Objectives: This study aimed to assess the performance of the LI-RADS tumor response algorithm in analyzing inter-reader agreement in patients with hepatocellular carcinoma (HCC) treated with Microwave Ablation (MWA) and Transarterial Embolization (TAE) and the relationship between inter-reader agreement and Neutrophils to Lymphocytes ratio dynamic variations at different time points to explore how inflammation influences tumor response and its interpretation on imaging. Methods: A retrospective analysis was conducted on 78 HCC patients treated with MWA or TAE. Two independent radiologists evaluated pre- and post-treatment imaging and assigned categories according to the LR-TRA. Inter-reader agreement was assessed with a focus on subgroup analysis considering the different locoregional treatments. NLR values, measured at baseline (T0), 72 h (T1), and 30 days post-procedure (T2), were compared with patients with concordant and discordant LR-TRA assessments. This analysis aimed to identify any association between NLR dynamics and inter-reader agreement on treatment response. Results: The inter-reader agreement in the LR-TRA application was “substantial” in the cases of MWA treatment evaluation (κ = 0.65), and “moderate” in the cases of TAE treatment evaluation (κ = 0.51). The differences in inter-reader agreement were found to be expressions of different levels of NLR mean values in the different time frames evaluated. Three days after treatment, NLR increased significantly in TAE groups. At 30 days, NLR had returned close to baseline levels but with NLR persisting higher in the TAE group. There was a statistically significant difference in NLR between the “mismatch” group (those with discrepant LR-TRA readings) and the “match” group at 3 days (p= 0.004) and late evaluation (30+ days). Conclusions: This study has shown that NLR levels can predict inter-reader discrepancies in LR-TRA assessment and may be translated into different levels of difficult imaging interpretation. Combining LR-TRA and NLR is promising for a more comprehensive assessment of tumor response and inflammatory dynamics.
目的：本研究旨在评估LI-RADS肿瘤疗效评估算法在分析经微波消融与经动脉栓塞治疗的肝细胞癌患者阅片者间一致性方面的表现，并探讨阅片者间一致性与不同时间点中性粒细胞-淋巴细胞比值动态变化之间的关系，以揭示炎症如何影响肿瘤疗效及其影像学解读。方法：对78例接受MWA或TAE治疗的HCC患者进行回顾性分析。由两名独立影像医师按照LR-TRA标准对治疗前后影像进行评估分类。重点针对不同局部治疗方案进行亚组分析以评估阅片者间一致性。将基线期、术后72小时及术后30天测量的NLR数值与LR-TRA评估结果一致组和不一致组进行比较，旨在探究NLR动态变化与治疗反应评估的阅片者间一致性是否存在关联。结果：在MWA治疗评估中，LR-TRA应用的阅片者间一致性为"高度一致"（κ=0.65）；在TAE治疗评估中为"中度一致"（κ=0.51）。阅片者间一致性的差异体现在不同评估时间窗内NLR平均值的水平差异。治疗后第3天，TAE组NLR显著升高；至30天时NLR虽恢复至接近基线水平，但TAE组NLR仍持续偏高。在治疗后第3天（p=0.004）及晚期评估（30天以上），LR-TRA评估"不一致组"与"一致组"的NLR存在统计学显著差异。结论：本研究表明NLR水平可预测LR-TRA评估中的阅片者间差异，这种差异可能转化为不同难度的影像解读。结合LR-TRA与NLR指标，有望为肿瘤疗效及炎症动态提供更全面的评估体系。
肿瘤(癌症)患者之家