Department of Radiology, Division of Life Sciences and Medicine, The First Affiliated Hospital of the University of Science and Technology of China, Huanhu Rd 107, Hefei, Anhui 230031, PRC. Address correspondence to J. Dong (
AJR Am J Roentgenol. 2021 May;216(5):1335-1343. doi: 10.2214/AJR.20.23516. Epub 2021 Mar 24.
The purpose of our study was to assess the value of combining quantitative dual-energy CT (DECT) parameters with qualitative morphologic parameters for the preoperative prediction of cervical nodal metastasis from papillary thyroid carcinoma (PTC). Thirty-five patients with pathologically proven PTC underwent single-phase contrast-enhanced DECT before thyroidectomy and cervical lymphadenectomy. Analyses of quantitative DECT parameters and qualitative morphologic features of metastatic and benign lymph nodes (LNs) were independently performed. The diagnostic performances of using only quantitative parameters, only morphologic features, and their combination for predicting cervical nodal metastasis were statistically calculated with ROC curves and logistic regression models. A total of 206 LNs, 80 metastatic and 126 benign, were included. The best single performer in DECT was the normalized iodine concentration in the venous phase, which had low sensitivity (62.5%) but high specificity (85.7%), for diagnosing metastatic cervical LNs. On the other hand, the best single performer in qualitative morphologic parameters was using the criterion of shortest diameter of greater than 5 mm, which had low specificity (69.8%) but high sensitivity (86.3%). Combining these two parameters improved the AUC, sensitivity, and specificity to 0.846, 86.3%, and 72.2%, respectively. The combination of multiple quantitative DECT parameters and all morphologic data further improved AUC, sensitivity, and specificity to 0.878, 87.5%, and 73.8%, respectively, which was significant compared with the use of any single parameter. The combination of quantitative DECT parameters with morphologic data improves performance in the preoperative diagnosis of metastatic cervical LNs in patients with PTC.
我们的研究目的是评估定量双能 CT（DECT）参数与定性形态学参数相结合对甲状腺乳头状癌（PTC）颈淋巴结转移的术前预测价值。35 例经病理证实的 PTC 患者在甲状腺切除术和颈淋巴结清扫术前行单相对比增强 DECT 检查。对转移和良性淋巴结（LNs）的定量 DECT 参数和定性形态特征进行独立分析。使用仅定量参数、仅形态特征及其组合来预测颈淋巴结转移的诊断性能通过 ROC 曲线和逻辑回归模型进行统计计算。共纳入 206 个 LNs，80 个转移性和 126 个良性。静脉期标准化碘浓度是 DECT 中预测转移性颈淋巴结的最佳单一指标，其敏感性（62.5%）较低，但特异性（85.7%）较高。另一方面，定性形态参数中最佳的单一指标是使用直径大于 5mm 的标准，其特异性（69.8%）较低，但敏感性（86.3%）较高。将这两个参数结合起来可以将 AUC、敏感性和特异性提高到 0.846、86.3%和 72.2%。将多个定量 DECT 参数和所有形态数据结合起来，可以将 AUC、敏感性和特异性分别提高到 0.878、87.5%和 73.8%，与使用任何单一参数相比，这均有显著改善。定量 DECT 参数与形态数据的结合提高了术前诊断 PTC 患者转移性颈淋巴结的性能。
