以多切面電腦斷層低劑量肺部影像推估冠狀動脈鈣化分數之研究

The Study of Deriving Coronary Calcium Score from Low-dose Multi-slice Computed Tomography for Lung Scan

廖育琳(Yu-Ling Liao)

指導教授 ： 蘇振隆

中原大學/工學院/生物醫學工程研究所/碩士(2015年)

https://doi.org/10.6840/cycu201500938

摘要

低劑量肺部電腦斷層掃描是目前用來篩檢肺癌的最佳工具，低劑量的應用主要是需保持檢測胸腔內結節敏感度，並合理抑低對受檢者的輻射劑量。在低劑量肺部電腦斷層檢查的影像裡也同時能夠發現冠狀動脈鈣化的現像。但如果只以肉眼評判，無法評估且量化冠狀動脈鈣化的程度。本文之目的為探討利用低劑量肺部電腦斷層癌症篩檢影像獲取冠狀動脈鈣化指數之可行性及修正模式。
採回溯性研究設計，以530人為研究樣本，於台灣北部健檢中心同時接受低劑量肺部電腦斷層以及標準冠狀動脈鈣化指數的受檢者為對象，分二階段進行，第一階段將原始的低劑量肺部電腦斷層影像改變參數重新組成新影像，與標準的冠狀動脈鈣化指數比較其可靠性。第二階段將標準鈣化的Threshold 130HU臨界值更改成90HU，並計算出鈣化分數，再與標準鈣化分析法比較其數值與鈣化分級，探討其一致性。
以第一階段研究結果顯示，年齡、性別和身體質量數這三個風險因子均為冠狀動脈鈣化嚴重度的有效指標，比較二種分析法測量鈣化分級，其一致性係數(Kappa)為0.74，在無鈣化分級的樣本群組，一致性為0.89，且在本研究樣本中無鈣化分級的樣本比例佔了78.6%，這表示有高比例的人口皆為無鈣化群組，進一步分析兩個方法的Pearson與Spearman相關係數分別為0.979與0.915，且利用低劑量分析法所得到的平均吸收劑量明顯比標準鈣化分析法所得到的吸收劑量減低1/3。依據第二階段實驗組研究結果發現，一致性係數(Kappa)為0.83，Pearson與Spearman相關係數分別為0.986與0.942，雖然一致性結果未達百分百吻合，但也呈現極高度吻合，且一致性與相關性係數，皆高於第一階段研究的結果。因此，以第二階段研究的結果可推論更改Agatston calcium score的Threshold為90，可增加利用低劑量肺部電腦斷層影像推估冠狀動脈鈣化指數的準確性。
綜合以上研究結果顯示，利用低劑量肺部影像推估冠狀動脈鈣化，透過修正模式是可適合運用在罹患冠狀動脈疾病率較低的人口或一般無冠狀動脈疾病症狀也不具有冠狀動脈風險因子的健康受檢者身上。當發現有冠狀動脈鈣化時，再做進一步檢查，這對於檢查成本、檢查時間、輻射劑量都可相對的減少。

關鍵字

低劑量肺部電腦斷層、冠狀動脈鈣化、標準鈣化指數

並列摘要

Low dose lung computed tomography (CT) scan is ideal for lung cancer screening. Maintaining sensitivity of detecting lung nodule and reasonably reducing radiation dose is the rationale behind low dose lung CT scan. However, coronary calcification, sometimes evidently identified, cannot be accurately quantified in the setting of low dose lung CT scan. The purpose of this study was to explore the feasibility of deriving coronary calcium score from low dose lung CT scan and to establish a post-processing correction model accordingly.
This was a retrospective study conducted in a Northern healthcare institute with a sample size of 530 persons who simultaneous underwent low dose lung CT scan and standard CT scan for estimating Agatston coronary calcium score. For first part of the study, parameters in low dose lung CT scan image reconstruction were modified and the reconstructed new images were used for calculating coronary calcium score and calcium score ranking. Calcium score and calcium score ranking derived from low dose lung CT scan and standard CT scan were compared with its agreement reported. For second part of the study, the threshold CT value of coronary calcification was lowered from 130 to 90 Hounsfield Unit (HU) and calcium score was recalculated. The recalculated calcium score and calcium score ranking were compared with those by standard CT and its agreement was explored.
In first part of the study, age, gender and body mass index significantly correlated with coronary calcium score. Calcium score ranking derived from low dose CT scan and standard CT scan showed modest agreement - Kappa value 0.74 for overall and 0.89 for zero calcium score subgroup (comprising 78.6% of the overall sample size). Calcium score and its ranking showed good correlation between two CT scans with Pearson and Spearman correlation coefficient of 0.979 and 0.915, respectively. The average radiation absorbed dose in low dose lung CT scan reduced one third compared with standard CT scan. In second part of the study, the agreement and correlation of calcium score and calcium score ranking in low dose lung CT and standard CT were further improved with a Kappa value of 0.83, Pearson and Spearman correlation coefficient of 0.986 and 0.942, respectively. It is postulated that adjusting threshold CT value for coronary calcification to 90 HU may improve accuracy of calcium score and calcium score ranking in low dose lung CT scan.
Summing up, low dose lung CT scan combined with the proposed correction model may be feasible for estimating coronary calcium score in low-risk or asymptomatic subjects without risk factors for coronary artery disease. Reserving further work-ups such as standard CT scan for those with significant coronary calcification seen on low dose lung CT not only reduces radiation dose but also may be more time-saving and cost-effective.

並列關鍵字

Low dose lung computed tomography, coronary calcification, coronary calcium score

參考文獻

[4] 曾揚達，羅偉業，陳良光，蘇誠道，冠狀動脈鈣化評分之臨床應用與分析，中華放射線技術學雜誌2005，29(1): 65-71。

[13] 卓瓊珍，陳建雄，梁羅沙，姚惠玲，吳銘庭，16 多切層螺旋電腦斷層低劑量肺癌篩檢影像評估冠狀動脈鈣化程度，中華放射線技術學雜誌2009，33(2): 99-104。

[25] 蔡書汎，田雨生，洪儀倫，廖英蘭，蔡惠予，電腦斷層掃描之自動管電流調控技術的劑量與影像品質，中華放射線技術學雜誌2010，34(2): 81-89。

[42] 郭嬿姝，六十四多切面電腦斷層冠狀動脈血管攝影與發生冠狀動脈疾病之相關性探討，國立台灣大學，2009年，碩士論文。

[45] 廖學鍵，劉益瑞，陳詩華，彭曉微，吳新華，利用動態假體研究冠狀動脈電腦斷層之移動性假影，中華放射線技術學雜誌2010，34(1): 42-52。