錐狀射束電腦斷層掃描與多切片電腦斷層掃描對於顱顏軟硬組織測量之比較

Abstract

目的 近年來，有越來越多學者使用錐狀射束電腦斷層掃描(cone-beam computed tomography, CBCT)進行各種研究。然而與多切片電腦斷層掃描(multislice computed tomography, MSCT)相比，CBCT對於偵測軟組織間的差異有其侷限且影像雜訊較多。此外，CBCT拍攝整個頭部需要較長的時間，可能因病患呼吸產生假影(artifacts)。儘管如此，回顧文獻發現單單2021年就有約百篇的研究是與CBCT測量呼吸道大小有關。有鑒於目前並沒有針對MSCT與CBCT在矯正領域中軟硬組織測量值差異以及可信度的的比較，因此本研究旨在針對MSCT和CBCT在矯正領域中軟硬組織測量值差異和可信度進行比較，以提供臨床實踐和決策支持。 研究程序及方法 本研究為一個回溯性研究，透過台大醫院的醫療整合資料庫，檢索2016-2021年期間，接受過兩次頭頸部電腦斷層掃描的病患，並且通過院內系統查證他們在此期間內沒有接受過任何可能導致顱顏結構明顯變化的治療。並將拍攝一次CBCT和一次MSCT的病患歸於BOTH組；拍攝兩次CBCT的病患歸於CBCT組；拍攝兩次MSCT的病患歸於MSCT組以及另外使用一顆頭顱骨分別拍攝一次CBCT與MSCT稱為SKULL組。使用Dolphin軟體和Amira軟體對CBCT和MSCT影像對四大組進行分析，並請6位觀察者測量BOTH組和SKULL組，而CBCT組與MSCT組找1位觀察者進行測量，測量包含矯正領域常用的軟硬組織變項，接著比較測量值的觀察者內(intra-rater)和觀察者間(inter-rater)的差異以及信度(reliability)等。 結果 大多數變項之觀察者間組內相關係數(Intraclass correlation coefficient, ICC)與觀察者內ICC為「幾乎完美一致」級別(0.81-0.99)，只有少數一些變項為「非常一致」級別(0.61-0.80)，包含：ANW(anterior nasal width)、PNW(posterior nasal width)、FM r-l(frontomaxillary suture right to left)、INC r-l(Inner nasal contour point right to left)以及MCA(minimum cross-sectional area in oropharynx)等。整體而言，這些結果顯示出ICC之一致性相當不錯，對於此研究的信度具有一定的支持。但從敘述性統計與檢定觀察，同種成像方式之同一受試者其測量變項，在測量軟體上有著顯著差異，尤其是軟組織類。另外也發現SKULL組全部變項之標準差都於1 個度量衡單位之內(如：1 mm、1度等)，而BOTH組中不論成像方式為CBCT或MSCT，當測量軟體為Dolphin時，其三大類變項，如：ANW、ANFW(anterior nasal floor width)、PNW、PNFW(posterior nasal floor width)、Inter-Mpa(intermolar width at the first molar palatal apex level)、Inter-Mcf(intermolar width at the first molar central fossa level)、ZMU r-l(zygomaticomaxillary upper right to left)、ZML r-l(zygomaticomaxillary lower right to left)、Co-A(condylion to point A)以及Co-Gn(condylion to gnathion)等之標準差在臨床測量上可能會超出1個度量衡單位；當測量軟體為Amira時，其三大類變項，如：PNFW、ZTU r-l(zygomaticotemporal suture upper right to left)、ZMU r-l以及ZML r-l等之標準差在臨床測量上可能會超出1個度量衡單位。 而針對MSCT組與CBCT組而言，使用Amira軟體時，Inter-Mcf和TOH在MSCT重拍之差異與CBCT重拍之差異有顯著差異；使用Dolphin軟體時，Inter-Mcf、INC r-l和OAV在MSCT重拍之差異與CBCT重拍之差異有顯著差異；進行MSCT拍攝時，ANW、ANFW、PNW、PNFW、EMW、ZTU r-l、FZ r-l、INC r-l、ZMU r-l、SNA、SNB以及SNGoGn在Amira重測之差異與Dolphin重測之差異有顯著差異；進行CBCT拍攝時，ANW、PNW、Inter-Mpa、FZ r-l、Co-Gn以及OAV在Amira重測之差異與Dolphin重測之差異有顯著差異。 結論 儘管統計上ICC有著相當不錯的一致性，但於臨床上不論哪項變項，其測量之誤差還是經常超出一個度量衡單位的黃金標準，本研究建議在進行齒顎矯正治療或是口腔顎面外科手術前之術前評估，成像方式上MSCT與CBCT都可以提供有用的資訊，測量軟體選擇上可使用Amira軟體，以獲得更加準確和可靠的測量結果。

Objective In recent years, cone-beam computed tomography (CBCT) has been used by many scholars for various research purposes. However, compared with multi-slice computed tomography (MSCT), CBCT has more image noise, which limits its ability to detect differences between soft tissues. In addition, CBCT takes longer to capture images of the entire head and may produce artifacts due to patient motion. Despite this, reviews of the literature show that about 100 studies in 2021 alone were related to CBCT measurements of airway size. Given the lack of comparison between MSCT and CBCT in terms of differences and reliability of soft and hard tissue measurements in the orthodontic field, this study aims to provide clinical practice and decision support by comparing the differences and reliability of these measurements between MSCT and CBCT. Material and methods This study is a retrospective study that involved retrieving patients from the integrated medical database of National Taiwan University Hospital who underwent two head and neck computed tomography (CT) scans between 2016 and 2021. It was verified through the hospital system that these patients did not receive any treatments during this period that could have led to significant changes in craniofacial structures. The patients who underwent one CBCT scan and one MSCT scan were categorized into the BOTH group. Those who had two CBCT scans were categorized into the CBCT group, while those who had two MSCT scans were categorized into the MSCT group. A A skull was used and undergone one CBCT scan and one MSCT scan, then categorized into the SKULL group. The CBCT and MSCT images were analyzed using Dolphin software and Amira software for the four groups. Six raters were assigned to measure the BOTH and SKULL groups, while one rater was assigned to measure the CBCT and MSCT groups. The measurements included commonly used soft and hard tissue variables in orthodontics. The differences and reliability (intra-rater and inter-rater) of the measurements were then compared. Results Most of the inter-examiner and intra-examiner’s intraclass correlation coefficient, (ICC) for the variables were in the "almost perfect agreement" category (0.81-0.99), with only a few variables falling into the "substantial agreement" category (0.61-0.80), including ANW(anterior nasal width), PNW(posterior nasal width), FM r-l(frontomaxillary suture right to left), INC r-l(Inner nasal contour point right to left) and MCA(minimum cross-sectional area in oropharynx). Overall, these results showed good consistency of the ICCs and provided some support for the reliability of this study. However, from descriptive statistics and tests, it was observed that there were significant differences in measurement variables within the same subject for the same imaging modality, particularly in the category of soft tissue, across different measurement software. It was also observed that the standard deviations of all variables in the SKULL group were within 1 unit of measurement (e.g., 1 mm, 1 degree). However, in the BOTH group, regardless of the imaging modality being CBCT or MSCT, when using the Dolphin software, the standard deviations of the three major categories of variables, such as ANW, ANFW(anterior nasal floor width), PNW, PNFW(posterior nasal floor width), Inter-Mpa(intermolar width at the first molar palatal apex level), Inter-Mcf(intermolar width at the first molar central fossa level), ZMU r-l(zygomaticomaxillary upper right to left), ZML r-l(zygomaticomaxillary lower right to left), Co-A(condylion to point A) and Co-Gn(condylion to gnathion) might exceed one unit of measurement in clinical measurements. When using the Amira software, the standard deviations of the three major categories of variables, such as PNFW, ZTU r-l(zygomaticotemporal suture upper right to left), ZMU r-l, and ZML r-l might exceed one unit of measurement in clinical measurements. As for MSCT group and CBCT group, when using Amira software, the Inter-Mcf and TOH showed a significant difference between MSCT and CBCT reshoots. When using Dolphin software, the Inter-Mcf, INC r-l, and OAV showed a significant difference between MSCT and CBCT reshoots. When performing MSCT scans, the ANW, ANFW, PNW, PNFW, EMW, ZTU r-l, FZ r-l, INC r-l, ZMU r-l, SNA, SNB, and SNGoGn showed a significant difference between Amira and Dolphin repeated measurements. When performing CBCT scans, the ANW, PNW, Inter-Mpa, FZ r-l, Co-Gn, and OAV showed a significant difference between Amira and Dolphin repeated measurements. Conclusion Although there is fairly good consistency in terms of ICC in statistical analysis, the measurement errors for all variables in clinical practice often exceed one unit of the gold standard measurement. This study suggests that in preoperative evaluations for orthodontic treatment or oral maxillofacial surgery, both MSCT and CBCT imaging can provide valuable information. In terms of measurement software selection, using Amira software is recommended to obtain more accurate and reliable measurement results.