基于能谱CT多参数术前预测结直肠癌神经侵犯  

Spectral CT multi-parameters for preoperatively predicting colorectal cancer perineural invasion

作　　者：卢婷 李昇霖 张玉婷[1,2,3,4] 袁隆 周俊林 LU Ting;LI Sheng-lin;ZHANG Yu-ting(Department of Radiology,Lanzhou University Second Hospital;Second Clinical School of Lanzhou University;Key Laboratory of Medical Imaging of Gansu Province;Gansu International Scientific and Technological Cooperation Base of Medical Imaging Artificial Intelligence,Lanzhou 730030,China)

机构地区：[1]兰州大学第二医院放射科,甘肃兰州730030 [2]兰州大学第二临床医学院 [3]甘肃省医学影像重点实验室 [4]医学影像人工智能甘肃省国际科技合作基地

出　　处：《放射学实践》2024年第12期1636-1642,共7页Radiologic Practice

基　　金：兰州大学第二医院萃英科技创新计划基金(CY2021-ZD-01)。

摘　　要：目的:探讨能谱CT定量参数术前预测结直肠癌(CRC)神经侵犯(PNI)的价值。方法:回顾性分析2021年3月-2023年5月经术后病理证实的123例CRC患者的临床和影像资料。其中,存在PNI者71例,无PNI(nPNI)者52例。术前所有患者行能谱CT全腹三期增强扫描。在各期图像上测量病灶的碘基值(IC)、标准化IC(nIC)与能谱曲线斜率(λH)并计算肿瘤细胞外容积分数(ECV)。比较PNI组与nPNI组之间各项能谱CT参数值的差异。将差异具有统计学意义的参数纳入二元logistic回归分析中,构建联合预测模型,并绘制受试者工作特征(ROC)曲线评估各项参数及联合模型对PNI和nPNI的鉴别诊断效能。结果:PNI组病灶在三期增强图像上的IC和λH、以及在静脉期和延迟期图像上的nIC和ECV的测量值均大于nPNI组,差异均具有统计学意义(P<0.05);而动脉期nIC在两组间差异无统计学意义(P>0.05)。单一参数预测PNI状态的AUC为0.632~0.820;联合模型预测PNI状态的AUC为0.859,敏感度和特异度分别为76.10%和86.50%。结论:能谱CT多参数能够在术前预测CRC-PNI,有望推进能谱CT在术前预测CRC-PNI中的进一步研究和远期临床运用。Objective:To investigate the value of spectral CT multi-parameters in the preoperative prediction of perineural invasion(PNI)in colorectal cancer(CRC).Methods:The clinical and imaging data of 123 CRC patients from March 2021 to May 2023 were retrospectively analyzed.Among them,71 cases were identified as PNI and 52 cases as non-PNI(nPNI).All CRC patients underwent whole-abdomen triple phase enhanced spectral CT examination before surgery.Iodine concentration(IC),normalized IC(nIC)and slope of the spectral curve(λH)at triple-phase contrast enhanced scanning were measured,and the tumor extracellular volume fraction(ECV)was calculated.The diffe-rences in each parameter were compared between the PNI group and the nPNI group.Parameters with statistically significant difference were included in binary logistic regression analyses to construct a combined prediction model,and effectiveness of differential diagnosis of each parameter and the combined model for differentiating between PNI and nPNI was tested by the receiver operating characte-ristic(ROC)curves.Results:The measured values of IC andλH on triple-phase enhanced images,and NIC and ECV on venous and delayed phase images in the PNI group were greater than those in the nPNI group,and the differences were statistically significant(P<0.05),whereas the difference between arterial phase nIC was not statistically significant between the two groups(P>0.05).The area under the curve(AUC)of the single parameter for predicting PNI status was 0.632~0.820,the AUC of the combined model for predicting PNI status was 0.859,with sensitivity and specificity of 76.10%and 86.50%,respectively.Conclusion:Spectral CT multi-parameters were able to predict preoperatively PNI in CRC,which is expected to promote further research and long-term clinical application of spectral CT in preoperative prediction of CRC-PNI.

关 键 词：结直肠癌 神经侵犯 能谱CT 细胞外容积分数 

分 类 号：R814.42[医药卫生—影像医学与核医学] R735.3[医药卫生—放射医学]