中国医学影像技术
中國醫學影像技術
중국의학영상기술
CHINESE JOURNAL OF MEDICAL IMAGING TECHNOLOGY
2010年
4期
760-763
,共4页
马聪%张礼荣%尉传社%王冬青%王德航
馬聰%張禮榮%尉傳社%王鼕青%王德航
마총%장례영%위전사%왕동청%왕덕항
磁共振成像%膝关节%软骨%体积%OsiriX
磁共振成像%膝關節%軟骨%體積%OsiriX
자공진성상%슬관절%연골%체적%OsiriX
Magnetic resonance imaging%Knee joint%Cartilage%Volume%OsiriX
目的 评估在3.0T MR上使用半自动软件OsiriX测量膝关节软骨体积的可重复性及准确性. 方法 在3.0T MR上使用MR轴位水激发3D-FLASH序列对30名健康受试者的右膝关节进行重复扫描.全部图像由3名观察者分别使用开放源软件OsiriX进行软骨的半自动分割及随机工作站进行人工分割,计算软骨体积,比较两种方法测量软骨体积所需时间、可重复性及测量结果. 结果 ①OsiriX软件分割比人工分割节省50%以上时间;②OsiriX软件分割及人工分割的观察者间可重复性误差分别为4.88%和9.82%,高年资观察者内部可重复性误差分别为0.77%和1.29%,个体内部可重复性误差范围分别为0.14%~1.11%和0.52%~1.61%.前者各项可重复性误差均低于后者(P<0.05);③OsiriX软件分割的系统误差为(-3.80±3.93)%,随机配对误差为(4.68±2.70)%,差异无统计学意义(t=0.92,P=0.36). 结论 与人工分割相比,OsiriX半自动分割测量软骨体积具有省时、观察者间及观察者内可重复性高、有相对固定标准等明显优势,可用于临床及多中心大样本量研究.
目的 評估在3.0T MR上使用半自動軟件OsiriX測量膝關節軟骨體積的可重複性及準確性. 方法 在3.0T MR上使用MR軸位水激髮3D-FLASH序列對30名健康受試者的右膝關節進行重複掃描.全部圖像由3名觀察者分彆使用開放源軟件OsiriX進行軟骨的半自動分割及隨機工作站進行人工分割,計算軟骨體積,比較兩種方法測量軟骨體積所需時間、可重複性及測量結果. 結果 ①OsiriX軟件分割比人工分割節省50%以上時間;②OsiriX軟件分割及人工分割的觀察者間可重複性誤差分彆為4.88%和9.82%,高年資觀察者內部可重複性誤差分彆為0.77%和1.29%,箇體內部可重複性誤差範圍分彆為0.14%~1.11%和0.52%~1.61%.前者各項可重複性誤差均低于後者(P<0.05);③OsiriX軟件分割的繫統誤差為(-3.80±3.93)%,隨機配對誤差為(4.68±2.70)%,差異無統計學意義(t=0.92,P=0.36). 結論 與人工分割相比,OsiriX半自動分割測量軟骨體積具有省時、觀察者間及觀察者內可重複性高、有相對固定標準等明顯優勢,可用于臨床及多中心大樣本量研究.
목적 평고재3.0T MR상사용반자동연건OsiriX측량슬관절연골체적적가중복성급준학성. 방법 재3.0T MR상사용MR축위수격발3D-FLASH서렬대30명건강수시자적우슬관절진행중복소묘.전부도상유3명관찰자분별사용개방원연건OsiriX진행연골적반자동분할급수궤공작참진행인공분할,계산연골체적,비교량충방법측량연골체적소수시간、가중복성급측량결과. 결과 ①OsiriX연건분할비인공분할절성50%이상시간;②OsiriX연건분할급인공분할적관찰자간가중복성오차분별위4.88%화9.82%,고년자관찰자내부가중복성오차분별위0.77%화1.29%,개체내부가중복성오차범위분별위0.14%~1.11%화0.52%~1.61%.전자각항가중복성오차균저우후자(P<0.05);③OsiriX연건분할적계통오차위(-3.80±3.93)%,수궤배대오차위(4.68±2.70)%,차이무통계학의의(t=0.92,P=0.36). 결론 여인공분할상비,OsiriX반자동분할측량연골체적구유성시、관찰자간급관찰자내가중복성고、유상대고정표준등명현우세,가용우림상급다중심대양본량연구.
Objective To evaluate the repeatability and accuracy of semi-automated software OsiriX in the assessment of cartilage volume in normal knees on 3.0T MR. Methods Thirty right knees of healthy subjects were scanned twice with 3.0T MR using 3D-FLASH sequence with selective water excitation. Cartilage volume of the patellar compartment was determined with a validated open-source software OsiriX and manual segmentation separately by 3 observers using MRI data sets. After calculating the cartilage volumes, the segmentation processing times, the repeatability and volume results were compared between two segmentations. Results ①Compared with the manual segmentation, a time saving of at least 50% for cartilage volume measurement was achieved with OsiriX software segmentation. ②The interobserver repeatability error was 4.88% and 9.82% with OsiriX software segmentation and manual segmentation, respectively;the intraobserver repeatability error of high experienced observer was 0.77% and 1.29%, the interindividual repeatability range were 0.14%-1.11% and 0.52%-1.61%, respectively. The repeatability error of OsiriX software segmentaion was proved to be significantly smaller than manual segmentation (P<0.05). ③System difference between results obtained with OsiriX software measurement and manual measurement was (-3.80±3.93)% and absolute pairwise difference was (4.68±2.70)%. There was no statistical difference between the volume results measured using two techniques (t=0.92, P=0.36). Conclusion Compared to manual segmentation, cartilage volume measurement with OsiriX semi-automated segmentation is faster and has higher inter- and intraobserver repeatability with relatively fixed standard. This technique may therefore be used for clinical and multi-center trials of large sample.