<p>To evaluate the feasibility of a low-cost augmented reality (AR) workflow for ultrasound-guided venipuncture and whether AR visualisation is non-inferior to conventional monitor guidance in total procedure time while improving user experience. Real-time ultrasound video from a conventional scanner was captured via an HDMI-to-USB device, pre-processed with a Python/OpenCV pipeline, and streamed over Wi-Fi to a HoloLens 2 headset using WebSocket. A turkey-thigh model with an embedded fluid-filled catheter simulated a peripheral vein. One experienced implantation nurse performed 58 venipuncture attempts with standard monitor guidance and 58 with AR (monitor blanked during AR). Total procedure time (TPT) was recorded for each attempt, and a questionnaire assessed perceived difficulty, image quality, usability, reliability, and ergonomic comfort. With all data included, overall TPT was comparable between standard and AR guidance (<InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(21.32 \pm 7.90\)</EquationSource> </InlineEquation> s vs. <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(21.64 \pm 10.30\)</EquationSource> </InlineEquation> s; <InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(p = 0.845\)</EquationSource> </InlineEquation>). After excluding predefined outliers, overall TPT remained non-significantly lower with AR (<InlineEquation ID="IEq4"> <EquationSource Format="TEX">\(21.40 \pm 6.78\)</EquationSource> </InlineEquation> s vs. <InlineEquation ID="IEq5"> <EquationSource Format="TEX">\(19.90 \pm 6.34\)</EquationSource> </InlineEquation> s; <InlineEquation ID="IEq6"> <EquationSource Format="TEX">\(p = 0.193\)</EquationSource> </InlineEquation>); on day 2 the <i>p</i>-value decreased to 0.103 (<InlineEquation ID="IEq7"> <EquationSource Format="TEX">\(20.45 \pm 6.32\)</EquationSource> </InlineEquation> s vs. <InlineEquation ID="IEq8"> <EquationSource Format="TEX">\(18.30 \pm 5.12\)</EquationSource> </InlineEquation> s). Questionnaire responses favoured AR for gaze continuity, perceived coordination, usability, and ergonomic comfort, with acceptable image quality. AR-assisted ultrasound-guided venipuncture was feasible and did not add measurable temporal overhead for an expert operator, while improving perceived ergonomics. Larger, multi-operator clinical studies are needed to confirm performance effects and clinical impact.</p>

Augmented reality for ultrasound-guided venipuncture: a pilot study on a simulated vascular access model

期刊论文
出版模式：

开放获取
发表日期：

2026年4月15日

查看全文

Cosimo Aliani,
Alberto Morelli,
Simona Cristina Aiello,
Fulvio Pinelli,
Gianluca Villa,
Stefano Romagnoli,
Leonardo Bocchi

摘要

To evaluate the feasibility of a low-cost augmented reality (AR) workflow for ultrasound-guided venipuncture and whether AR visualisation is non-inferior to conventional monitor guidance in total procedure time while improving user experience. Real-time ultrasound video from a conventional scanner was captured via an HDMI-to-USB device, pre-processed with a Python/OpenCV pipeline, and streamed over Wi-Fi to a HoloLens 2 headset using WebSocket. A turkey-thigh model with an embedded fluid-filled catheter simulated a peripheral vein. One experienced implantation nurse performed 58 venipuncture attempts with standard monitor guidance and 58 with AR (monitor blanked during AR). Total procedure time (TPT) was recorded for each attempt, and a questionnaire assessed perceived difficulty, image quality, usability, reliability, and ergonomic comfort. With all data included, overall TPT was comparable between standard and AR guidance ( \(21.32 \pm 7.90\) s vs. \(21.64 \pm 10.30\) s; \(p = 0.845\) ). After excluding predefined outliers, overall TPT remained non-significantly lower with AR ( \(21.40 \pm 6.78\) s vs. \(19.90 \pm 6.34\) s; \(p = 0.193\) ); on day 2 the p-value decreased to 0.103 ( \(20.45 \pm 6.32\) s vs. \(18.30 \pm 5.12\) s). Questionnaire responses favoured AR for gaze continuity, perceived coordination, usability, and ergonomic comfort, with acceptable image quality. AR-assisted ultrasound-guided venipuncture was feasible and did not add measurable temporal overhead for an expert operator, while improving perceived ergonomics. Larger, multi-operator clinical studies are needed to confirm performance effects and clinical impact.