肺臟斷層影像之電腦輔助分析

Computer-aided Quantification for The CT Image of Lung

趙一平(Yi-Ping Chao)

指導教授 ： 蘇振隆

中原大學/工學院/醫學工程研究所/碩士(2002年)

https://doi.org/10.6840/cycu200200464

摘要

高解析度的胸部電腦斷層(computed tomography, CT)影像對於胸部影像及解剖學提供了三維的了解，對於各種不同肺臟實質的病變，提供了臨床診斷的依據、處置疾病的方向與追蹤疾病的變化。本研究發展一肺臟CT影像肺功能參數擷取系統，能夠提供臨床上診斷決策所需之輔助參數。
系統透過擷取肺部DICOM影像直接進行分析，針對不同厚度影像採用不同閥值來選取出肺部區域，並進行半自動邊緣選取、去除雜訊、區域成長、Ray-casting、Voxel-highlight等影像處理函式來分析出肺活量、氣管壁厚度、氣管截面積、肺氣腫程度參數及氣管走向及角度。此外，也配合OpenGL API及Marching cube演算法表現出肺部Surface render三維影像，提供解剖學上之需要。研究中使用玻璃管假體及影像假體來測試系統參數的正確性，並對於正常者及病患的肺部CT影像實際進行分析探討。
假體測試評估中均證實參數之正確性後，研究便針對正常人及病患之CT影像進行分析。在實際肺活量參數方面，將程式計算值與傳統肺功能檢查結果比較所得R.R參數為0.910，其相關性不高的原因在於測量姿勢的不同及部分體績效應等因素所造成；對於正常人及支氣管擴張病患的”氣管截面積/氣管厚度”參數進行比較，經T-test統計之結果，其p值均小於0.05可知，利用此參數可以正確診斷出支氣管擴張及氣喘之病患；在BI參數研究上則是針對肺氣腫病患與正常人的1mm厚度影像進行分析，結果顯示計算BI參數時所呈現的class 3 及class 4部分對於肺氣腫的偵測相當重要，經過T-test統計方式所得P值均小於0.05，因此可以利用class 3及class 4的點數及比例判斷是否患有肺氣腫，並透過BI參數來對於肺氣腫程度進行量化。
本研究之完成能夠提供三維肺部影像暨肺功能參數分析系統，並能實際運用於肺部疾病的診斷上，透過開發之程式進行指標性參數擷取，並且配合三維成像概念來顯示出病灶處及空間關係。未來也可針對病患影像加以收集並擷取參數至資料庫內，透過資料探索及分析技術來輔助臨床上之肺部疾病診斷。

關鍵字

肺功能參數 ； 肺 ； 三維影像 ； 斷層掃瞄影像

並列摘要

In clinical, high-resolution chest computed tomography(CT) image can supports three-dimensional(3D) viewpoint and understanding in anatomy. According to the image texture, it could give the basics of clinical decision and the directions. Thus it could deal with pulmonary diseases and the changes of disease’s transfer. In this thesis, we developed a system which could acquire the pulmonary function parameters from chest spiral CT image sets and helped physician using those to make decisions in clinical diagnosis.
The system read serial chest CT DICOM image sets directly for analysis and used different threshold to process different thickness of the image slice that could separate the lung area from that. The semi-automatic image segmentation technology was used for this system were adopted masks to reduce the noise and image processing functions were used to analysis image and find the pulmonary function parameters. The image processing functions included Region-grow, Ray-casting, and Voxel-highlighting technology. The pulmonary function parameters included lung capacity, the thickness of airway tree, the lumen of airway tree, bullae index (BI), and the angle and direction of airway tree. Besides, this system could show 3D surface render image which composed of the OpenGL API and Marching cube algorithm, was needed in anatomy. Finally, this system could analysis phantom image, normal and patient image. The accuracy and the applicability of system parameter also were evaluated.
By the result, the accuracy of system could be evaluated by using the phantom image. Applying this system to analysis normal and patient image could get following results. Firstly, the R.R value in lung capacity was 0.910 between the traditional lung function test and system calculated, was low as a result of the different of the pose in measure and the effort of partial volume. Secondly, the p-value in t-test of the “airway lumen /airway thickness” parameter which compared by the normal and Chronic Obstructive Pulmonary Disease (COPD) patient was lower than 0.05. Finally, the BI index could present the class 3 and class 4 parts were important for detecting the degree of emphysema. Because the p-value was lower than 0.05, it showed number of point and percent of class 3 and class 4 could distinguish between normal and patient also provided the quantification in the degree of emphysema.
This research could provide a system to complete the 3D object reconstructed and pulmonary function parameters acquired which could be applied the clinical application to diagnose the lung disease. In the further, the system might combine with the database, and then using the data mining and other analysis technology assist the clinical applications.

並列關鍵字

Lung ； Three-dimension ； Computed Tomography ； Pulmonary function parameter

參考文獻

[1] N.L. Muller, R.R Miller, ”Diseaes of the bronchioles: CT and histopathologic findings.” Radiology 1995;196:3-12.

[2] W.R. Webb ,“Radiology of obstructive pulmonary disease”. AJR 1997;169:637-647.

[3] M.T. Wu, J.M. Chang, A.A. Chiang, ”Use of quantitative CT to predict postoperative lung function in patients with lung cancer.”,Radiology 1994;191:257-262.

[5] K.T. Bae, R.M. Slone, and D.S Gierada, ”Patients with emphysema: quantitative Ct analysis before and after lung volume reduction surgery.”, Radiology 1997;203:705-714.

[6] H.U. Kauczor, C.P. Heussel, and B. Fischer, ”Assessment of lung volumes using helical CT at inspiration and expiration: comparison with pulmonary function tests.”, AJR 1998:171:1091-1095.