以多頻近似模型評估超音波對比劑低頻影像之訊雜比

Abstract

微氣泡的低頻反應是最近才開始被研究，在頻譜上可以觀察到低頻反應的中心頻率是低於基頻反應(fundamental response)與次諧振反應(subharmonic response)，為並以此低頻反應為基礎運用於成像上，可以得到很好的微氣泡與組織的信雜比(SIR)。 目前在理論上有雙頻近似模型可以預估微氣泡低頻反應的頻寬，但是卻無法描述完整連續的頻譜並推估重疊的程度。因此，推展了多頻近似模型(Multi-Frequency Approximation Model)從頻譜的角度來描述微氣泡的低頻反應和基頻反應的頻譜，並藉以得到理論上的微氣泡與組織的信雜比。 在本研究中先以電腦模擬並運用多頻近似模型，配合不同的頻寬、不同的微氣泡半徑分佈與參數設定，得到模擬總微氣泡反應的頻譜，可以觀察到不同的微氣泡半徑分佈及頻寬對於頻譜及SIR的影響。再以實驗將模擬人體的仿體及水管內的微氣泡分佈取得低頻反應的影像來驗證，SIR大致會隨著頻寬增大而變差，並可以多頻近似模型做為濾波器設計、觀察微氣泡半徑分佈影響等的選擇。

Low-frequency response of microbubbles was found recently, the center frequency of low-frequency response was lower than ones of fundamental and subharmonic response. Based on the low-frequency response, good SIR could be made. Two-frequency approximation might evaluate the bandwidth of spectrum but could not describe the continuous spectrum of response and the extent of overlap. Hence, the multi-frequency approximation model was developed to evaluate the signal-to-interference ratio and to describe the continuous spectrum of fundamental and low-frequency response. Computer simulations were performed toinvestigate the spectrum of microbubbles and the bandwidth of signal. Experiments on phantoms could make imaging by low-frequency response and got a SIR from ROI. SIR almost were worse when bandwidth was larger. Multi-frequency approximation model could provide a new technique to evaluate the filter and the influence of bubble distributions.