Special Article

Optimising artificial intelligence ultrasound tools in

anaesthesiology and perioperative medicine: The
next frontier for advanced technology application

Address for correspondence: Anastasia Jones1,2, Ryan Tang2, Anahita Dabo‑Trubelja3, Cindy B. Yeoh1,2,
Dr. Anastasia Jones,
MD12902 Magnolia Drive
Leshawn Richards1,2, Vijaya Gottumukkala4
MCB‑Anest Tampa, FL 33612.
1
Department of Anesthesiology, Moffitt Cancer Center, Tampa, FL, 2Department of Oncologic Sciences,
Morsani College of Medicine, University of South Florida, Tampa, FL, 3Department of Anesthesiology,
E‑mail: Anastasia.Jones@
Memorial Sloan Kettering Cancer Center, New York, NY, 4Department of Anesthesiology, MD Anderson
moffitt.org
Cancer Center, Houston, Texas, USA
Submitted: 31-May-2024
Revised: 20-Aug-2024
Accepted: 22-Aug-2024 ABSTRACT
Published: 26-Oct-2024
Artificial intelligence (AI) was once considered avant‑garde. However, AI permeates every industry
today, impacting work and home lives in many ways. While AI‑driven diagnostic and therapeutic
applications already exist in medicine, a chasm remains between the potential of AI and its
Access this article online clinical applications. This article reviews the status of AI‑powered ultrasound (US) applications
in anaesthesiology and perioperative medicine. A literature search was performed for studies
Website: https://journals.lww.
com/ijaweb examining AI applications in perioperative US. AI applications for echocardiography and regional
anaesthesia are the most robust and well‑developed. While applications are available for lung
DOI: 10.4103/ija.ija_578_24
imaging and vascular access, AI programs for airway and gastric US imaging solutions have yet
Quick response code
to be available. Legal and ethical challenges associated with AI applications need to be addressed
and resolved over time. AI applications are beneficial in the context of education and training. While
low‑resource settings may benefit from AI, the financial burden is a considerable limiting factor.

Keywords: Anaesthesiology, artificial intelligence, machine learning, medication, monitoring,

perioperative medicine, point‑of‑care ultrasound, regional anaesthesia

INTRODUCTION without human intervention [Figure 1]. Examples

of deep learning applications include programming
What exactly is artificial intelligence (AI)? If you ask suggestions by streaming services and language
ChatGPT 3.5, AI is the simulation of human‑intelligence processing by virtual assistants. Details of techniques in
processes by computer systems, and the simulation AI and specific types of learning are beyond the scope
enables the systems to perform tasks that require of this review; thus, the authors refer an interested
human intelligence, such as learning, reasoning, reader to a review by Hashimoto et al.[3]
problem‑solving, perception, and language processing.[1]
Professor John McCarthy, credited with coining the term The expansive growth of AI has generated useful
artificial intelligence, defined AI in simpler terms in 1955 tools for optimising operator experience and task
as ‘the science and engineering of making intelligent
This is an open access journal, and articles are distributed under the terms of
machines’.[2] Today, AI is any technique that enables the Creative Commons Attribution‑NonCommercial‑ShareAlike 4.0 License,
computers to mimic human intelligence. Machine which allows others to remix, tweak, and build upon the work non‑commercially,
as long as appropriate credit is given and the new creations are licensed under
learning is a subset of AI and includes techniques that the identical terms.
enable machines, through experience, to improve at For reprints contact: [email protected]
specific tasks [Figure 1]. Examples of machine learning
include credit card fraud detection and automatic How to cite this article: Jones A, Tang R, Dabo‑Trubelja A,
Yeoh CB, Richards L, Gottumukkala V. Optimising artificial
friend‑tagging suggestions on social media. Machine intelligence ultrasound tools in anaesthesiology and perioperative
learning also includes deep learning, which relies on medicine: The next frontier for advanced technology application.
artificial neural networks to reach accurate conclusions Indian J Anaesth 2024;68:1016-21.

1016 © 2024 Indian Journal of Anaesthesia | Published by Wolters Kluwer - Medknow

Page no. 88
 Jones, et al.: Ultrasound in anaesthesiology: Powered by AI

USA), the first FDA‑approved AI device for flagging

irregularities in colonoscopy.[7] Initial studies showed
GI Genius to be reasonably helpful—non‑inferior to
expert endoscopists (clinicians who performed >2000
colonoscopies). Using GI Genius, non‑experts showed
a 21% improvement in detecting adenomas.[8,9]

In anaesthesiology, we rely heavily on the US as

a diagnostic tool. US imaging, also affectionately
known as (POCUS), enables real‑time imaging for
immediate bedside assessment and decision‑making.
Furthermore, developing handheld US devices has
Figure 1: Conceptual relationships of deep and machine learning under
the study and development of artificial intelligence been highly favourable because they are affordable,
portable, and accessible in various healthcare
automation. Potential benefits of AI‑powered programs settings (e.g. in‑patient hospitals, ambulatory surgical
in anaesthesiology have been outlined in six major centres, and rural and remote settings).[10]
themes: (1) anaesthesia monitoring, (2) medication
administration, (3) event and risk prediction, (4) Analysis of US‑acquired images with AI has posed
ultrasound (US) guidance, (5) pain management, a challenge because of the dynamic nature of image
and (6) operating‑room logistics.[3] This review focuses acquisition, variability of technique and anatomy,
on AI applications in ultrasonography for the clinical and intrinsic noise. The development of machine
anaesthesiologist. learning at the turn of the century and of deep learning
in the 2010s has made it possible to develop image
METHODS procurement and interpretation algorithms to identify
objects in the image and interrogate the video clip
We performed a literature search on PubMed using for functional information, such as lung movement,
the keywords ‘artificial intelligence’, ‘ultrasound’, and ejection fraction, valve function, and tracking a needle
‘anest*’ over the past five years (January 2020–May tip. Major US manufacturers are racing to develop
2024), resulting in 305 publications. The number of AI algorithms to augment US image acquisition and
publications dramatically increased over the past analysis, and some of these algorithms are already
three years—67 in 2021, 78 in 2022, and 79 in 2023. available in a limited capacity.
There were already 76 in 2024. Articles unrelated to
US or anaesthesiology (e.g. cancer diagnosis) were Features of AI algorithms with the greatest potential
excluded. All articles were in English. They included to augment US imaging include image optimisation
articles describing the use of AI in ultrasonography in assistance, anatomy identification, estimation of
perioperative anaesthesiology, specifically applications cardiac function, cardiac‑valve interrogation, and
in regional anaesthesia (RA), echocardiogram, lung quantification of fluid volume.
imaging, and vascular access.
Transthoracic echocardiogram
AI‑POWERED APPLICATIONS IN ANAESTHESIOLOGY Although technically challenging, a transthoracic
echocardiogram is a great tool for an expedient
AI in point‑of‑care ultrasound (POCUS) cardiac examination to estimate cardiac function and
One of the most dazzling applications of AI is the characterise valvular disease with limited information.
capacity for image and video analyses. One of the Fortunately, much of the progress in image analysis has
first AI‑powered applications involved medical been specific to the cardiac examination, with most US
imaging. Numerous assistive AI programs have manufacturers offering at least some assistive features.
been proposed for imaging in radiology, such as One comprehensive AI product that is commercially
computed tomography–enhanced cancer detection, available is a proprietary algorithm named Kosmos,
and some of these programs are already commercially developed by EchoNous (Redmond, WA, USA).
available.[4‑6] In addition, significant progress has been Kosmos uses AI to assist the anaesthesiologist with
made in analysing video imaging in medicine; one identifying and interrogating cardiac anatomy,
example is GI Genius (Medtronic, Minneapolis, MN, calculating ejection fraction, and assessing ventricular

Indian Journal of Anaesthesia | Volume 68 | Issue 11 | November 2024 1017

Page no. 89
 Jones, et al.: Ultrasound in anaesthesiology: Powered by AI

function [Figure 2].[11] Another helpful feature of interpretation of prospective real‑time lung US from
Kosmos is the immediate feedback during image an average of 68.1% to 93.4% (P < 0.001).[15] In the
acquisition regarding probe placement, tilt, and wake of the coronavirus disease 2019 (COVID‑19)
rotation. Feedback on image optimisation helps the pandemic, another study found that the AI application
user become skilful over time, and the feedback helps of guided B‑line quantification was used to detect
in a setting where POCUS experts are not available COVID‑19 pneumonia.[16] Kuroda et al. found that
for consultation.[12] In a single‑centre study by Baum this AI function correlated well with computed
et al.,[13] a randomised investigation showed that tomography findings of pneumonia, with an accuracy
novice users with POCUS devices equipped with of 94.5% for 12 zones (bilateral anterior, lateral, and
AI functionality had a shorter apical 4‑chamber posterior chest) and 83.9% for eight zones (bilateral
acquisition time and higher image‑quality scores. anterior and lateral chest).

US imaging of the lungs Airway and gastric US

POCUS facilitates the assessment of critical lung Emerging trends in using POCUS for the gastric
pathologies in the operating room and the intensive care US include evaluating a patient’s aspiration risk to
unit. Most novices can appreciate the absence of normal mitigate aspiration events, particularly for patients
lung‑sliding motion in the pneumothorax. Hence, with delayed gastric emptying, patients using
some of the most helpful features of AI algorithms in glucagon‑like peptide‑1 receptor agonists, and acutely
lung US include quantifying B‑lines in pleural oedema and critically ill children.[17,18] Mathematical models
and quantifying pleural effusions to assist clinical have been developed to estimate gastric volume, but
diagnosis and management decisions.[14] A recent these do not account for additional risk factors like
prospective study showed a potential supportive role the particulate nature of gastric contents. An AI‑tool
of AI algorithms for lung US for non‑expert clinicians solution would greatly benefit perioperative care
in low‑resource settings in Vietnam.[15] In this study, if it enabled the anaesthesiologist to quantify and
clinicians were asked to identify the findings in a characterise these risk factors.[19,20] Pre‑anaesthetic
series of 10 US video clips for patients hospitalised airway evaluation using the US is another developing
with dengue fever and septic shock, and the clinician’s area of POCUS. Literature suggests the distance
responses were compared to the expert’s assessment. from skin to epiglottis is currently the most accurate
The study found that when supported by the real‑time predictor index of a difficult laryngoscopy.[21]
AI‑assisted lung US system for interpreting lung‑US However, there is a dearth of literature on using AI
clips, non‑experts improved their accuracy of for US airway examinations; thus, explorative studies
will prove beneficial.

Vascular‑access US
In the operating room, the US is commonly used for
vascular access. For practised clinicians, using the US
for common procedures, such as central venous and
arterial access, is second nature. For less‑practised
clinicians (i.e. those with no US or interventional
skills to obtain central femoral venous access,
especially during critical haemorrhage), a handheld
AI‑guided US device has been developed to direct
them.[22] This device (1) identifies the femoral vein, (2)
directs the user in image optimisation, (3) identifies
a safe needle‑insertion point, (4) deploys the needle,
and (5) prompts the user to advance the guidewire
for catheter placement. So far, this device has only
Figure 2: Transthoracic echocardiogram parasternal long‑axis
image obtained with Kosmos. AI‑powered labeling appropriately
been tested on phantoms and porcine models.[22]
labels structures‑ right ventricle, left ventricular outflow track, left Yet, continued research and improvement on similar
ventricle, mitral valve, and left atrium. AO = aorta, AV = aortic valve, AI‑guided technologies could provide support
IVS = intraventricular septum, LA = left atrium, LV = left ventricle,
LVOT = left ventricular outflow track, MV = mitral valve, RV = right for trainees or novice clinicians in low‑resource
ventricle settings. However, the potential benefits of AI‑guided

1018 Indian Journal of Anaesthesia | Volume 68 | Issue 11 | November 2024

Page no. 90
 Jones, et al.: Ultrasound in anaesthesiology: Powered by AI

technologies must be compared to the risks for with 95%–100% accuracy.[27] Nerveblox, developed in
inexperienced operators. 2020, presents colour‑labelled anatomical structures
of US images to guide RA procedures.[26,28] Various
AI in US‑guided regional anaesthesia machine learning models are also being developed
The field of RA relies heavily on the US to acquire to further improve target detection and tracking
images and guide procedures. Pros of using US in algorithms.[24]
RA include portability, absence of radiation, direct
and real‑time visualisation of anatomic structures, One of the main advantages of AI in US‑guided RA
and local anaesthetic spread.[23,24] However, the is structure identification for non‑expert users.[27,29,30]
success of a block is often operator‑dependent, with However, more than accurate structure identification
outcomes affected by factors such as a physician’s is needed to ensure a safe and effective block. Accurate
anatomic knowledge, inattentional blindness, fatigue, imaging and identifying pertinent structures are
technique, and anatomical challenges, such as merely a foundational component of RA procedures.
obesity, trauma, and subcutaneous emphysema.[25] Safe needle trajectory, needle visualisation,
Introducing AI guidance for US imaging with real‑time placement, and injection of local anaesthetic into
detection of key structures, such as nerves, muscles, the appropriate location are all crucial components
fascia, and blood vessels, could significantly improve of an operator‑dependent, safe, and effective block.
physician performance. Currently, AI models for NeedleTrainer was built into ScanNav Anatomy US
these applications include deep convolutional neural systems to bridge this gap in developing procedural
networks. For example, the U‑Net architecture, skills. The programme uses retractable needles and
developed in 2015 by Olaf Ronneberger at the augmented‑reality technology to simulate procedural
University of Freiburg in Germany, can segregate conditions in a human patient. This system could
and segment grey images.[26] With this technology, be used to improve needling skills in regional nerve
machine learning platforms integrated into US blocks as well as vascular access.[26,31]
systems, such as ScanNav Anatomy Peripheral Nerve
Block (Intelligent Ultrasound, Cardiff, UK) and A study by Bowness et al.[30] explored the utility of
Nerveblox (Pajunk, Geisingen, Germany), have been assistive AI for US scanning in RA, and it showed
developed. ScanNav Anatomy Peripheral Nerve that both experts and non‑experts valued learning and
Block was approved by the FDA in 2022; it creates teaching US scanning for RA. Interestingly, non‑experts
colour overlays of key anatomical structures to assist were more likely to give positive feedback on using
physicians in performing RA [Figure 3]. In a clinical assistive AI, while experts observing non‑experts in
study from April 2021 on ScanNav, experts reviewed a procedure were more likely to report increased risk
the application of the technology and found that the and safety concerns.[30] Risk and safety concerns are
programme identified and recognised key structures expected with implementing any new technology. By
contrast, proceduralists performing RA blocks should
be aware of potential risks and understand how to
mitigate and address complications. Regrettably, this
type of understanding is often relegated to experienced
providers and is outside the arsenal of a novice
clinician.

CHALLENGES AND LIMITATIONS

Through assistive AI and augmented‑reality training,

image‑optimisation features with instructions for
probe placement, tilt, and rotation with repetitive
use may be helpful in provider education. However,
the traditional learning model under an experienced
clinician’s supervision is indispensable. Assistive AI
Figure 3: Ultrasound view for supraclavicular block with Nerveblox. is not intended to substitute sound clinical judgment,
Despite a grainy image, the software appropriately labels the brachial
plexus, subclavian artery, first rib, and pleural surfaces. BP = brachial which cannot be developed without ample exposure
plexus, FR = first rib, PL = pleura, SA = subclavian artery to real clinical cases.

Indian Journal of Anaesthesia | Volume 68 | Issue 11 | November 2024 1019

Page no. 91
 Jones, et al.: Ultrasound in anaesthesiology: Powered by AI

As AI‑assisted technologies are still in the preliminary improve patient care. In the future, AI will integrate
stages, detection and tracking errors and image into daily clinical practice as a tool for clinicians;
misinterpretation are risks, especially in patients with therefore, anaesthesiologists must collaborate
anatomical variations or abnormal anatomy, such as closely with engineers and scientists in AI systems
in trauma. There are also limitations in the technology development.
used for detecting osseous images. Block complications
may include block failure, needle trauma, hematoma, CONCLUSION
nerve injury, pleural injury, peritoneal injury, and local
anaesthetic systemic toxicity.[24] In addition to clinical Assistive AI has great potential for successful
challenges, the legal, financial, and ethical questions and widespread implementation in the US in
associated with implementing AI are still evolving. anaesthesiology because of its capacity to aid image
There are multiple unresolved ethical questions— acquisition, image optimisation, video interrogation,
such as who is responsible for complications when AI and feedback. While incorporation into clinical
is being utilised? In addition, what if AI and clinician practice has been lagging because of multiple barriers,
assessments are directly opposed? Additional causes AI tools in the US will succeed first as educational
of concern are legal responsibility for system errors and personal improvement tools. As AI solutions
and the use of faulty data when training these models. become increasingly robust, the AI programs for the
The White House of the United States released US have the potential to revolutionise and transform
the ‘Blueprint for AI Bill of Rights’ as a guide to AI anaesthesiology and perioperative medicine. Still,
developers and users; however, legal regulation is still they will require practising clinicians to collaborate
limited.[32] with AI developers to engage actively.

FUTURE DIRECTIONS Financial support and sponsorship

The authors have no sources of funding to declare for
A few potential benefits of AI technologies in this manuscript.
anaesthesiology include facilitating teaching
Conflicts of interest
and learning, quality improvement, reducing the
There are no conflicts of interest.
clinician’s cognitive workload, improving care
recommendations, and providing feedback relating ORCID
to standard care. However, most AI applications in Anastasia Jones: https://orcid.org/0000-0003-1699-3433
the US are still in the initial stages of research and Ryan Tang: https://orcid.org/0000-0002-6573-7893
development; thus, drawing any conclusions is Anahita Dabo: https://orcid.org/0000-0001-9473-0917
difficult.[3,33,34] Multiple barriers to clinical application Cindy B. Yeoh: https://orcid.org/0000-0002-3135-4181
remain, such as uncertainties about the generalisation Leshawn Richards: https://orcid.org/0009-0009-1268-
of algorithms, financial benefits, legal implications, 9603
and susceptibility to bias. Vijaya Gottumukkala: https://orcid.org/0000-0002-
6941-4979
Regardless of improvements in AI technology,
anaesthesiologists will continue to be essential REFERENCES
because of varied clinical cases, AI technological
errors, and physician‑patient relationships. Because 1. ChatGPT (Version 3.5) [Large language model] OpenAI;
of the wide variability in clinical case type and c2015‑2024. Available from: https://chat.openai.com/.
2. Andresen SL. John McCarthy: Father of AI. IEEE Intelligent
complexity, anaesthesia‑delivery systems should Systems 2002;17:84‑5.
never function autonomously without clinician 3. Hashimoto DA, Witkowski E, Gao L, Meireles O, Rosman G.
Artificial intelligence in anesthesiology: Current techniques,
supervision. Provider‑rescue interventions will always clinical applications, and limitations Anesthesiology
be needed if AI makes an unexpected decision or gives 2020;132:379‑94.
a warning during treatment. Technological errors are 4. Kiraly AP, Cunningham CA, Najafi R, Nabulsi Z, Yang J, Lau C,
et al. Assistive AI in lung cancer screening: A retrospective
also inevitable with AI systems; therefore, AI systems multinational study in the United States and Japan. Radiol
need constant clinician supervision.[33] AI systems will Artif Intell 2024:6:e230079. doi: 10.1148/ryai.230079
not replace anaesthesiologists in the future. Still, these 5. Chassagnon G, De Margerie‑Mellon C, Vakalopoulou M,
Marini R, Hoang‑Thi TN, Revel MP, et al. Artificial intelligence
systems have the potential to reduce the cognitive load in lung cancer: Current applications and perspectives. Jpn J
on physicians so they can focus on pertinent tasks that Radiol 2023;41:235‑44.

1020 Indian Journal of Anaesthesia | Volume 68 | Issue 11 | November 2024

Page no. 92
 Jones, et al.: Ultrasound in anaesthesiology: Powered by AI

6. Sechopoulos I, Teuwen J, Mann R. Artificial intelligence assessment of gastric volume by gastroscopic examination.
for breast cancer detection in mammography and digital Anesth Analg 2013;116:357-63.
breast tomosynthesis: State of the art. Semin Cancer Biol 20. Tacken MCT, van Leest TAJ, van de Putte P, Keijzer C,
2021;72:214-25. Perlas A. Ultrasound assessment of gastric volumes of thick
7. Repici A, Badalamenti M, Maselli R, Correale L, Radaelli F, fluids: Validating a prediction model. Eur J Anaesthesiol
Rondonotti E, et al. Efficacy of real-time computer-aided 2021;38:1223-9.
detection of colorectal neoplasia in a randomized trial. 21. Carsetti A, Sorbello M, Adrario E, Donati A, Falcetta S. Airway
Gastroenterology 2020;159:512-20.e7. ultrasound as predictor of difficult direct laryngoscopy:
8. Repici A, Spadaccini M, Antonelli G, Correale L, Maselli R, A systematic review and meta-analysis. Anesth Analg
Galtieri PA, et al. Artificial intelligence and colonoscopy 2022;134:740-50.
experience: Lessons from two randomised trials. Gut 22. Brattain LJ, Pierce TT, Gjesteby LA, Johnson MR, DeLosa ND,
2022;71:757-65. Werblin JS, et al. AI-enabled, ultrasound-guided handheld
9. Biffi C, Salvagnini P, Dinh NN, Hassan C, Sharma P, robotic device for femoral vascular access. Biosensors (Basel)
Group GIGCS, Cherubini A. A novel AI device for real-time 2021;11:522. doi: 10.3390/bios11120522
optical characterization of colorectal polyps. NPJ Digit Med 23. Marhofer P, Greher M, Kapral S. Ultrasound guidance in
2022;5:84. doi: 10.1038/s41746-022-00633-6 regional anaesthesia. Br J Anaesth 2005;94:7-17.
10. Naji A, Chappidi M, Ahmed A, Monga A, Sanders J. 24. Viderman D, Dossov M, Seitenov S, Lee MH. Artificial
Perioperative point-of-care ultrasound use by anesthesiologists. intelligence in ultrasound-guided regional anesthesia:
Cureus 2021;13:e15217. doi: 10.7759/cureus.15217 A scoping review. Front Med (Lausanne) 2022;9:994805. doi:
11. Kosmos by EchoNous. EchoNous, Inc; c2015-2024. 10.3389/fmed. 2022.994805
Avaliable from: https://echonous.com/kosmos-plus/?utm_ 25. Lloyd J, Morse R, Taylor A, Phillips D, Higham H,
term=ultrasound%20ai&utm_campaign=&utm_ Burckett-St Laurent D, et al. Artificial intelligence:
source=adwords&utm_medium=ppc&hsa_acc=3477071792 Innovation to assist in the identification of sono-anatomy for
&hsa_cam=20562115886&hsa_grp=161634935400 &hsa_ ultrasound-guided regional anaesthesia. Adv Exp Med Biol
ad=691771426280&hsa_src=g&hsa_tgt=kwd-530725551486 2022;1356:117-40.
&hsa_kw=ultrasound%20ai&hsa_mt=e&hsa_ 26. Mika S, Gola W, Gil-Mika M, Wilk M, Misiołek H. Artificial
net=adwords&hsa_ver=3&gad_source=1&gclid=Cj0KCQjw8 intelligence-supported ultrasonography in anesthesiology:
J6wBhDXARIsAPo7QA-XaK4nmUH-PGegl0nqgtpjA2FpfHwM Evaluation of a patient in the operating theatre. J Pers Med
WhPTwYpKf7OuuJa3Ya_4PJYaAjjxEALw_wcB. Last assessed 2024;14:310. doi: 10.3390/jpm14030310
on July 2024. 27. Bowness J, Varsou O, Turbitt L, Burkett-St Laurent D.
12. Serrano RA, Smeltz AM. The promise of artificial intelligence- Identifying anatomical structures on ultrasound: Assistive
assisted point-of-care ultrasonography in perioperative care. artificial intelligence in ultrasound-guided regional anesthesia.
J Cardiothorac Vasc Anesth 2024;38:1244-50. Clin Anat 2021;34:802-9.
13. Baum E, Tandel MD, Ren C, Weng Y, Pascucci M, Kugler J, et al. 28. Larkin HD. FDA approves artificial intelligence device for
Acquisition of cardiac point-of-care ultrasound images with guiding regional anesthesia. JAMA 2022;328:2101. doi:
deep learning: A randomized trial for educational outcomes 10.1001/jama.2022.20029
with novices. CHEST Pulmonary 2023;1:100023. 29. Gungor I, Gunaydin B, Oktar SO, M Buyukgebiz B, Bagcaz S,
14. Bhoil R, Ahluwalia A, Chopra R, Surya M, Bhoil S. Signs and Ozdemir MG, et al. A real-time anatomy ıdentification via
lines in lung ultrasound. J Ultrason 2021;21:e225-33. tool based on artificial ıntelligence for ultrasound-guided
15. Nhat PTH, Van Hao N, Tho PV, Kerdegari H, Pisani L, Thu LNM, peripheral nerve block procedures: An accuracy study.
et al. Clinical benefit of AI-assisted lung ultrasound in a J Anesth 2021;35:591-4.
resource-limited intensive care unit. Crit Care 2023;27:257. 30. Bowness JS, El-Boghdadly K, Woodworth G, Noble JA,
doi: 10.1186/s13054-023-04548-w Higham H, Burckett-St Laurent D. Exploring the utility of
16. Kuroda Y, Kaneko T, Yoshikawa H, Uchiyama S, Nagata Y, assistive artificial intelligence for ultrasound scanning in
Matsushita Y, et al. Artificial intelligence-based point-of- regional anesthesia. Reg Anesth Pain Med 2022;47:375-9.
care lung ultrasound for screening COVID-19 pneumoniae: 31. Shevlin SP, Turbitt L, Burckett-St Laurent D, Macfarlane AJ,
Comparison with CT scans. PLoS One 2023;18:e0281127. doi: West S, Bowness JS. Augmented reality in ultrasound-guided
10.1371/journal.pone.0281127 regional anaesthesia: An exploratory study on models with
17. Baldawi M, Ghaleb N, McKelvey G, Ismaeil YM, Saasouh W. potential implications for training. Cureus 2023;15:e42346.
Preoperative ultrasound assessment of gastric content in doi: 10.7759/cureus.42346
patients with diabetes: A meta-analysis based on a systematic 32. Hine E, Floridi L. The blueprint for an AI bill of rights: In
review of the current literature. J Clin Anesth 2024;93:111365. search of enaction, at risk of inaction. Minds and Machines
18. Valla FV, Tume LN, Jotterand Chaparro C, Arnold P, 2023;33:285-92.
Alrayashi W, Morice C, et al. Gastric point-of-care ultrasound 33. Naaz S, Asghar A. Artificial intelligence, nano-technology and
in acutely and critically ill children (POCUS-ped): A scoping genomic medicine: The future of anaesthesia. J Anaesthesiol
review. Front Pediatr 2022;10:921863. doi: 10.1016/j.jclinane. Clin Pharmacol 2022;38:11-7.
2023.111365 34. Char DS, Burgart A. Machine-Learning Implementation in
19. Perlas A, Mitsakakis N, Liu L, Cino M, Haldipur N, Davis L, Clinical Anesthesia: Opportunities and Challenges. Anesth
et al. Validation of a mathematical model for ultrasound Analg 2020;130:1709-12.

Indian Journal of Anaesthesia | Volume 68 | Issue 11 | November 2024 1021

Page no. 93