale Metropolitano Niguarda, 20162 Milan, Italy
| | - Patrizia Pedrotti
- Cardiac Magnetic Resonance Laboratory, Cardiologia 4, ASST Grande Ospedale Metropolitano Niguarda, 20162 Milan, Italy
| | | | | | - Gianluca Pontone
- Department of Perioperative Cardiology and Cardiovascular Imaging, Centro Cardiologico Monzino IRCCS, 20138 Milan, Italy
- Department of Biomedical, Surgical and Dental Sciences, University of Milan, 20122 Milan, Italy
| | - Ciro Indolfi
- Division of Cardiology, Magna Graecia University, 88100 Catanzaro, Italy
- Center for Cardiovascular Research, Magna Graecia University, 88100 Catanzaro, Italy
- Mediterranea Cardiocentro, 80122 Naples, Italy
| | - Santo Dellegrottaglie
- Advanced Cardiovascular Imaging Unit, Ospedale Medico-Chirurgico Accreditato Villa dei Fiori, 80011 Acerra, Italy
| | - Massimo Lombardi
- Multimodality Cardiac Imaging Section, IRCCS Policlinico San Donato, 20097 Milan, Italy
| | - Juerg Schwitter
- Division of Cardiology, Cardiovascular Department, University Hospital Lausanne—CHUV, 1011 Lausanne, Switzerland
- CMR Center, University Hospital Lausanne—CHUV, 1011 Lausanne, Switzerland
- Faculty of Biology & Medicine, University of Lausanne—UniL, 1015 Lausanne, Switzerland
| | - Giovanni Donato Aquaro
- Academic Radiology Unit, Department of Surgical Medical and Molecular Pathology and Critical Area, University of Pisa, 56126 Pisa, Italy
| | | | | |
Collapse
|
2
|
Neural Network Detection of Pacemakers for MRI Safety. J Digit Imaging 2022; 35:1673-1680. [PMID: 35768751 DOI: 10.1007/s10278-022-00663-2] [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: 10/31/2021] [Revised: 04/23/2022] [Accepted: 05/30/2022] [Indexed: 10/17/2022] Open
Abstract
Flagging the presence of cardiac devices such as pacemakers before an MRI scan is essential to allow appropriate safety checks. We assess the accuracy with which a machine learning model can classify the presence or absence of a pacemaker on pre-existing chest radiographs. A total of 7973 chest radiographs were collected, 3996 with pacemakers visible and 3977 without. Images were identified from information available on the radiology information system (RIS) and correlated with report text. Manual review of images by two board certified radiologists was performed to ensure correct labeling. The data set was divided into training, validation, and a hold-back test set. The data were used to retrain a pre-trained image classification neural network. Final model performance was assessed on the test set. Accuracy of 99.67% on the test set was achieved. Re-testing the final model on the full training and validation data revealed a few additional misclassified examples which are further analyzed. Neural network image classification could be used to screen for the presence of cardiac devices, in addition to current safety processes, providing notification of device presence in advance of safety questionnaires. Computational power to run the model is low. Further work on misclassified examples could improve accuracy on edge cases. The focus of many healthcare applications of computer vision techniques has been for diagnosis and guiding management. This work illustrates an application of computer vision image classification to enhance current processes and improve patient safety.
Collapse
|