Artificial Intelligence for Ultrasound-Guided Peripheral Nerve and Plane Blocks
Study Details
Study Description
Brief Summary
The goal of this observational study is to test the accuracy of an artificial intelligence tool used for identifying ultrasound-guided block regions in healthy volunteer participants.
The main question aims to answer is:
• Is the artificial intelligence tool effective for identifying selected ultrasound-guided nerve block regions and their anatomical landmarks?
Three anesthesiology trainees perform ultrasound scanning for 8 nerve block regions on participants. Peripheral nerve and plane block regions are;
-
Adductor canal block region
-
Axillary brachial plexus block region
-
ESP (erector spinae plane) block region
-
Femoral block region
-
PECS (pectoral) block region
-
Popliteal block region
-
Rectus sheath block region
-
Superficial cervical plexus block region
| Condition or Disease | Intervention/Treatment | Phase |
|---|---|---|
|
Detailed Description
Sonoanotomy knowledge is essential for ultrasound-guided regional anesthesia (UGRA) procedures. We aimed to assess the accuracy of artificial intelligence (AI) software used to assist sonoanatomy interpretation by highlighting anatomical structures in peripheral nerve and plane blocks in recognizing anatomical structures.
All scans were performed with an ultrasound device (GE Logiq, Wisconsin, USA) having AI software (Nerveblox, Smart Alfa Teknoloji San. Ve Tic. A.Ş., Ankara, Turkey). Using this setup, when a user performs an ultrasound scan, the AI software provides the user with real-time feedback about the identification of anatomical structures/landmarks.
The AI software is designed to provide three major feedback signals to the user in real-time;
-
name tags for each anatomical structure
-
color overlays for each anatomical structure
-
scan success rate for the entire image
Color overlays and name tags are transparency-adjusted highlights and dots that provide the user with more general spatial feedback on the anatomical layout. The plane completeness rate is visualized with a "scan success" gauge, which guides the user in a way that shows how close the current image is to the ideal visualization of predefined landmarks.
The Peripheral nerve and plane block regions (their anatomical landmarks) that the AI software can identify are;
-
Adductor Canal Block Region (Femoral artery, Sartorius muscle, Saphenous nerve, and Vastus medialis muscle)
-
Axillary Brachial Plexus Block Region (Axillary artery, Biceps brachii muscle, Coracobrachialis muscle, Conjoint tendon, Musculocutaneous nerve, and Triceps brachii muscle)
-
ESP (Erector Spinae Plane) Block Region (Erector spinae muscle, Pleura, Rhomboid muscle, Trapezius muscle, and Transverse process)
-
Femoral Block Region (Femoral artery, Femoral vein, Femoral nerve, and Iliopsoas muscle)
-
PECS (Pectoral) Block Region (Pleura, Pectoralis minor muscle, Pectoralis major muscle, Rib, and Serratus anterior muscle
-
Popliteal Block Region (Common peroneal nerve, Popliteal artery, Popliteal vein, and Tibial nerve)
-
Rectus Sheath Block Region (Peritoneal cavity, Rectus abdominis muscle, Anterior rectus sheath, and posterior rectus sheath)
-
Superficial Cervical Plexus Block Region (Anterior scalene muscle, Carotid artery, Cervical plexus, Jugular vein, and Sternocleidomastoid muscle)
For the study, three anesthesiology trainees who were trained in regional anesthesia and qualified to perform UGRA techniques will scan each volunteer with the guidance of AI software. In total, three residents will perform scans of 8 block types for all 40 volunteers. All scan images will be saved. Using this procedure, 960 ultrasound images will be acquired in both raw and AI-processed forms for expert assessment.
An anesthesiologist with expert knowledge of ultrasound-guided regional anesthesia techniques, and a radiologist with extensive experience in ultrasound will review and score the accuracy of the AI software on the acquired ultrasound images. To obtain a more precise result, the assessment of the AI software will be performed separately for each anatomical structure of the selected block regions.
The experts are asked to evaluate and rate (0: mislocated, 1: very poor, 2: poor, 3: good, 4:
very good, 5: excellent) the name tags and color overlays placed by the AI software. If a name tag (represented by a dot and abbreviation of the structure name) for an anatomical structure is located in a way that it is not within the visual boundaries of the anatomical structure, then the score will be "0: mislocated." If a name tag for an anatomical structure is correctly placed within the visual boundaries and able to represent the anatomical structure, then the score should be between "1: very poor" and "5: excellent," according to the consistency of the surrounding color overlay and the underlying anatomical structure.
Data will be analysed by using SPSS 26 software at a 95% confidence level. For the measurements, the mean, standard deviation (SD), minimum, maximum, and median statistics will be provided. Because the "score" variable is an ordinal measurement between 0 and 5 and does not provide a normal distribution in the regions, non-parametric methods will be used in the analysis.
Study Design
Outcome Measures
Primary Outcome Measures
- Validation of real-time identification of anatomical landmarks associated with selected peripheral nerve and plane blocks via AI supported ultrasound practice [After collecting and saving all scans/images performed by the anesthesiology trainees in one day, rating/scoring of all these saved raw and highlighted ultrasound scans/images by the experts in one day, single point]
In 40 healthy volunteer participants, AI supported ultrasound was used to scan each peripheral nerve and plane block to highlight the block-specific anatomical landmarks (by the three anesthesiology trainees). Then, expert practitioners score/rate the accuracy of color overlays using a 6-point scale (between 0 to 5) for a total of 4,440 anatomical landmarks by assessing raw and highlighted ultrasound images.
Secondary Outcome Measures
- Difference according to BMI and gender [After saving all ultrasound scans/images in one day, rating/scoring in one day, single point,]
To evaluate whether there is a difference in score according to BMI and gender
Eligibility Criteria
Criteria
Inclusion Criteria:
- Volunteers over the age 18
Exclusion Criteria:
-
anatomical deformity in the selected regions
-
psychiatric or neurological diseases that would impair understanding of the consent form
-
inability to lie flat
Contacts and Locations
Locations
| Site | City | State | Country | Postal Code | |
|---|---|---|---|---|---|
| 1 | Gazi University | Ankara | Turkey |
Sponsors and Collaborators
- Gazi University
- Smart Alfa Teknoloji San. ve Tic. A.S.
Investigators
- Principal Investigator: Irfan Gungor, Asoc.prof., Gazi University
Study Documents (Full-Text)
None provided.More Information
Publications
- Bowness J, Varsou O, Turbitt L, Burkett-St Laurent D. Identifying anatomical structures on ultrasound: assistive artificial intelligence in ultrasound-guided regional anesthesia. Clin Anat. 2021 Jul;34(5):802-809. doi: 10.1002/ca.23742. Epub 2021 May 11.
- Gungor I, Gunaydin B, Oktar SO, M Buyukgebiz B, Bagcaz S, Ozdemir MG, Inan G. A real-time anatomy identification via tool based on artificial intelligence for ultrasound-guided peripheral nerve block procedures: an accuracy study. J Anesth. 2021 Aug;35(4):591-594. doi: 10.1007/s00540-021-02947-3. Epub 2021 May 19.
- E.595466