摘要
目的探讨通过多层螺旋CT轴位图像预测游离气胸肺压缩百分比的准确性。资料与方法回顾性分析88例游离气胸CT扫描数据,分左肺及右肺,用VR容积测量法测量气胸量及胸腔体积并计算气胸肺压缩百分比Y;在轴位图像气管隆突层面测量胸廓最大垂直距离D_1及该线上气胸脏层胸膜至胸廓的垂直距离d_1,并计算垂直距离比值X_1;同理测量胸廓水平距离D_2及气胸边缘至胸廓的水平距离d_2,并计算水平距离比值X_2;将Y与X_1及X_2进行回归分析,挑选出系数显著且解释度最大的回归方程,并验证方程准确性。结果所有模型差异均有统计学意义(P<0.001)。左侧气胸最佳方程:Y=0.0758+2.8623X_1—2.8049X_12,R^2=0.7460;右侧气胸最佳方程:Y=0.0708+2.5960X_1—2.1178X_12,R^2=0.8575。方程预测左侧气胸肺压缩百分比误差范围为(0.01±9.68)%,右侧为(0.12±8.02)%。结论通过气管隆突层面轴位图像测量垂直距离压缩比X_1及水平距离压缩比X_2均对预测游离气胸肺压缩百分比有意义,尤其是X_1更能反映气胸的压缩百分比。
Purpose To evaluate the accuracy of predicting lung compression percentage in free pneumothorax using axial images of multi-slice spiral CT. Materials and Methods The chest CT data of 88 cases with free pneumothorax were retrospectively analyzed. Volume rendering measurement was used to calculate the volume of pneumothorax and the thoracic cavity in the left and right lungs to calculate compression percentage Y. At the level of carina on axial view, the maximum vertical distance D_1 and the distance from visceral layer of compressed lung to the parietal pleural d_1 were measured to calculate the vertical distance ratio X_1; similarly, a maximum horizontal distance D_2 and the horizontal distance d_2 from visceral layer to the parietal pleural were measured to calculate the horizontal distance ratio X_2. Regression analysis was performed to select the equation with significant coefficient and fitness, and the accuracy of the equation was verified. Results All regression models were statistically significant（P0.001）. The best equation for the left pneumothorax was： Y=0.0758＋2.8623X_1- 2.8049X_12, R^2=0.7460; the best equation for the right pneumothorax was： Y=0.0708＋2.5960X_1- 2.1178X_12, R^2=0.8575. The deviation of predicting the percentage of pneumothorax of left pneumothorax was（0.01±9.68）%, and the right was（0.12±8.02）%. Conclusion Measurement of the vertical compression ratio X_1 and the horizontal compression ratio X_2 at the carina plane can predict lung compression percentage in free pneumothorax, with the vertical compression ratio X_1 more accurately reflecting compression percentage.
出处
《中国医学影像学杂志》
CSCD
北大核心
2016年第10期750-754,共5页
Chinese Journal of Medical Imaging
