Abstract:

Objective　To evaluate the value of normalized iodine concentrations (NIC) in energy spectrum enhanced CT in the diagnosis of lymph node metastasis of colorectal cancer by meta analysis. Methods　Pubmed, Embase, Cochrane, Web of Science, CBM, CNKI, Wanfang, and VIP databases were searched from the establishment of the databases to October 2022, and the literatures related to NIC in the diagnosis of lymph node metastasis of colorectal cancer were screened. The risk of bias was evaluated in the literatures using the Quality Assessment of Diagnostic Accuracy Studies tool-2, the Meta-DiSc1.4 was used to perform data statistics, the Spearman correlation coefficient was calculated, and the summary receiver operating characteristic curve (sROC) was plotted to analyze the heterogeneity caused by the presence or absence of threshold effects. Fixed-effects combined effect sizes were used for I² <50%; random-effects combined effect sizes were used for I² ≥50%. The sensitivity analysis was performed by Stata 14.0, and the publication bias was detected by Deeks funnel plot. Results　Seven literatures with 717 lymph nodes were included, including 314 metastatic lymph nodes and 403 non-metastatic lymph nodes. They were divided into two subgroups according to arterial phase (AP) and venous phase (VP) of energy spectrum enhanced CT. Meta DiSc1.4 showed that there was no heterogeneity caused by threshold effects in both groups (both P>0.05). Further, according to the I² results, the corresponding model was selected to combine the effect sizes. The sensitivities of AP and VP were 0.76 (0.71-0.80) and 0.67 (0.61-0.73), the specificities were 0.71 (0.66-0.75) and 0.87 (0.82-0.91), the positive likelihood ratios were 2.83 (1.85-4.31) and 5.39 (3.79-7.67), the negative likelihood ratios were 0.33 (0.23-0.46) and 0.34 (0.22-0.54), the diagnostic odds ratios (DORs) were 10.00 (4.38-22.82) and 17.27 (10.39-28.69), and the areas under the curve (AUCs) were 0.835 4 and 0.894 5. Subgroup-AP and subgroup-VP sensitivity analysis showed stable studies and no publication bias (P>0.05). Conclusions　NIC helps to distinguish metastatic and non-metastatic lymph nodes in patients with colorectal cancer. In terms of sensitivity, the NIC of AP is better; in terms of specificity, positive likelihood ratio, negative likelihood ratio, DOR, and AUC, the NIC of VP has higher diagnostic value.

Key words:

Colorectal cancer, Lymph node, Spectral CT, NIC, Meta-analysis

摘要：

目的　采用meta分析探讨能谱增强CT中标准化碘密度（normalized iodine concentrations，NIC）诊断结直肠癌淋巴结转移的价值。方法　检索建库至2022年10月期间PubMed、Embase、Cochrane、Web of Science、CBM、CNKI、万方、维普数据库，筛选NIC诊断结直肠癌淋巴结转移相关的文献。采用诊断准确性研究的质量评价工具-2对文献进行偏倚风险评价，运用Meta-DiSc1.4进行数据统计，计算Spearman相关系数并绘制综合受试者工作特征曲线（sROC）以分析有无阈值效应引起的异质性。I²<50%采用固定效应合并效应量；I²≥50%采用随机效应合并效应量。采用Stata 14.0进行敏感性分析，以Deeks漏斗图检测发表偏倚。结果　纳入7篇文献，共计717枚淋巴结，转移淋巴结314枚、非转移淋巴结403枚。按照能谱增强CT中的动脉期（arterial phase，AP）和静脉期（venous phase，VP）将其分为两个亚组。Meta DiSc1.4显示，两亚组均不存在阈值效应引起的异质性（均P>0.05）。进而根据I²结果，选择对应的效应模型合并效应量。结果（AP、VP）分别为：灵敏度=0.76（0.71～0.80）、0.67（0.61～0.73）；特异度=0.71（0.66～0.75）、0.87（0.82～0.91）；阳性似然比=2.83（1.85～4.31）、5.39（3.79～7.67）；阴性似然比=0.33（0.23～0.46）、0.34（0.22～0.54）；诊断比值比（diagnostic odds ratio，DOR）=10.00（4.38～22.82）、17.27（10.39～28.69）；曲线下面积（area under the curve，AUC）=0.835 4、0.894 5。亚组-AP及亚组-VP敏感性分析显示研究稳定，并且不存在发表偏倚（P>0.05）。结论　NIC有助于区分结直肠癌患者转移性和非转移性淋巴结。在灵敏度方面，AP的NIC更优；在特异度、阳性似然比、阴性似然比、DOR、AUC方面，VP的NIC具有更高的诊断价值。

关键词:

结直肠癌, 淋巴结, 能谱CT, NIC, meta分析