甲状腺微小乳头状癌淋巴结转移的准确术前预测:迈向患者的最佳管理
本研究利用1004例接受甲状腺切除术加中央颈淋巴结清扫的甲状腺微小乳头状癌(PTMC)患者数据,开发并验证了术前预测淋巴结转移(LNM)的列线图模型。模型在训练队列(n=702)和验证队列(n=302)的AUC分别为0.733和0.771,校准和临床效用良好。该模型有助于细化「低风险」分类,为个体化治疗提供工具。
要点
- PTMC的LNM独立预测因素包括年龄、肿瘤最大直径、结节-包膜距离、包膜接触长度、双侧可疑病灶、淋巴门缺失、微钙化和性别。
- 肿瘤大于7 mm、结节距包膜小于3 mm、包膜接触长度大于1 mm与更高的LNM率相关。
- 预测模型在训练队列和验证队列的AUC分别为0.733和0.771,校准和临床效用良好。
核心结论
- 本研究提出并验证了PTMC患者LNM的术前预测模型,提高了「低风险」分类的精确性,为个体化治疗提供了有价值的工具。
研究设计与样本
- 研究纳入1004例接受甲状腺切除术加中央颈淋巴结清扫的PTMC患者,分为训练队列(n=702)和验证队列(n=302)。
- 采用多变量逻辑回归识别PTMC的独立LNM预测因素,并构建预测列线图模型。
- 通过ROC分析、校准曲线分析和决策曲线分析评估模型性能。
主要发现
- PTMC的LNM预测因素包括年龄、肿瘤最大直径、结节-包膜距离、包膜接触长度、双侧可疑病灶、淋巴门缺失、微钙化和性别。
- 肿瘤大于7 mm、结节距包膜小于3 mm、包膜接触长度大于1 mm的患者LNM率更高。
- 模型在训练队列的AUC为0.733,在验证队列的AUC为0.771,校准和临床效用良好。
关键数值与推荐
- 肿瘤大于7 mm、结节距包膜小于3 mm、包膜接触长度大于1 mm与更高的LNM率相关。
- 模型在训练队列AUC为0.733,验证队列AUC为0.771。
临床怎么用
- 该模型为临床医生提供了个体化治疗的工具,有助于对原本可能建议观察或消融治疗的患者进行更好的管理。
局限性
- 原文未报告局限性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
OBJECTIVE: To enhance the accuracy in predicting lymph node metastasis (LNM) preoperatively in patients with papillary thyroid microcarcinoma (PTMC), refining the "low-risk" classification for tailored treatment strategies. METHODS: This study involves the development and validation of a predictive model using a cohort of 1004 patients with PTMC undergoing thyroidectomy along with central neck dissection. The data was divided into a training cohort (n = 702) and a validation cohort (n = 302). Multivariate logistic regression identified independent LNM predictors in PTMC, leading to the construction of a predictive nomogram model. The model's performance was assessed through ROC analysis, calibration curve analysis, and decision curve analysis. RESULTS: Identified LNM predictors in PTMC included age, tumor maximum diameter, nodule-capsule distance, capsular contact length, bilateral suspicious lesions, absence of the lymphatic hilum, microcalcification, and sex. Especially, tumors larger than 7 mm, nodules closer to the capsule (less than 3 mm), and longer capsular contact lengths (more than 1 mm) showed higher LNM rates. The model exhibited AUCs of 0.733 and 0.771 in the training and validation cohorts respectively, alongside superior calibration and clinical utility. CONCLUSION: This study proposes and substantiates a preoperative predictive model for LNM in patients with PTMC, honing the precision of "low-risk" categorization. This model furnishes clinicians with an invaluable tool for individualized treatment approach, ensuring better management of patients who might be proposed observation or ablative options in the absence of such predictive information.
引用
- Lin SY, Li MY, Zhou CP, et al. Accurate preoperative prediction of nodal metastasis in papillary thyroid microcarcinoma: Towards optimal management of patients. Head Neck. 2024. PMID: 38441255. doi:10.1002/hed.27720
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。