摘要
目的:根据40s动脉晚期(LAP)的CT检查结果,使用列线图术前预测胃癌分化程度。方法:回顾性收集本院188例胃癌患者的病例资料,将其分为训练组(85例低分化和41例中分化/高分化)和验证组(42例低分化和20例中分化/高分化)。分析经手术切除的胃癌患者术前CT图像,评估40s LAP的12个形态学特征。建立基于多参数二元逻辑回归模型的列线图来预测低分化胃癌并用ROC曲线评价诊断效能。结果:训练组40s LAP的6个形态学特征、6个常规CT值参数和年龄在两组间差异均有统计学意义(均P<0.05)。多参数模型由年龄(P=0.003)、浸润性(P=0.001)、形态(P<0.001)、"C"征(P=0.009)、延迟期平均CT值(DP value mean,P=0.005)和延迟期最小CT值(DP value min,P=0.046)组成。训练组和验证组基于多参数模型预测低分化胃癌的AUC值分别达到0.849和0.762。结论:40s动脉晚期CT图像的多个形态学特征及CT值参数在低分化和中/高分化胃癌组间存在显著差异。此外,联合形态学特征、常规CT值参数和年龄的列线图可术前预测胃癌分化程度。
Objective:To predict the differentiation of gastric cancer(GC) preoperatively using a nomogram based on CT findings in 40 s late arterial phase(LAP).Methods:A total of 188 patients with GC in our hospital were retrospectively collected and divided into training cohort(85 poorly and 41 moderately/well differentiated GCs) and validation cohort(42 poorly and 20 moderately/well differentiated GCs).Twelve CT morphological characteristics in 40 s LAP were evaluated.A nomogram based on the multivariate binomial logistic regression model was built to predict poorly differentiated GCs.ROC curve was performed to evaluate the diagnostic efficacy.Results:In the training cohort, six morphological characteristics in 40 s LAP,six traditional CT value parameters, and age differed significantly between the two groups(all P<0.05).The multivariate model consists of age(P=0.003),infiltrative(P=0.001),morphology(P<0.001), "C" sign(P=0.009),DP value mean(P=0.005) and DP value min(P=0.046).Based on the multivariate model, the AUCs of the training and the validation cohort for predicting poorly differentiated GCs were 0.849 and 0.762,respectively.Conclusion:Multiple morphological characteristics and CT value parameters in 40 s LAP were significantly different between poorly and moderately/well differentiated GCs.Furthermore, a nomogram integrating morphological characteristics, traditional CT value parameters, and age might hold promise in predicting the differentiation of GC preoperatively.
作者
王玉婷
谭玲玉
乔香梅
季长风
李琳
刘松
WANG Yu-ting;TAN Ling-yu;QIAO Xiang-mei(Department of Radiology,Nanjing Drum Tower Hospital,the Affiliated Hospital of Nanjing University Medical School,Nanjing 210008,China)
出处
《放射学实践》
CSCD
北大核心
2022年第3期338-343,共6页
Radiologic Practice
关键词
胃肿瘤
分化程度
体层摄影术
X线计算机
列线图
Stomach neoplasms
Differentiation degree
Tomography,X-ray computed
Nomogram
