A Study to Collect Imaging Data for the Validation of the Intelligent Ultrasound's ScanNav Anatomy Peripheral Nerve Block (PNB) - US v1.0

Background:

The American Society of Regional Anesthesia and Pain Medicine (ASRA) has published evidence-based assessment of ultrasound-guided regional anesthesia (Neal et al., 2010) concluding that ultrasound guidance is superior or equal to other non-ultrasound nerve localization techniques. A subsequent publication from ASRA (Neal et al., 2016), has strengthened their position of ultrasound guidance being superior than other methods, including for the reduction of local anesthetic systemic toxicity. However, ultrasound-guided regional anesthesia (UGRA) is a difficult technique to master. A key activity of UGRA is ultrasound image interpretation (Sites et al., 2009) which ScanNav Anatomy PNB is designed to support. The data collected during this study will be assessed by a panel of intended users (experts in UGRA) to evaluate the performance and safety of ScanNav Anatomy PNB device highlighting.

Statistical Methodology:

Validation analyses will be conducted once the data collection has been complete. The collected scans will be processed, and the device output will be generated post hoc. Device output will be presented with raw ultrasound side-by-side. A panel of at least three expert anesthesiologists will review and evaluate each processed scan. The majority view of the panel will be used to evaluate each endpoint for any given structure ScanNav Anatomy PNB is intended to highlight.

Data collection and scan processing:

40 different subjects will be scanned. The dataset will be balanced to contain approximately equal numbers of subjects with BMI<30 and BMI>= 30 kg/m2.

All data collection will be performed with FDA cleared general purpose ultrasound machine, ScanNav Anatomy PNB device will not be used during data collection.

Data characteristics for scan subjects (e.g., age and BMI) will be reported. Ultrasound scans for all 9 supported anatomical regions will be collected from both sides of each subject.

90 x 10s clips per supported anatomical region will be generated, consisting of:

• 80 x 10s scene clips: the block view (chosen by the expert scanner) together with the preceding 10 seconds of ultrasound scanning will be recorded (without the use of ScanNav Anatomy PNB)
• 10 x 10s non-scene clips: 10 second ultrasound scans will be recorded at non-optimum block views, chosen by the expert scanner to represent plausible scanning errors (without the use of ScanNav Anatomy PNB) Scenes and non-scenes will be analyzed separately. Unmodified ultrasound video and highlighted video (color overlay produced by ScanNav Anatomy PNB generated post-hoc) will be presented side-by-side to independent experts for data analysis.

Data analysis:

Every clip will be presented to a minimum of 3 independent expert reviewers. All clips from a single anatomical region will be reviewed by the same 3 reviewers. Experts may review more than one anatomical region, but not necessarily all anatomical regions. Thus, a range of experts will review all anatomical regions.

Reviewers will be asked structured questions to assess the highlighting of safety critical anatomical structures (see definitions later in document) and the performance of ScanNav Anatomy PNB output for each individual clip.

The majority opinion (at least 2/3) will be obtained to establish the overall panel opinion (e.g., yes/yes/no = yes) for each structure on each clip.

Data will be evaluated and presented by structure in each anatomical region and overall (i.e., total for each class; nerve, artery etc.).

Inter-rater agreement between the reviewers will be reported on an anatomical region basis.

Data will be presented as frequencies and presented as percent of total clips analyzed.

Data analysis will include stratification by subject age, BMI, and ultrasound machine to ensure consistency across these variables.