Abstract
Focused Assessment with Sonography in Trauma (FAST) exam is the standard of care for pericardial and abdominal free fluid detection in emergency medicine. Despite its life saving potential, FAST is underutilized due to requiring clinicians with appropriate training and practice. To aid ultrasound interpretation, the role of artificial intelligence has been studied, while leaving room for improvement in localization information and computation time. The purpose of this study was to develop and test a deep learning approach to rapidly and accurately identify both the presence and location of pericardial effusion on point-of-care ultrasound (POCUS) exams. Each cardiac POCUS exam is analyzed image-by-image via the state-of-the-art YoloV3 algorithm and pericardial effusion presence is determined from the most confident detection. We evaluate our approach over a dataset of POCUS exams (cardiac component of FAST and ultrasound), comprising 37 cases with pericardial effusion and 39 negative controls. Our algorithm attains 92% specificity and 89% sensitivity in pericardial effusion identification, outperforming existing deep learning approaches, and localizes pericardial effusion by 51% Intersection Over Union with ground-truth annotations. Moreover, image processing demonstrates only 57 ms latency. Experimental results demonstrate the feasibility of rapid and accurate pericardial effusion detection from POCUS exams for physician overread.
Graphical Abstract
Similar content being viewed by others
References
Rose JS et al (2005) The fast is positive, now what? Derivation of a clinical decision rule to determine the need for therapeutic laparotomy in adults with blunt torso trauma and a positive trauma ultrasound. J Emerg Med 29(1):15–21
Moylan M et al (2007) Association between a positive ED FAST examination and therapeutic laparotomy in normotensive blunt trauma patients. J Emerg Med 33(3):265–71
Helling TS, Wilson J, Augustosky K (2007) The utility of focused abdominal ultrasound in blunt abdominal trauma: a reappraisal. Am J Surg 194(6):728–32 (discussion 732-3)
Quinn AC, Sinert R (2011) What is the utility of the Focused Assessment with Sonography in Trauma (FAST) exam in penetrating torso trauma? Injury 42(5):482–487
Melniker LA et al (2006) Randomized controlled clinical trial of point-of-care, limited ultrasonography for trauma in the emergency department: the first sonography outcomes assessment program trial. Ann Emerg Med 48(3):227–235
American college of surgeons committee on trauma (1997) advanced trauma life support course for physicians. American College of Surgeons, Chicago
Tayal VS et al (2004) FAST (focused assessment with sonography in trauma) accurate for cardiac and intraperitoneal injury in penetrating anterior chest trauma. J Ultrasound Med 23(4):467–472
Ma OJ et al (1995) Prospective analysis of a rapid trauma ultrasound examination performed by emergency physicians. J Trauma 38(6):879–885
Brooks A et al (2004) Prospective evaluation of non-radiologist performed emergency abdominal ultrasound for haemoperitoneum. Emerg Med J 21(5):580–581
Rozycki GS et al (1995) A prospective study of surgeon-performed ultrasound as the primary adjuvant modality for injured patient assessment. J Trauma 39(3):492–8 (discussion 498-500)
Soundappan SV et al (2005) Diagnostic accuracy of surgeon-performed focused abdominal sonography (FAST) in blunt paediatric trauma. Injury 36(8):970–975
Kimura A, Otsuka T (1991) Emergency center ultrasonography in the evaluation of hemoperitoneum: a prospective study. J Trauma 31(1):20–23
Biffl WL et al (2009) Management of patients with anterior abdominal stab wounds: a Western Trauma Association multicenter trial. J Trauma 66(5):1294–1301
O’Connor G et al (2013) Looking beyond Morison’s pouch in focused assessment with sonography for trauma: penetrating hepatobiliary trauma and a new sign for emergency physicians. Emerg Med J 30(9):778–779
Hoffmann B, Nguyen H, Hill HF (2009) Diaphragmatic laceration after penetrating trauma: direct visualization and indirect findings on focused assessment with sonography for trauma in the emergency department. J Ultrasound Med 28(9):1259–1263
Boulanger BR et al (2001) The routine use of sonography in penetrating torso injury is beneficial. J Trauma 51(2):320–325
Kirkpatrick AW et al (2004) The hand-held ultrasound examination for penetrating abdominal trauma. Am J Surg 187(5):660–665
Soffer D et al (2004) A prospective evaluation of ultrasonography for the diagnosis of penetrating torso injury. J Trauma 56(5):953–7 (discussion 957-9)
Branney SW et al (1995) Quantitative sensitivity of ultrasound in detecting free intraperitoneal fluid. J Trauma 39(2):375–380
Bode PJ et al (1993) Abdominal ultrasound as a reliable indicator for conclusive laparotomy in blunt abdominal trauma. J Trauma 34(1):27–31
Jehle D, Guarino J, Karamanoukian H (1993) Emergency department ultrasound in the evaluation of blunt abdominal trauma. Am J Emerg Med 11(4):342–346
McGahan JP et al (1997) Use of ultrasonography in the patient with acute abdominal trauma. J Ultrasound Med 16(10):653–62 (quiz 663-4)
Maitra S et al (2008) When FAST is a FAFF: is FAST scanning useful in non-trauma patients? Ultrasound 16(3):165–168
Moore C et al (2007) Free fluid in Morison’s pouch on bedside ultrasound predicts need for operative intervention in suspected ectopic pregnancy. Acad Emerg Med 14(8):755–758
Volpicelli G et al (2013) Point-of-care multiorgan ultrasonography for the evaluation of undifferentiated hypotension in the emergency department. Intensive Care Med 39(7):1290–1298
Gaspari R, Weekes A, Adhikari S et al (2016) Emergency department point-of-care ultrasound in out-of-hospital and in-ED cardiac arrest. Resuscitation 109:33–39
Tayal VS, Beatty MA, Marx JA, Tomaszewski CA, Thomason MH (2004) FAST (focused assessment with sonography in trauma) accurate for cardiac and intraperitoneal injury in penetrating anterior chest trauma. J Ultrasound Med 23(4):467–472. https://doi.org/10.7863/jum.2004.23.4.467
Netherton S, Milenkovic V, Taylor M, Davis P (2019) Diagnostic accuracy of eFAST in the trauma patient: A systematic review and meta-analysis. Can J Emerg Med 21(6):727–738. https://doi.org/10.1017/cem.2019.381
Hall MK, Omer T, Moore CL, Taylor RA (2016) Cost-effectiveness of the cardiac component of the focused assessment of sonography in trauma examination in blunt trauma. Acad Emerg Med 23(4):415–423. https://doi.org/10.1111/acem.12936
Moore CL, Molina AA, Lin H (2006) Ultrasonography in community emergency departments in the United States: access to ultrasonography performed by consultants and status of emergency physician-performed ultrasonography. Ann Emerg Med 47(2):147–153
Akhtar S et al (2009) Resident training in emergency ultrasound: consensus recommendations from the 2008 Council of Emergency Medicine Residency Directors Conference. Acad Emerg Med 16(Suppl 2):S32–S36
Counselman FL et al (2003) The status of bedside ultrasonography training in emergency medicine residency programs. Acad Emerg Med 10(1):37–42
Freitas ML, Frangos SG, Frankel HL (2006) The status of ultrasonography training and use in general surgery residency programs. J Am Coll Surg 202(3):453–458
Wehbe RM, Thomas JD (2022) Validating deep learning to distinguish Takotsubo syndrome from acute myocardial infarction-beware of shortcuts, human supervision required. JAMA Cardiol 7(5):477–479. https://doi.org/10.1001/jamacardio.2022.0193
Pokaprakarn T, Prieto JC, Price JT, Kasaro MP, Sindano N, Shah HR, Peterson M, Akapelwa MM, Kapilya FM, Sebastião YV, Goodnight W (2022) AI estimation of gestational age from blind ultrasound sweeps in low-resource settings. N Engl J Med Evid 1(5):EVIDoa2100058
Wilson M, Chopra R, Wilson MZ et al (2021) Validation and clinical applicability of whole-volume automated segmentation of optical coherence tomography in retinal disease using deep learning. JAMA Ophthalmol 139(9):964–973. https://doi.org/10.1001/jamaophthalmol.2021.2273
Van Sloun RJ, Cohen R, Eldar YC (2019) Deep learning in ultrasound imaging. Proc IEEE 108(1):11–29
Diniz PHB, Yin Y, Collins S (2020) Deep learning strategies for ultrasound in pregnancy. Eur Med J Reprod Health 6(1):73–80
Liu S, Wang Y, Yang X et al (2019) Deep learning in medical ultrasound analysis: a review. Engineering 5(Generic):261–275. https://doi.org/10.1016/j.eng.2018.11.020
Akkus Z, Cai J, Boonrod A et al (2019) A survey of deep-learning applications in ultrasound: artificial intelligence-powered ultrasound for improving clinical workflow. J Am Coll Radiol 16(9 Pt B):1318–1328. https://doi.org/10.1016/j.jacr.2019.06.004
Madani A, Arnaout R, Mofrad M, Arnaout R (2018) Fast and accurate view classification of echocardiograms using deep learning. NPJ Digit Med 1(1):6
Dong S, Luo G, Wang K, Cao S, Li Q, Zhang H (2018) A combined fully convolutional networks and deformable model for automatic left ventricle segmentation based on 3d echocardiography. Biomed Res Int 2018:5682365
Moradi M, Guo Y, Gur Y, Negahdar M, Syeda-Mahmood T (2016) A cross-modality neural network transform for semi-automatic medical image annotation. In Medical Image Computing and Computer-Assisted Intervention–MICCAI 2016: 19th International Conference, Athens, Greece. Proceedings, Part II 19 2016 (pp. 300–307). Springer International Publishing
Laumer F, Di Vece D, Cammann VL et al (2022) Assessment of artificial intelligence in echocardiography diagnostics in differentiating Takotsubo syndrome from myocardial infarction. JAMA Cardiol 7(5):494–503. https://doi.org/10.1001/jamacardio.2022.0183
Litjens G, Ciompi F, Wolterink JM, de Vos BD, Leiner T, Teuwen J, Išgum I (2019) State-of-the-art deep learning in cardiovascular image analysis. JACC: Cardiovasc Imaging 12(8 Part 1):1549–1565
Cheng CY, Chiu IM, Hsu MY, Pan HY, Tsai CM, Lin CR (2021) Deep learning assisted detection of abdominal free fluid in Morison’s pouch during focused assessment with sonography in trauma. Front Med (Lausanne) 23(8):707437. https://doi.org/10.3389/fmed.2021.707437
Lin Z, Li Z, Cao P et al (2022) Deep learning for emergency ascites diagnosis using ultrasonography images. J Appl Clin Med Phys 23(7):e13695. https://doi.org/10.1002/acm2.13695
Nayak A, Ouyang D, Ashley EA (2020) A deep learning algorithm accurately detects pericardial effusion on echocardiography. J Am Coll Cardiol 75(11_Supplement_1):1563–1563
Wu CC, Cheng CY, Chen HC, Hung CH, Chen TY, Lin CHR, Chiu IM (2022) Development and validation of an end-to-end deep learning pipeline to measure pericardial effusion in echocardiography. medRxiv pp.2022–08
Guidelines U (2017) Emergency, Point-of-care and clinical ultrasound guidelines in medicine. Ann Emerg Med 69(5):e27–e54. https://doi.org/10.1016/j.annemergmed.2016.08.457
NCH Software Inc (2019) PhotoPad Image Editor [Computer software]. Retrieved from https://www.nchsoftware.com/
Redmon J, Farhadi A (2018) “YOLOv3: An incremental improvement.” arXiv preprint arxiv:1804.02767
Diwan T, Anirudh G, Tembhurne JV (2023) Object detection using YOLO: Challenges, architectural successors, datasets and applications. Multimed Tools Appl 82(6):9243–75
Han R, Liu X, Chen T (2022) Yolo-SG: Salience-Guided Detection Of Small Objects In Medical Images. In 2022 IEEE International Conference on Image Processing (ICIP) (pp. 4218–4222)
Bengio Y, Courville A, Vincent P (2013) Representation learning: a review and new perspectives. IEEE Trans Pattern Anal Mach Intell 35(8):1798–1828
Lin TY, Maire M, Belongie S, Hays J, Perona P, Ramanan D, Dollár P, Zitnick CL (2014) Microsoft coco: common objects in context. In Computer Vision–ECCV 2014: 13th European Conference, Zurich, Switzerland, Proceedings, Part V 13 2014 (pp. 740–755). Springer International Publishing
Kingma DP, Ba J (2014) “Adam: a method for stochastic optimization.” arXiv preprint arXiv:1412.6980
Dey P, Gopal M, Pradhan P, Pal T (2019) On robustness of radial basis function network with input perturbation. Neural Comput Appl 31(2):523–537
Krogh A, Hertz J (1991) A simple weight decay can improve generalization. Adv Neural Inf Process Syst 4
Geirhos R, Rubisch P, Michaelis C, Bethge M, Wichmann FA, Brendel W (2019) ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness. Proceedings of the 7th International Conference on Learning Representations, New Orleans, LA
Ford N, Gilmer J, Carlini N, Cubuk D (2019) Adversarial examples are a natural consequence of test error in noise. arXiv preprint arXiv:1901.10513
Rusak E, Schott L, Zimmermann RS, Bitterwolf J, Bringmann O, Bethge M, Brendel W (2020) A simple way to make neural networks robust against diverse image corruptions. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, Proceedings, Part III 16 2020 (pp. 53–69). Springer International Publishing
Pinton GF, Trahey GE, Dahl JJ (2011) Sources of image degradation in fundamental and harmonic ultrasound imaging using nonlinear, full-wave simulations. IEEE Trans Ultrason Ferroelectr Freq Control 58(4):754–765
Ioffe S, Szegedy C (2015) Batch normalization: Accelerating deep network training by reducing internal covariate shift. In International conference on machine learning (pp. 448–456). Proceedings of Machine Learning Research
Eykholt K, Evtimov I, Fernandes E, Li B, Rahmati A, Tramèr F, Prakash A, Kohno T, Song D (2018) Physical adversarial examples for object detectors. In Proceedings of the 12th USENIX Conference on Offensive Technologies (pp. 1–1)
Redmon J, Farhadi A (2017) YOLO9000: better, faster, stronger. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 7263–7271)
Power M, Fell G, Wright M (2013) Principles for high-quality, high-value testing. BMJ Evid-Based Med 18(1):5–10
Liu W, Anguelov D, Erhan D, Szegedy C, Reed S, Fu CY, Berg AC. SSD: Single shot multibox detector. In Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11–14, 2016, Proceedings, Part I 14 2016 (pp. 21–37). Springer International Publishing
Elgendi M, Nasir MU, Tang Q, Fletcher RR, Howard N, Menon C, Ward R, Parker W, Nicolaou S (2020) The performance of deep neural networks in differentiating chest X-rays of COVID-19 patients from other bacterial and viral pneumonias. Front Med 7:550
Sumit SS, Watada J, Roy A, Rambli DRA (2020) In object detection deep learning methods, YOLO shows supremum to Mask R-CNN. J Phys: Conf Ser 1529(4):042086 (IOP Publishing)
Liu J, Cai J, Chellamuthu K, Bagheri M, Lu L, Summers RM (2018) Cascaded coarse-to-fine convolutional neural networks for pericardial effusion localization and segmentation on CT scans. In 2018 IEEE 15th international symposium on biomedical imaging (ISBI 2018) (pp. 1092–1095)
Wilder-Smith AJ, Yang S, Weikert T, Bremerich J, Haaf P, Segeroth M, Ebert LC, Sauter A, Sexauer R (2022) Automated detection, segmentation, and classification of pericardial effusions on chest ct using a deep convolutional neural network. Diagnostics 12(5):1045
Azarbal A, LeWinter MM (2017) Pericardial effusion. Cardiol Clin 35:515–524. https://doi.org/10.1016/j.ccl.2017.07.005
Vakamudi S, Ho N, Cremer PC (2017) Pericardial effusions: causes, diagnosis, and management. Prog Cardiovasc Dis 59:380–388. https://doi.org/10.1016/j.pcad.2016.12.009
Tran HV, Charles M, Garrett RC, Kempe PW, Howard CA, Khorgami Z (2020) Ten-year trends in traumatic cardiac injury and outcomes: a trauma registry analysis. Ann Thorac Surg 110(3):844–848. https://doi.org/10.1016/j.athoracsur.2019.12.038
Acknowledgements
We thank the Department of Surgery Section of Trauma and Acute Care Surgery and Ms. Heidi A. Wing, Trauma Registry Supervisor at Boston Medical Center as well as research assistants Samantha Roberts, MPH, Tyler Pina, Shinelle Kirk, Haley Connelly and all the research staff who contributed countless hours to this study. Ms. Ijeoma Okafor MPH assisted in the data analysis.
Funding
Research reported in this publication was supported by the National Institute of General Medical Sciences of the National Institutes of Health under Award Number R44GM123821. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. National Center for Advancing Translational Sciences, National Institutes of Health, through BU-CTSI Grant Number 1UL1TR001430 provided support for this study through the REDCap electronic data capture tools hosted at Boston University. Dr. Feldman is supported in part by UL1TR001430.
Author information
Authors and Affiliations
Contributions
All of the listed authors have participated actively in the entire study project, including study design, data acquisition, analysis, and manuscript preparation. ML, AV, JF developed the design and conduct of the study. IYP, ML, AV, JF participated in the data analysis, interpretation, and manuscript preparation. IYP drafted the original manuscript. All authors participated in and approved the final submission. IYP assumes responsibility for the paper as a whole.
Corresponding author
Ethics declarations
Ethics approval
This retrospective study was deemed Exempt from review by the Institutional Review Board of Boston Medical Center/ Boston University Medical Campus.
Competing interests
Dr. Megan M. Leo and Dr. Ilkay Yildiz Potter declare that they have no financial interests. Dr. Ashkan Vaziri and Dr. James Feldman received the research grants funding this work as investigators.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Yıldız Potter, İ., Leo, M.M., Vaziri, A. et al. Automated detection and localization of pericardial effusion from point-of-care cardiac ultrasound examination. Med Biol Eng Comput 61, 1947–1959 (2023). https://doi.org/10.1007/s11517-023-02855-6
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11517-023-02855-6