利用電腦輔助放射影像技術進行下顎骨生長之測量與型態學分析

Measurement and Morphological Analysis of Mandibular Growth Using Computer-Aided Radiological Approaches

林顯書(Hsien-shu Lin)

指導教授 ： 陳韻之

共同指導教授 ： 呂東武(Tung-Wu Lu)

國立臺灣大學/牙醫專業學院/臨床牙醫學研究所/博士(2014年)

https://doi.org/10.6342/NTU.2014.02192

摘要

背景：
下顎骨的生長模式在現代牙醫學基礎中扮演著十分重要之角色。過去學者常使用放射平面影像量測之方法記錄下顎骨的生長模式，但其透過平面觀點來描述立體的形態學參數，其中存在許多的盲點與干擾因素。隨著科技的進步，經由三維電腦斷層攝影(CT)的技術可記錄精確的下顎骨生長。但其昂貴的價格與高幅射劑量，使得相關的研究數量極少。近年來，改良過後的錐狀光束電腦斷層攝影(CBCT)解決了舊有CT的問題；因此，透過CBCT記錄下顎骨生長等相關之分析與研究便開始受到重視。而透過影像學探討下顎骨的生長分析，最重要的莫過於量測的可靠度，亦即辨識同一解剖標記點的重複性。其中，現行臨床應用中最廣泛使用的放射影像在拍攝時，存在著頭部擺位的問題。亦即在立體空間中，頭部解剖座標系統與攝影系統之座標系統於三軸任一方向有些微的旋轉偏差，皆可能影響影像的判讀與量測。因此根據二維影像在絕對標準的位置下與偏移的位置下，對於辨識解剖標記點的重複性有進一步探討的必要性。待量測的可靠度被釐清後，便可以在二維或三維的影像上做精確的標點，可提供型態學量測的精確性。
然而，鑑於醫學倫理的問題，長期對人類活體施予放射性的照射以取得下顎骨的影像資料尚不適當，因此，透過動物模型來了解相關研究則有其必要性。在廣泛的科學研究之中，迷你豬的下顎骨型態、咀嚼方式及骨頭代謝率皆與人類相似。因此本研究選用李宋系迷你豬做為下顎骨生長觀測的實驗動物。
綜上所述，本研究將使用電腦輔助放射影像技術進行下顎骨生長之測量與型態學分析。以進行下列的研究目的：(1)在絕對標準、隨機偏移角度及特定旋轉角度的下顎骨情況下，使用CBCT合成的測顱射影像來評估量測者之間、量測者之中與不同量測週期之間的人類下顎骨型態學參數之重複性。(2)使用CBCT長期量測迷你豬的下顎骨生長，以建立一個基本的資料庫，供未來進一步研究使用。(3)研究迷你豬下顎骨的連續性幾何變化。(4)比較下顎骨三維CBCT和CBCT合成的二維影像兩者之間的型態學量測參數誤差。

材料、方法：
本研究所使用的實驗儀器為牙科錐狀光束電腦斷層掃描系統。實驗第一部份使用12個因進行矯正評估所拍攝的三維影像，利用數位重組放射影像技術，重組為只有下顎骨塊的二維影像，在決對標準位置、隨機偏移位置與特定選轉角度位置的情況下，由一位資深與一位資淺的牙醫師，針對下顎骨上十一個標記點，進行間隔五天的兩輪點選。並將此資料做量測者之間、量測者之中與不同量測週期之間的人類下顎骨型態學參數之重複性。
第二部份選用八隻李宋系迷你豬，每隻豬於出生後四週起，開始持續每四週一次的牙科錐狀光束電腦斷層掃描，共拍攝12次，隨後利用軟體重建影像為三維立體模型，並選用17個標記點，做為型態學量測參數。根據標記點所相連成的線段與角度，量測迷你豬下顎骨的生長模式，並建立基本生長資料庫。
第三部份採用第二部份的三維立體模型，利用電腦程式自動分析，尋找每兩個量測週期之間骨表面上最相近的點，進行生長變化量的分析，以獲得一序列迷你豬下顎骨的幾何生長模型。
第四部份，利用上述三維立體模型，採用數位重組放射影像技術，取得每一個三維立體模型的二維模擬影像，並選用五個標記點兩兩連線，做為型態學參數，進行三維影像與二維影像之間，量測誤差的比較與分析。

結果：
研究結果顯示在標準位置的人類下顎骨影像上，資深與資淺的牙醫師，其量測者之間、量測者之中與不同量測週期之間的型態學參數之組內相關係數(intra-class correlation)均為非常好，只有資淺牙醫師的少數幾個參數的為好的。而在隨機旋轉的角度下，其結果近似於在標準位置下量測的結果。但是總體而言，組內相關係數稍差於在標準位置下量測的結果，顯示偏移的頭部擺放位置的確會對型態學參數的量測有所影響。
在長期迷你豬下顎骨生長型態記錄的研究中，其生長資料庫被成功的建立。並發現到下顎骨的生長在骨頭內部為非等向性且為非均質的，且隨著時間變化而顯示非線性的結果。在此12個月的觀察期中，下顎支生長較下顎體為快速。此下顎骨的生長模式顯示和齒列的發育有關，主要是指在上下方向的生長提供牙齒向上長以建立咬合關係，而在前後向的生長則是提供給後牙萌發的空間。
而在迷你豬下顎骨的幾何生長模型研究中，此為文獻中第一次提出迷你豬在生長期前十二個月，利用連續的色彩圖來呈現下顎骨表面的三維型態學變化。此為一個整合CBCT的全新方法來量化整體下顎骨及其不同區域非線性生長模式與非線性生長速率。
關於三維影像與二維影像之間，量測誤差的比較與分析研究中。使用二維影像量測方法，顯示明顯的誤差。主要是會低估下顎骨的尺寸以及生長早期的每月變化量及下顎骨的總生長量。這個結果告訴我們，若使用廣泛性的二維影像時，在牙科治療起始的治療計劃應該被仔細考量可能被低估的因素以獲得更精確的治療過程與結果。

關鍵字

牙科錐型電腦斷層 ； 顎顏面發育 ； 下顎骨生長 ； 迷你豬 ； 模擬的放射影像 ； 下顎骨 ； 信賴度 ； 敏感度分析 ； 測顱攝影片 ； 組內相關係數分析

並列摘要

Background:
Mandibular growth patterns play a very important role in the foundation of modern dentistry. Orthodontists and oral surgeons are especially interested in this topic. For many years, the earliest scholars used a measurement method based on two-dimensional radiographic images to record mandibular growth. However, this method could not provide a complete picture of three-dimensional bone mass because it was limited to only a two-dimensional plane view.

With advances in technology, three-dimensional computer tomography has been developed and used in the study of mandibular growth. However, little related research on mandibular growth has been conducted because of its high costs and high radiation doses. In recent years, improved cone beam computed tomography (CBCT) has been developed, eliminating the problems with computer tomography (CT). Therefore, many scholars have begun to use CBCT to record and analyze mandibular growth.

In using radiological imaging technology to investigate mandibular growth, the most important factor is the repeatability of the identification of anatomical marker points. In contrast, two-dimensional radiography entails a problem with head-positioning. In three-dimensional space, any slight deviation in the X, Y, or Z axis is bound to affect interpretation of the images.

Therefore, when using two-dimensional images in the absolute standard position and in the offset position, it is necessary to further investigate the repeatability of identification of anatomical marker points. Once this problem is clarified, it can be done accurately in two-dimensional or three-dimensional images to provide accuracy in morphological measurement.

Since medical ethics do not permit long-term radioactive irradiation of humans to obtain mandibular radiographic images, animal models must be used. Previous scientific research has determined that the miniature pig’s mandible patterns, chewing patterns, and bone metabolic rate are very similar to those of humans. Therefore, in this study, Lee Sung strain miniature pigs were used as the experimental animal for the observation of mandibular growth in this study.

This study used computer-assisted radiological imaging techniques to measure mandibular growth and conduct morphological analysis for the following research purposes:
(1) to assess the repeatability of the morphological parameters of human mandibles between measurements and those measurements among different measurement cycles under various mandibular conditions of the absolute standard, random offset angle and certain angle of rotation using CBCT.
(2) to use CBCT to collect measurements of mandibular growth in miniature pigs over the long term and thereby create a basic database for use in further study,
(3) to study the continuity of the geometric changes of the mini pig’s mandibles, and
(4) to compare the measurement error of morphological parameters between
three-dimensional CBCT and the synthetic two-dimensional CBCT images of mandibles.

Materials and Methods:
The experimental apparatus used in this study was a dental cone beam computed tomography (CBCT) system. In the first part of the study, CBCT data of twelve mandibles were obtained and used to generate synthetic cephalograms via a digitally reconstructed radiography (DRR) technique. Eleven landmarks describing the key morphological features of the mandible on each DRR-synthesized cephalogram were identified manually 6 times by one senior and one junior dentist in the neutral and randomized rotated positions. The procedure was repeated 5 days later. Twelve parameters based on inter-landmark line segments and their angles were calculated. Test-retest reliability was assessed in terms of intra-class correlation coefficient (ICC) and coefficient of variation (CV) using a two-way mixed-effects model. The paired-sample t-test was used to compare differences between examiners and sessions. A one-sample t-test was employed to assess whether differences between the examiners were significantly different from zero.

Eight Lee-Sung strain miniature pigs were chosen in the second part of the study. The mandibles of each of the pigs were scanned with CBCT 12 times, once every 4 weeks beginning at the age of four weeks, to accumulate a total of twelve sets of CT data. Each of the CT data sets was used to reconstruct a 3D model of the mandible. In total, 17 anatomical landmarks on the mandible were marked as morphological measurement parameters by an experienced dentist. Line segments were then defined for selected pairs of markers and were used to measure the growth patterns of the miniature pigs' mandibles and build up baseline data of mandibular growth.

The third part was adapted from the three-dimensional model of the second part. A
new method was developed to search for corresponding points on two consecutive
models with the highest likelihood of the anatomical and morphological features.
The growth rate patterns of the mandibles for each month were described using color
maps on the models over the monitoring period.

In the fourth part of the study, mandibles of six miniature pigs were scanned with CBCT at one-month intervals for 12 months, and the data were used to reconstruct 3D bone models. The DRR technique was used to generate simulated 2D cephalograms. Five anatomical landmarks were identified on each bone model, and the inter-marker distances, monthly distance changes, and their errors were calculated with the 3D measurements as the gold standard. Differences in the variables measured using the 2D and 3D methods were compared using a paired t-test.

Results:
The results indicated very good intra-rater (senior: ICC>0.93; junior: ICC=0.78 for CdP-GOP, ICC>0.91 for other parameters), inter-rater (ICC=0.62 for CdP-GOP, ICC>0.84 for other parameters), and inter-session reliability (ICC>0.84 for all parameters and examiners; ICC=0.74 for CdP-GOP for junior examiner) in measuring mandibular morphological parameters in the neutral position. These results suggest that very good reliability can be achieved via manual identification of the anatomical landmarks, without the effects of factors such as malpositioning of the head during imaging. The results for reliability of the randomized rotated position were not quite as high, especially in the trials measured by the junior dentist. This difference can be explained by the malpositioning of the head during imaging, which indeed affected identification of anatomical landmarks and thus the accuracy of morphological measurements.

In the study of long-term miniature pigs’ mandibular growth patterns, the baseline data of the mandibular growth has been successfully established. The mandibular volume increased nonlinearly with time, growing rapidly during the first five months and more slowly in subsequent months. The growth of the mandible was found to be anisotropic and non-homogeneous within the bone and non-linear over time, with faster growth in the ramus than in the body. These growth patterns appeared to be related to the development of dentition, providing necessary space for the teeth to grow upward for occlusion and for the posterior teeth to erupt.

This is the first study to track the continuous morphological changes of the mandible in miniature pigs during growth in the first 12 months in three dimensions with continuous color maps over the surface of the mandible. This was achieved by integrating CBCT and the new analytical approach, which quantifies the nonlinear growth patterns and the nonlinear rate of their changes in different growth regions and the whole mandible.

Comparison and analysis of measurement errors between the three-dimensional images and two-dimensional images revealed significant errors in measurements using 2D imaging. The mandibular dimensions, their monthly changes in the early stages of growth, and total annual growth were underestimated. These results should be considered in dental treatment planning at the beginning of treatment in order to control more precisely the treatment process and outcome.

並列關鍵字

Cone-Beam Computed Tomography ； CBCT ； Maxillofacial Development ； Mandible Growth ； Swine ； Miniature Pig ； Simulated radiograph ； mandible ； reliability ； sensitivity analysis ； ICC ； cephalograms

參考文獻

Kumar, V. and K. Satheesh (2013). "Applications of cone beam computed tomography (CBCT) in implant treatment planning." JSM 1(1008).

(2007). "Recommendations of the International Commission on Radiological Protection ICRP Publication 103." Ann ICRP: 37.

(Ed.), A. A. (1997). "Orthodontic Cephalometry." Mosby Wolfe,London, .

AD, W. (2008). "New ICRP recommendations." Journal of Radiological Protection 28: 161.

Adams, G. L., S. A. Gansky, A. J. Miller, W. E. Harrell and D. C. Hatcher (2004). "Comparison between traditional 2-dimensional cephalometry and a 3-dimensional approach on human dry skulls." American journal of orthodontics and dentofacial orthopedics : official publication of the American Association of Orthodontists, its constituent societies, and the American Board of Orthodontics 126(4): 397-409.