Li Wei, Pan Yuanming, Cai Siyu, Xie Shenglong, He Bin
Department of Thoracic Surgery, Sichuan Provincial People's Hospital, University of Electronic Science and Technology of China, Chengdu, China.
Cancer Research Center, Beijing Chest Hospital, Capital Medical University, Beijing Tuberculosis and Thoracic Tumor Research Institute, Beijing, China.
Transl Cancer Res. 2025 Aug 31;14(8):4649-4661. doi: 10.21037/tcr-2025-12. Epub 2025 Aug 22.
Some lung cancer patients are pathologically confirmed to have lung adenocarcinoma with neuroendocrine differentiation (LUAD-ND). However, research on this subtype remains limited. This study aimed to systematically investigate the metastatic patterns and prognosis-related factors of LUAD-ND, and construct neural network-based prediction models for survival outcomes.
By analyzing the Surveillance, Epidemiology, and End Results (SEER) database, we employed the Cox proportional hazards model to investigate prognostic factors for overall survival (OS) and cancer-specific survival (CSS) in patients with LUAD-ND. We calculated hazard ratios (HRs) and 95% confidence intervals (CIs) and detailed the median survival time and specific time survival probabilities for different features in the LUAD-ND population. Finally, using a neural network algorithm, we developed a predictive model for forecasting LUAD-ND's OS and CSS, evaluating its performance using the area under the receiver operating characteristic curve (AUC).
Most patients with LUAD-ND were diagnosed at an advanced stage. The OS time of patients with LUAD-ND was 12 (95% CI: 10-14) months, and the CSS time was 14 (95% CI: 12-16) months. The most common distant metastatic sites were bone, followed by liver, brain, and lung. Surgery (HR =0.51; 95% CI: 0.31-0.82; P=0.006) and chemotherapy (HR =0.33; 95% CI: 0.21-0.50; P<0.001) were associated with improved OS. Similarly, surgery (HR =0.49; 95% CI: 0.28-0.84; P=0.01) and chemotherapy (HR =0.31; 95% CI: 0.19-0.49; P<0.001) were linked to better CSS. The neural network-based tool can effectively predict the prognosis of LUAD-ND, achieving an AUC of 0.852-0.864 for 6-month OS and 0.835-0.883 for 6-month CSS.
Patients with LUAD-ND face a dismal prognosis, yet chemotherapy and surgical interventions can ameliorate their outcomes. The neural network tool developed in this study yields precise prognostic estimations.
部分肺癌患者经病理确诊为具有神经内分泌分化的肺腺癌（LUAD-ND）。然而，关于该亚型的研究仍然有限。本研究旨在系统地探究LUAD-ND的转移模式和预后相关因素，并构建基于神经网络的生存结局预测模型。
通过分析监测、流行病学和最终结果（SEER）数据库，我们采用Cox比例风险模型来研究LUAD-ND患者总生存（OS）和癌症特异性生存（CSS）的预后因素。我们计算了风险比（HR）和95%置信区间（CI），并详细列出了LUAD-ND人群中不同特征的中位生存时间和特定时间生存概率。最后，使用神经网络算法，我们开发了一个预测模型来预测LUAD-ND的OS和CSS，并使用受试者操作特征曲线下面积（AUC）评估其性能。
大多数LUAD-ND患者在晚期被诊断出来。LUAD-ND患者的OS时间为12（95%CI：10 - 14）个月，CSS时间为14（95%CI：12 - 16）个月。最常见的远处转移部位是骨，其次是肝、脑和肺。手术（HR = 0.51；95%CI：0.31 - 0.82；P = 0.006）和化疗（HR = 0.33；95%CI：0.21 - 0.50；P < 0.001）与OS改善相关。同样，手术（HR = 0.49；95%CI：0.28 - 0.84；P = 0.01）和化疗（HR = 0.31；95%CI：0.19 - 0.49；P < 0.001）与更好的CSS相关。基于神经网络的工具可以有效地预测LUAD-ND的预后，6个月OS的AUC为0.852 - 0.864，6个月CSS的AUC为0.835 - 0.883。
LUAD-ND患者预后不佳，但化疗和手术干预可以改善其结局。本研究开发的神经网络工具能得出精确的预后估计。
