^(99m)Tc-DTPA利尿劑腎圖中因曲線湊合造成半廓清時間(T_(1/2))數據偏差的陷阱

Pitfalls in Diuretic Renography: T_(1/2) Calculated by Curve Fitting

宋美慧(Mei-Hui Song) ； 鄭雯文(Wen-Wen Cheng) ； 康依婷(Yi-Ting kang) ； 杜高瑩(Kao-Yin Tu)

《臺灣醫事放射期刊》 1卷1期 (2012/11) Pp. 25-32

https://doi.org/10.6717/JTMRT.201211_1(1).0004

摘要

1960年Nordyke開始將131I-OIH用於腎圖檢查，隨著^(99m)Tc的發展及儀器、電腦技術的廣泛應用，使得腎圖術不僅能顯示泌尿系統代謝的影像外還可以藉由定量分析獲得相關定量參數，提供臨床有意義的訊息。在腎圖術中，我們常計算注射利尿劑(Lasix®)後的曲線來決定病人是否有機械性或非機械性阻塞水腎，故該如何選擇適當的湊合方式來描述我們所得的曲線，是很重要的；因此，希望經由此篇文章，提醒放射師(電腦分析師)在處理腎圖的曲線湊合(Curve- fitting)時能慎選湊合方式，方可得到合理的定量數值。材料與方法：常規的腎圖檢查是利用偵測頭(detector)紀錄放射性藥物到達兩側腎臟、在腎臟聚集以及離開腎臟排到膀胱的過程，並將此過程記錄成時間活度曲線(Time Activity Curve)。當在注射完放射性藥物開始攝影後約12-15分鐘，若發現病人腎盂有擴張積水的現象，則會施打利尿劑(Lasix®)來分辨機械性或非機械性阻塞，而這兩種的差別在於施打利尿劑後至腎臟代謝排至膀胱的活度曲線所得到的T_(1/2)。一般來說，依照數學方程式的不同，fitting 可分為：外插法(Extrapolate)、內插法(Interpolate)、定義模式(Define model)、數學式運算(Mathematically manipulated)等幾種方式。此次我們使用數學式運算中的線性湊合(linear fits)、對數湊合(logarithmic fits)、指數湊合(exponential fits)、和冪次湊合(power fits)試算曲線，比較哪種方式較能正確地描述病人的身體狀況。一般商用軟體計算基準為利尿劑注射時間到檢查終止時間，且使用的曲線湊合需事先選定，然而病人的狀況不盡相同，若以相同的曲線湊合方式來計算T_(1/2)，會造成計算結果與曲線實際情況不一致，因此我們採用線性湊合方式並訂出兩個曲線湊合的基準點來進行曲線湊合，藉此方式來減少因曲線湊合計算所造成的誤差。結果與討論：根據試算後的T_(1/2)值，可發現線性湊合最能提供臨床有意義的定量數值。在試算的過程中，我們觀察到線性湊合使得曲線完整且下降斜率間距一致，計算T_(1/2)時較容易而且準確性高；對數湊合和指數湊合，曲線雖然完整但斜率間距不等距，所以得到的T_(1/2)值過高；而冪次湊合會使得曲線太過於平坦，所推算的T_(1/2)值會偏低。結論：腎功能造影(腎圖術Renography)是常用來評估腎臟功能的檢查，但執行這項檢查時卻容易因為曲線湊合(Curve Fitting)關係而造成數據偏差的陷阱。希望可藉由此篇文章，提醒放射師(電腦分析師)在處理腎圖時，慎選湊合方式，方可得到合理的定量數值。

關鍵字

Curve Fitting ； Renography

並列摘要

Nordyke started to use ^(131)I-OIH for renography in 1960, with the development of ^(99m)Tc and the extensive application of instruments and computer technology, making renography not only display the images about metabolism of the urinary system but also get related quantitative values by quantitative analyze to provide clinical diagnose. In renography, we use to calculate the curve after injection diuretics (lasix) to determine whether a patient has a mechanical or non-mechanical obstruction hydronephrosis. How we choose the appropriate fitting to describe the data curve we have data is very important. Therefore, this article reminds radiologists (or computer analysts) to choose right curve fitting to get reasonable quantitative values when they process the renography. Methods: Conventional renography was used to detect radiopharmaceutical to flow into both sides of the kidneys, accumulation in the kidneys and excretion from the kidney into the bladder, and this process was recorded into a time activity curve. After injecting radiopharmaceutical, we begin to image for 12-15 minutes. If there is dilation of the pelvis, we will inject diuretics (lasix) to distinguish mechanical and non-mechanical obstruction hydronephrosis, depending on T_(1/2) value received from time activity curve. In general, according to the different mathematical equations, the fitting can be defined in several ways, for example, Extrapolation, Interpolation Define model, and Mathematical manipulation. We used linear fits, logarithmic fits, exponential fits, and power fits of the Mathematical manipulation to compare which method is better to correctly describe the patient's physical condition. General commercial software calculated T_(1/2) from the diuretic injection time to the end of examination, and we should pre-select the curve fitting. Different patients have different conditions, if we use the same method to calculate the curve, it will be inconsistent with the real curve. We chose the linear fits and set the benchmark for the curve, therefore reducing errors caused by calculation. Results and discussion: According the T_(1/2) values, we find that linear fits is the best way to provide significant quantitative values for clinical diagnosis. In the spreadsheet, we observed linear fits would make the curve completed and spacing regular calculate T_(1/2) easily and high accuracy. Although the logarithmic and exponential fits made the curve completed, the slope of spacing was irregular and caused T_(1/2) value too high; while the power fits would make the curve so flat that T_(1/2) value would be too low. Conclusion: Renal function examination (Renography) is used to estimate kidney function, but we fall into pitfalls easily caused by the wrong value from curve fitting when we do the examination. This article reminds radiologists (or computer analysts) to choose right curve fitting to get reasonable quantitative values when they process the renography.

並列關鍵字

Curve Fitting ； Renography