Jiang Xia, Cao Zhenyu, Fu Weiqi, Xie Zongyu, Shi Hengfeng, Tian Fengjuan, Yang Yang, Wang Jian, Zhao Li
Department of Radiology, Tongde Hospital of Zhejiang Province Afflicted to Zhejiang Chinese Medical University (Tongde Hospital of Zhejiang Province), Hangzhou, China.
Department of Radiology, Taizhou Municipal Hospital, Taizhou, China.
Quant Imaging Med Surg. 2025 Jun 6;15(6):5410-5423. doi: 10.21037/qims-24-2016. Epub 2025 Jun 3.
Non-small cell lung cancer (NSCLC) is a leading cause of cancer-related death worldwide. The accurate preoperative prediction of lymph node (LN) status is crucial for reducing NSCLC mortality rates and formulating treatment strategies. This study aimed to develop and validate a nomogram for the individualized prediction of lymph node metastasis (LNM) risk in NSCLC patients, thereby guiding therapeutic decision making.
The complete clinical, computed tomography (CT) imaging, and pathological data of patients with pathologically confirmed NSCLC at three independent tertiary medical centers from January 2014 to October 2023 were retrospectively collected and analyzed. In total, 2,725 patients were enrolled from Centers 1 and 2, and 112 patients were enrolled from center 3. The patients were randomly divided into training and test datasets at an 8:2 ratio. A clinical CT imaging predictive model was established by binary logistic regression analysis, and a corresponding nomogram chart was constructed. The predictive ability, discriminative ability, accuracy, and clinical benefit of the model were evaluated by receiver operating characteristic (ROC) curve, calibration curve, decision curve, and clinical impact curve (CIC) analyses.
Mixed ground-glass opacity (mGGO), obstructive inflammation, emphysema or bullae, a short diameter (SD), and heterogeneous ventilation or perfusion (HVP) were identified as independent risk factors for LN positivity in NSCLC. The model had an area under the curve (AUC) of 0.893, an accuracy of 76.8%, a sensitivity of 91.6%, and a specificity of 74.2% on the training dataset. While it had an AUC of 0.907, an accuracy of 86.2%, a sensitivity of 89.3%, and a specificity of 80.9% on the test dataset.
We developed and validated a robust clinical CT imaging nomogram model that effectively predicts LN positivity in NSCLC.
非小细胞肺癌（NSCLC）是全球癌症相关死亡的主要原因。准确的术前淋巴结（LN）状态预测对于降低NSCLC死亡率和制定治疗策略至关重要。本研究旨在开发并验证一种列线图，用于个体化预测NSCLC患者的淋巴结转移（LNM）风险，从而指导治疗决策。
回顾性收集并分析了2014年1月至2023年10月在三个独立的三级医疗中心病理确诊为NSCLC患者的完整临床、计算机断层扫描（CT）影像和病理数据。中心1和中心2共纳入2725例患者，中心3纳入112例患者。患者按8:2的比例随机分为训练集和测试集。通过二元逻辑回归分析建立临床CT影像预测模型，并构建相应的列线图。通过受试者操作特征（ROC）曲线、校准曲线、决策曲线和临床影响曲线（CIC）分析评估模型的预测能力、鉴别能力、准确性和临床获益。
混合性磨玻璃影（mGGO）、阻塞性炎症、肺气肿或肺大疱、短径（SD）以及不均匀通气或灌注（HVP）被确定为NSCLC患者LN阳性的独立危险因素。该模型在训练集上的曲线下面积（AUC）为0.893，准确率为76.8%，灵敏度为91.6%，特异度为74.2%。在测试集上，其AUC为0.907，准确率为86.2%，灵敏度为89.3%，特异度为80.9%。
我们开发并验证了一个强大的临床CT影像列线图模型，该模型能有效预测NSCLC患者的LN阳性。
