Tetanus
Conditions
Keywords
Muscle Wasting, Intensive Care Unit, Muscle ultrasound, Artificial Intelligence, Deep Learning
Brief summary
This study aims to investigate the feasibility of using a real-time artificial intelligent (AI)-assisted tool for Rectus Femoris cross sectional area measurement from muscle ultrasound to improve reliability, reduce inter- and intra-observer variability and reduce operator time spent on ultrasound examination
Detailed description
This project proposes to develop computational methods to automatically analyze conventional 2D muscle ultrasound images in real time to assist operators circumvent achieve high quality reproducible views and measurements specifically for Rectus Femoris muscle. Study design: This is a prospective observational study to test the reliability of AI-assisted muscle ultrasound at the patient's bedside compared to standard RFCSA ultrasound. All measurements will be performed in adult patients with severe tetanus (Ablett Grade 3 or 4) admitted to the Adult ICU at HTD expected to stay at least 5 days. All patients are on mechanical ventilation, muscle relaxation and neuromuscular blockers following the Ministry of Health guidelines. Study procedures: Three ultrasound examinations will be carried out according to a standard operating procedure where patients are in the supine position with the leg in neutral rotation. Measurements will be taken using 12L-RS linear probe, Venue Go ultrasound machine (General Electric Healthcare, London, UK). Statistical analysis: Study will compare the intra- and interobserver variability of measurements and examination duration. All statistical analysis was performed with R version 4.0.4.
Interventions
DEVICEReal-time AI-assisted muscle ultrasound
RAIMUS software provides automatic segmentation and size measurement for the RFCSA
Sponsors
King's College London
Oxford University Clinical Research Unit, Vietnam
Study design
Allocation
RANDOMIZED
Intervention model
PARALLEL
Primary purpose
SUPPORTIVE_CARE
Masking
NONE
Intervention model description
The developed AI assistant, named RAIMUS, was deployed in real-time using the PRETUS tool. The ultrasound machine HDMI output was connected to the laptop via a USB framegrabber. This allowed the user to use an external screen with an AI overlay instead of the screen of the ultrasound machine. The interface to RAIMUS is as follows. On the right of the screen, there is a widget containing information from the automatic muscle segmentation, including the muscle delineation continuously overlaid onto the ultrasound image and the corresponding cross-sectional area in cm2. The segmentation overlay and related information can be enabled or disabled by the user.
Eligibility
Sex/Gender
ALL
Age
16 Years to No maximum
Healthy volunteers
No
Inclusion criteria
* Age ≥16 years * Written informed consent * Staff and equipment available for ultrasound * Admitted to Viet Anh Ward ICU with a diagnosis of meningitis or encephalitis or Ablett Grade 3 or 4 tetanus * Within 72 hours of ICU admission * Duration of ICU stay expected at least 5 days
Exclusion criteria
* Informed consent not given * Contraindication to ultrasound scan
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Reproducibility of RFCSA measurements | during the study procedure | In this trial, the users are randomly assigned to scan muscle ultrasound with and without AI-assisted software to measure the size of the Rectus Femoris muscle. The investigators will compare the reliability and agreement metrics of the RF measurement |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Time spent on ultrasound examination | during the study procedure | In this trial, the users are randomly assigned to scan muscle ultrasound with and without AI-assisted software to measure the size of the Rectus Femoris muscle. The investigators will record the time needed to carry out the muscle ultrasound examinations |
Countries
Vietnam
Outcome results
None listed