旨在调查指示分化型甲状腺癌 (DTC) 淋巴结 (LN) 强烈增强的定量 CT 参数的诊断性能和观察者间一致性，并将其与具有不同经验的放射科医生的定性分析进行比较。本研究纳入了 399 名 DTC 患者的 463 个 LN。三位放射科医生独立分析了 CT 上的强淋巴结增强。强增强的定性分析被定义为 LN 皮质在动脉期显示出比邻近肌肉更大的增强。定量分析包括 LN 在动脉期 (LNA) 和静脉期 (LNV) 上的平均衰减值 (MAV)、标准化为颈总动脉 (NAVCCA)、颈内静脉 (NAVIJV) 和胸锁乳突肌 (NAVSCM) 的 LNA，衰减差[AD; (LNA - MAVSCM)]和相对洗脱率[((LNA - LNV)/LNA) × 100]。对定量和定性分析的观察者间一致性和诊断性能进行了评估。所有定量 CT 参数（ICC，0.83-0.94）的观察者间一致性非常好，定性评估也很重要（κ = 0.61）。除 LNV 之外的所有 CT 参数均对转移性 LN 表现出良好的诊断性能（AUC，0.81-0.85）。 NAVCCA（0.85，95% CI：0.8-0.9）和 AD（0.85，95% CI：0.81-0.89）的 AUC 最高。除 NAVIJV 之外的所有定量参数的 AUC 均显着高于缺乏经验的放射科医生的定性评估，与经验丰富的放射科医生的评估没有显着差异。动脉期 CT 上 LN 增强的定量评估显示出比缺乏经验的放射科医生的定性分析更高的观察者间一致性和 AUC 值，支持基于标准化定量 CT 参数的模型来确定强 LN 增强的需要。在评估 DTC 中强 LN 增强时，指示强增强的定量 CT 参数可以提高观察者间的一致性，无论经验水平如何。因此，建立基于定量CT参数的标准化诊断模型可能是必要的。术前准确评估甲状腺癌淋巴结转移至关重要。表明 LN 强烈增强的定量 CT 参数表现出良好的观察者间一致性和良好的诊断性能。对比增强的定量评估为识别转移性 LN 提供了更客观的模型。© 2024。作者获得欧洲放射学会的独家许可。

To investigate the diagnostic performance and interobserver agreement of quantitative CT parameters indicating strong lymph node (LN) enhancement in differentiated thyroid cancer (DTC), comparing them with qualitative analysis by radiologists of varying experience.This study included 463 LNs from 399 patients with DTC. Three radiologists independently analyzed strong LN enhancement on CT. Qualitative analysis of strong enhancement was defined as LN cortex showing greater enhancement than adjacent muscles on the arterial phase. Quantitative analysis included the mean attenuation value (MAV) of LN on arterial phase (LNA) and venous phase (LNV), LNA normalized to the common carotid artery (NAVCCA), internal jugular vein (NAVIJV), and sternocleidomastoid muscle (NAVSCM), attenuation difference [AD; (LNA - MAVSCM)], and relative washout ratio [((LNA - LNV)/LNA) × 100]. The interobserver agreement and diagnostic performance of the quantitative and qualitative analyses were evaluated.Interobserver agreement was excellent for all quantitative CT parameters (ICC, 0.83-0.94) and substantial for qualitative assessment (κ = 0.61). All CT parameters except for LNV showed good diagnostic performance for metastatic LNs (AUC, 0.81-0.85). NAVCCA (0.85, 95% CI: 0.8-0.9) and AD (0.85, 95% CI: 0.81-0.89) had the highest AUCs. All quantitative parameters except for NAVIJV had significantly higher AUCs than qualitative assessments by inexperienced radiologists, with no significant difference from assessments by an experienced radiologist.Quantitative assessment of LN enhancement on arterial phase CT showed higher interobserver agreement and AUC values than qualitative analysis by inexperienced radiologists, supporting the need for a standardized quantitative CT parameter-based model for determining strong LN enhancement.When assessing strong LN enhancement in DTC, quantitative CT parameters indicating strong enhancement can improve interobserver agreement, regardless of experience level. Therefore, the development of a standardized diagnostic model based on quantitative CT parameters might be necessary.Accurate preoperative assessment of LN metastasis in thyroid cancer is crucial. Quantitative CT parameters indicating strong LN enhancement demonstrated excellent interobserver agreement and good diagnostic performance. Quantitative assessment of contrast enhancement offers a more objective model for the identification of metastatic LNs.© 2024. The Author(s), under exclusive licence to European Society of Radiology.