基於混合實境導引之低能量聚焦式超音波脊椎神經調控系統

Abstract

聚焦式超音波治療以其高選擇性與非侵入性特性，成為重要的無創醫療技術。然而，現行影像導引多以平面螢幕呈現影像，醫師須在病患與螢幕間頻繁切換視線，增加操作負擔與配準誤差。本研究提出一套基於混合實境（Mixed Reality, MR）的低能量聚焦式超音波（LIFU）脊椎神經調控導引平台，旨在提升導引精度並簡化臨床流程。系統結合基於聲場模擬與商用診斷探頭整合所設計之共軸一體化雙模探頭，低延遲即時影像擷取與串流至MicrosoftHoloLens2之虛實疊合顯示，以及以Kabsch註冊演算法為核心之多模態配準。經由水槽聲場量測驗證，所設計探頭之實測聲場與模擬結果高度一致；在水箱導引實驗中，本平台達到90%的導引成功率，且HoloLens2端畫面更新率可維持在50–60fps，提供流暢的操作體驗。研究結果顯示，本平台於成像—治療配準、即時互動與操作便利性方面具有可行性，為LIFU應用於脊椎神經調控提供一條具臨床潛力的新方向。

Focused ultrasound (FUS) therapy, with its high selectivity and noninvasive nature, has emerged as a significant technique in modern noninvasive medicine. However, conventional image-guided systems typically display imaging information on flat monitors, forcing physicians to frequently shift their gaze between the patient and the screen, which increases cognitive load and introduces potential registration errors. This study proposes a mixed reality (MR)-based low-intensity focused ultrasound (LIFU) spinal neuromodulation guidance platform, aiming to enhance targeting accuracy while simplifying clinical workflow. The system integrates a co-axial dual-mode probe designed through acoustic field simulation and commercial diagnostic ultrasound integration, a low-latency real-time image acquisition and streaming pipeline for holographic visualization on Microsoft HoloLens 2, and a multimodal registration framework based on the Kabsch algorithm. Acoustic field measurements demonstrated strong consistency between experimental results and simulations. In water-tank guidance experiments, the proposed platform achieved a 90% targeting success rate, while maintaining a frame refresh rate of 50–60 fps on the HoloLens 2, ensuring a smooth interactive experience. These results demonstrate the feasibility of the proposed platform in improving imaging–therapy registration, real-time interaction, and operational efficiency, offering a promising direction for applying LIFU in spinal neuromodulation.