非酒精性脂肪肝臟疾病(Nonalcoholic Fatty Liver Disease, NAFLD)診斷主要是藉由非侵入式超音波影像判讀或侵入式肝組織切片來執行，而肝臟組織中的脂肪分佈並不均勻，且不同部位肝組織切片結果可能有差異度，因此當我們使用肝組織切片進行診斷的可靠度不佳。然而判讀超音波影像往往依賴醫師的經驗累積與主觀判定，因此如何針對超音波影像的組織特性進行定量分析，發展出具有臨床應用價值之組織特性描述(Tissue characterization)技術，為現階段的研究主軸。
本研究使用聲學結構量化(acoustic structure quantification, ASQ)與Nakagami兩種散射統計模型對實驗射頻訊號資料(Radiation Frequency signal data, RF data)進行分析，實驗包含三個部分；模擬實驗為了驗證Nakagami模型與ASQ模型在描述散射子為按照雷利分佈(Rayleigh distribution)之隨機均勻分佈的狀態下兩者之差異；仿體實驗使用高頻超音波系統取得不同散射濃度之碳粉仿體的RF data並使用散射統計模型分析，為了驗證模擬實驗之結果，並找出計算參數影像時最合適的窗型方框尺寸(Window size)，更進一步探討聚焦、頻率及散射濃度效應；在動物實驗中，以高脂食物餵食小鼠誘發非酒精性脂肪肝臟疾病(NAFLD)，進行12週觀察及試驗後犧牲，使用高頻超音波影像系統取得肝臟超音波影像之RF data，並進行肝臟的病理分析，最後比較病理變化與散射統計模型之分析結果。
實驗結果發現，ASQ統計模型相對較為穩定且靈敏度可能較高；本研究所使用參數分析法之窗型方框尺寸(window size)為兩倍的理論值(3*pulse length)；聚焦、頻率與散射濃度效應皆會影響參數估計結果，主要是受取樣空間及散射子數目的影響；在動物實驗的應用中，本研究使用的方法可以區分出正常肝組織與非酒精性脂肪肝組織之差異。

The diagnosis of Nonalcoholic Fatty Liver Disease was depended on noninvasive ultrasound (US) and invasive fine needle biopsy. But the distribution of adipocyte was not uniform in liver tissue. When we use fine needle biopsy the accuracy of diagnosis for liver disease may be unreliable and probably correlate with different locations of tissue. However, interpreting the US image depends on experience and subjective judgment of physician. Therefore, how to quantitative the characteristics of tissue of US image and develop a tissue characterization technology which has clinical value is a very important study.
We proposed two statistical models of scatters which called Nakagami and Acoustic Structure Quantification (ASQ) to analyze the experiment Radiation Frequency signal data (RF data). We have three experiment parts. For simulation experiment, we create data which scatters are random number and following Rayleigh distribution, try to find out the difference of these two statistical models. For phantom experiment, we use high-frequency ultrasound system to get RF data from different concentration of scatters agar phantom and analyzed by the statistical model of scattering. To verify the result of simulation experiment and find the optimal size of window used to calculate the parameter. Further exploration focal effect, frequency effect and scattering concentration effect. For animal experiment, mice are fed high fat diet to induce NAFLD (Nonalcoholic Fatty Liver Disease) and continue observation and test up to 12 weeks, and after that mice were scarified. Then analyzing the pathological of liver and compared with the results of liver (via US system) RF data which were analyzed by statistical model of scattering after experiment.
The results from experiment demonstrated that ASQ statistical model is more stable and probably more sensitive. We use two folds of theoretical window size for parametric analysis method. All of focus, frequency and scatters concentration effect would affect the result of parameter estimation. It’s mainly due to sample volume and number of scatters. In animal experiment, the method we used on this research can distinguish normal liver tissue and NAFLD liver tissue.

[1] K. Cusi,”Role of obesity and lipotoxicity in the development of non-alcoholic steatohepatitis: pathophysiology and clinical implications. “Gastroenterology, vol. 142, pp. 711-725(2012).

[2] E.M. Brunt, C.G. Janney, A.M. Di Bisceglie, et al.” Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions. “Am J Gastroenterol,vol. 94, pp. 2467-2474(1999).

[3] T. Karlas, J. Wiegand, T. Berg, “Gastronintestinal complications of obesity:Non-alcoholic fatty liver disease (NAFLD) and its sequelae.” Best Practice & Research Clinical Endocrinology & Metabolism, vol.27, pp. 195-208(2013).

[4] C. B. Burckhardt, “Speckle in ultrasound B-mode scan”, IEEE Transactions on Sonics and Ultrasonics, vol.su-25, no.1(1978).

[5] T. A. Tuthill, R. H. Sperry ,and K. J. Parker , “Deviation from Rayleigh Statistical in Ultrasonic Speckle,” Ultrason. Imag., vol. 10, pp. 81-89(1988).

[6] L. Weng, J.M. Reid, P.M. Shankar, K. Soetanto, “Ultrasound speckle analysis based on the K distribution,” J. Acoust. Soc. Am., vol. 89, pp. 2992-2995(1991).

[7] P.M. Shankar, J.M. Reid, H. Ortega, C.W. Piccoli, and B.B. Goldberg, “Use of Non-Rayleigh Statistics for the Identification of Tumors in Ultrasonic B-Scans of the Breast,” IEEE Trans. Med. Imag., vol. 12, pp. 687-692(1993).

[8] V.M. Narayanan, P.M. Shankar, and J.M. Reid, “Non-Rayleigh Statistics of Ultrasonic Backscattered Signals,” IEEE Trans. Ultrason. Ferr. Freq. Cont., vol. 41, pp. 845-852(1994).

[9] R.C. Molthen, P.M. Shankar, J.M. Reid, K. Soetanto, X.M. Lu, “Characterization of ultrasonic B-modes using non-Rayleigh statistics”, Ultrasound in Med. Bio., vol. 21, pp. 161-170(1995).

[10] P.M. Shankar, R. Molthen, V.M. Narayanan, etc. “Studies on the Use of Non-Rayleigh Statistics for Ultrasonic Tissue Characterization,” Ultrason. Med. Bio., vol. 22, pp. 873-882(1996).

[11] R. Molthen, P.M. Shankar, J.M. Reid,F.Forsberg, E.J.Halpern, C.W. Piccoli, B.B. Goldberg,” Comparisons of the Rayleigh and K-distribution models using in vivo breast and liver tissue”, Ultrasound Med. Bio., vol. 24, pp. 93-100(1998).

[12] P.M. Shankar, “A General Statistical Model for Ultrasonic Backscattering from Tissues,” IEEE Trans. Ultrason. Ferr. Freq. Cont., vol. 47, pp. 727-736(2000).

[13] P. H. Tsui and C. C. Chang "Imaging local scatterer concentrations by the Nakagami statistical model", Ultrasound Med. Biol., vol. 33, pp. 608-619(2007).

[14] W.C. Holfman, “Statistical Methods in Radio Wave Propagation,” (1960).

[15] T. Yamaguchi ,H. Hachiya, N. Kamiyama,F.Moriyasu “ Examination of the spatial correlation of statistics informationin the ultrasonic echo from diseased liver”. Jpn J ApplPhys,vol. 41, pp. 3585-3589(2002).

[16] T.Yamaguchi ,H. Hachiya , N. Kamiyama ,F.Moriyasu“Estimation of characteristic of echo envelope using RF echo signal from the liver”. Jpn J ApplPhys, vol. 40, pp. 3900-3904(2001).

[17] H.Toyoda , T. Kumada, N. Kamiyama, K. Shiraki, K. Takase, T. Yamaguchi, H. Hachiya,“B-mode ultrasound with algorithmbased on statistical analysis of signals: evaluation of liver fibrosis in patients with chronic hepatitis C.”AJR Am J Roentgenol, vol. 193, pp. 1037-1043(2009).

[18] M.L. Olleros, M.L. Martin, et al.” Fat diet and alcohol-induced steatohepatitis after LPS challenge in mice: Role of bioactive TNF and Th1 type cytokines.” Cytokine, vol. 44, pp. 118-125(2008).

[19] TOSHIBA. “A Guide to Acoustic Structure Quantification ( ASQ) Aplio XG V5 Rev2.”