近來已有研究證實核磁共振Dixon 方法可以應用於臨床上定量肝脂肪分數，利用核磁共振造影MRDixon 方法以量化脂肪肝假體的取樣訊號，某些情況可能會影響使用MR Dixon 方法的肝脂肪分數的讀取。本研究是利用MR Dixon 方法，於不同採樣點下評估其是否影響脂肪分數的讀值。利用試管製備五個不同脂肪比例的均勻假體，脂肪所占的重量百分比分別為0％，10％，20％，30％和40％。使用三維平行加速同相位和反相位容積插入式閉氣檢查技術(VIBE)，其TR / TE 反相/ TE 同相分別為7.48 / 2.38 / 4.76 ms。利用四個計算脂肪指數的公式，依序是：同相位信號強度-反相位信號強度(SIin-SIopp)、(同相位信號強度-反相位信號強度)/同相位信號強度((SIin-SIopp)/SIin)、同相位信號強度/反相位信號強度(SIin/SIopp)、脂肪影像訊號強度/水影像訊號強度(SIfat/SIwater)。在前三個公式的對角線和脂肪指數圖上發現了轉折現象，最後一個配方的SIfat和脂肪指數隨脂肪分數的增加而線性增加且無轉折點。從左到右方向的採樣位置通常對SI 和脂肪指數有輕微的影響，但從頭到尾方向的採樣位置則無影響。建議使用第四種公式的脂肪指數SIfat / SIwater 定量脂肪肝，而沒有轉折問題。由於從左到右的方向對脂肪指數造成微小偏差，因此建議在左到右方向上固定採樣點以解決此問題。

MR Dixon method has been proposed for quantifying fatty liver and showed great potential in clinical practice. However, certain conditions encountered in clinical practice may influence the reading of liver fat fraction using the MR Dixon method. This study is focused on if the site of sampling would influence the readings of fat fraction by using the MR Dixon method. Five phantoms of 0%, 10%, 20%, 30%, and 40% fat in weight percentage were prepared in vitro. A three-dimensional parallel-accelerated in-phase and opposed-phase (opp-phase) volumetric interpolated breath-hold examination (VIBE) sequence with TR/TEopp-phase/TEin-phase of 7.48/2.38/4.76 ms was used. Four formulas calculating fat indexes were included in our study. The four formulas are in order: (in-phase signal strength)-(anti-phase signal strength) (i.e. SIin-SIopp), (in-phase signal strength)- (anti-phase signal strength)/(in-phase signal strength) (i.e. SIin-SIopp)/SIin), (in-phase signal strength)/(inverse-phase signal strength) (i.e. SIin/SIopp), (fat image signal strength)/(water image signal strength) (i.e. SIfat/SIwater). We tested that if the site of sampling would influence the SIs on the MR images and the fat indexes calculated using the four formulas. The phenomenon of turning point on the plots of SIopposite and fat indexes of the former three formulas was noted. The SIfat and fat index of the last formula increased linearly with the increase of fat fraction without a turning point. The site of sampling on the left to right direction generally mildly influenced the SIs and fat indexes, but the site of sampling on the cranial to caudal direction generally did not. The fat index of the fourth formula, SIfat/SIwater, is suggested to quantify fatty liver for no observing turning point. Since the left to right direction caused a trivial bias on fat indexes, a fixed sampling site on the left to right direction is suggested to overcome this problem.