胃癌CT体积测量在预测其术后病理分期中的应用

CT volumetry of gastric cancer can potentially predict the pathologic stage

目的:探讨胃癌患者CT肿瘤体积与术后病理分期的相关性及临床应用价值。方法:搜集2015年4月-2016年1月确诊的胃癌患者105例(男69例,女36例,年龄34~84岁,平均60.96±9.95岁)。术前1周内行CT增强扫描,通过人工测量门脉期每一层面的肿瘤面积乘以层厚叠加获得肿瘤体积,与术后病理分期进行关联性分析,将T分期分为T1-2组和T3-4组,N分期分为N0组和≥N1组后进行U-检验及ROC分析。结果:CT肿瘤体积与术后病理T分期及N分期的相关性分别为r=0.80(P<0.001)和r=0.66(P<0.001),呈显著相关,不同T或N分期的CT肿瘤体积中位数值随着T或N分期的增加呈增长趋势,各组间差异均具有统计学意义(P<0.001)。T3-4组或≥N1组的CT肿瘤体积分别明显>T1-2组或N0组(P<0.001)。CT肿瘤体积预测T1-2期的ROC曲线下面积AUC=0.96(95%CI0.90~0.99),95%置信区间的相伴概率P<0.001,如果将CT肿瘤体积≤24.5mL作为预测T1-2期的阈值时,其敏感度为92.9%,特异度为90.5%,准确度为91.4%。CT肿瘤体积预测N0期的ROC曲线下面积为0.84(95%CI0.76~0.90),95%置信区间的相伴概率P<0.001。如果将CT肿瘤体积≤23.4mL作为预测N0期的阈值时,其敏感度为75.6%,特异度为85.0%,准确度为81.0%。CT肿瘤体积预测T1-2N0期的ROC曲线下面积为0.80(95%CI0.65~0.91),95%置信区间的相伴概率P<0.001。如果将CT肿瘤体积≤10.8mL作为预测T1-2N0期的阈值时,其敏感度为71.9%,特异度为80.0%,准确度为73.8%。结论:CT肿瘤体积与术后病理T、N分期均具有显著相关性,通过选取恰当的阈值,可为胃癌的术前临床分期提供良好的参考价值。

Purpose: To evaluate the value of CT tumor volumetry of gastric cancer for predicting T and N stage with pathologic results as the reference standard. Methods: This study retrospectively evaluated 105 patients diagnosed with gastric cancer who underwent CT before surgery.CT tumor volumes were measured in portal venous phase.The correlations between CT tumor volumes and pathologic stages were analyzed.Mann-Whitney U -test was performed to assess differences between CT tumor volumes of different T-stages (T1-2 vs.T3-4) and N-stages (N0 vs.N≥1).Receiver-operating curve (ROC) characteristic analysis was further conducted to assess the diagnostic performance of significant parameters for prediction of T-stage and N-stage. Results: CT tumor volumes were significantly correlated with T stage and N stage (rho:0.80,0.66,respectively).The median value of CT tumor volumes in different T and N stage is a growing trend with the increase of T and N stage,which are statistically significant difference between the groups ( P <0.001).The differences between T1-2/N0 and T3-4/≥N1 stages in CT tumor volumes were statistically significant ( P <0.001).The area under the ROC curves (AUCs) for identification of T1-2 groups were 0.96,the CT tumor volumes of 24.5mL predicted T1-2 stage with 92.9% sensitivity,90.5% specificity and 91.4 accuracy.AUCs were 0.84 for identification of N0 groups.the CT tumor volumes of 23.4mL predicted N0 stage with 75.6% sensitivity,85.0% specificity and 81.0 accuracy.AUCs were 0.80 for identification of T1-2N0 groups.the CT tumor volumes of 10.8mL predicted T1-2N0 stage with 71.9% sensitivity, 80.0% specificity and 73.8 accuracy. Conclusion: CT tumor volumetry demonstrates potential value for predicting histopathologic T and N stage in patients with gastric cancer.