以雙能量電腦斷層血管攝影自動去骨技術量化頸動脈狹窄度及斑塊形態

Automatic Subtract of Bone and Calcified Plaques using Dual Energy CT for Carotid Artery Angiography to Quantify the Stenosis Grade and Morphology

林郁珮(Yu-Pei Lin) ； 陳威良(Wei-Liang Chen) ； 王景祥(Ching-Hsiang Wang) ； 趙敏(Min Chao)

《台灣應用輻射與同位素雜誌》 10卷3期 (2014/09) Pp. 901-910

摘要

MDCTA對於頸動脈狹窄的檢出率頗高，但礙於空間解析度限制，對於顯著性狹窄（≥50%）的頸動脈伴隨硬化斑塊時，在影像判讀上鈣化所致部分容積效應會影響密度的測量，限制了鑑別斑塊的可靠性進而影響血管狹窄管腔的測量。故本研究旨在評估DECT的3D-MIP_(PBS)技術在診斷頸動脈狹窄及斑塊形態的臨床價值。採回溯性方式收集超音波篩檢頸動脈狹窄度呈陽性之40位受檢者，以3D-MIP_(PBS)影像、NASCET規範判定頸動脈狹窄度、及狹窄處之表面斑塊形態，並以頸動脈超音波檢查報告為準則。共取得80條頸動脈分叉處之血管片段，經專家模式評分其3D-MIP_(PBS)平均狹窄率分別為75.2%±19.1、75.1%±19.3，Kappa為0.891。有7段血管為假性閉鎖，決定係數(R^2)為0.716，截距為11.168，所以會高估狹窄率。另以ROC曲線判別各後處理軟體在頸動脈狹窄率≥50%、≥70%的敏感度及特異性，其AVA、sagMPR、MIP_(PBS)的AUC分別為0.98、0.96、0.93 和0.93、0.91、0.88，以AVA的鑑別力最好。而血管表面鈣化斑塊與管徑狹窄度並無直接關係（Spearman相關係數-0.101，P＞0.05），但將鈣化結構移除後，造成中度以上狹窄多為潰瘍形斑塊（48%），次為不規則形斑塊（33%），而平滑斑塊僅（19%）。故DECT3D-MIP_(PBS)演算法雖會高估血管直徑，但卻可克服鈣化性斑塊所致的部分容積效應，增加鑑別表面斑塊形態的可靠性。

關鍵字

雙能量電腦斷層血管攝影 ； 頸動脈狹窄 ； 三維最大強度投影

並列摘要

MDCTA has high detection rate for carotid stenosis, but the spatial resolution limit for significant carotid stenosis (≥50%) with plaques is limited. The partial volume effects caused by calcification on image interpretation will affect the measured density, limiting the identification plaque, and impact the measurement of vascular lumen. This study aims to assess the DECT 3D-MIP_(PBS) clinical value in the diagnosis of carotid artery stenosis technology and plaque morphology. The method to identify retrospectively 40 patients with carotid artery CDS is positive. Stenosis was quantified according to NASCET criteria on ultrasound report and surface plaques morphology images by two experienced radiologists for each modality on the same planes. To quantify inter-reader reliability, linear weighted Cohen's kappa was calculated. To avoid overestimation of reliability due to the high number of categories linear weights were chosen over quadratic weights. And to assess 3D-MIP_(PBS) images value of diagnostic, the ROC curve was calculated. Achieved a total of 80 carotid artery, the 3D-MIP_(PBS) images rated by the expert mode ,the average stenosis rate was 75.2%±19.1, 75.1%±19.3, kappa value is 0.891. There are seven sections of carotid artery was pseudo atresia, the R^2 of 0.716, an intercept of 11.168, it will overestimate stenosis rate. ROC curve with each other post-processing software to determine the rate of carotid artery stenosis ≥ 50%, ≥ 70%. The sensitivity and specificity of the AVA, sagMPR, 3D-MIP_(PBS) the AUC were 0.98, 0.96, 0.93 and 0.93, 0.91, 0.88, the AVA discernment best. Surface calcified plaque and diameter stenosis were no statistical relationship (Spearman correlation coefficient -0.101, P＞0.05). But the calcified structures removed, moderate stenosis the mostly ulcerated plaque (48%), followed by irregular patches (33%), smooth plaque was only 19%. PBS facilitated the evaluation of grade of the stenosis in all cases. Nevertheless, after PBS stenosis were overrated in 3D-MIP_(PBS) in comparison to ultrasound and axMPR. Moreover, plaque morphology, as an independent risk factor for stroke, can be evaluated even in calcified plaques after PBS. Therefore dual energy CTA with plaque subtraction has the potential to identify patients with vulnerable plaques better than conventional CTA.

並列關鍵字

Dual-Energy CT Angiography ； carotid artery stenosis ； Three-Dimensional Maximum Intensity Projection ； Subtraction of Bone and Calcified Plaques