Construction and validation of a predictive model for benefits of tunnel-type mediastinal lymph node dissection in lung cancer patients based on the SEER database

To identify the independent factors influencing the benefits of tunnel-type mediastinal lymph node dissection in lung cancer patients, construct and validate a nomogram model that can quantitatively predict the benefits of this surgical procedure and provide an evidence-based tool for individualized clinical surgical decision-making. Data on demographics, clinicopathology, surgical methods, and survival of stage N0-N2 lung cancer diagnosed between 2010 and 2022 were extracted from the SEER database for analysis. A total of 156 lung cancer patients admitted to our Hospital from July 2020 to June 2022 were included in the clinical external validation cohort. Propensity score matching (PSM) was used to balance confounding factors between groups, and the SEER data were divided into a training set and a validation set at a ratio of 7:3. The training set was used to construct the predictive model, and the validation set was used to evaluate the model. Univariate and multivariate Cox regression analyses were conducted to screen for independent predictive factors for post-surgical benefits in patients with lung cancer, and a nomogram for predicting lung cancer-specific survival (LCSS) was constructed based on these independent factors. The discriminative ability, calibration degree, and clinical utility of the model were evaluated using concordance index (C-index), ROC curve (AUC value), calibration curve, and decision curve analysis (DCA). Differences in surgical benefits among patients with different risk levels were analyzed using risk stratification. A total of 75,940 patients from the SEER database were included, and 24,230 patients were screened after PSM, including 12,115 patients in the tunnel-type mediastinal lymph node dissection group and 12,115 patients in the non-tunnel-type mediastinal lymph node dissection group. The patients were divided into a training set (16,961 cases) and validation set (7,269 cases) at a ratio of 7:3. Multivariate Cox regression analysis showed that surgical method, age, marital status, family income, residential type, TNM stage, and histological type were independent predictive factors for LCSS (all P < 0.05). Among these factors, tunnel-type dissection reduced the risk of lung cancer-specific death by 17.4% (HR = 0.826, 95% CI: 0.776–0.879). The C-index was 0.6585 (SE = 0.004), indicating that the model’s discriminative ability for LCSS was moderate. This performance is consistent with common observational predictive models based on the SEER database, reflecting the complexity of lung cancer prognosis affected by tumor heterogeneity and unmeasured confounding factors. The core clinical value of the model lies was in risk stratification (low/medium/high risk) rather than in precise individual-level death probability prediction. The AUC values of the training and validation sets for predicting the 1-, 3-, and 5-year LCSS were 0.674–0.679, 0.682–0.693, and 0.688–0.697, respectively. The calibration curve showed that the predicted values were consistent with the actual survival outcomes, and DCA confirmed that the model had significant net benefits within the commonly used clinical risk thresholds. Risk stratification showed that patients in the high-risk group benefited most significantly from tunnel-type dissection (HR = 0.779, P < 0.001), with a 4.3% increase in the 3-year survival rate. Clinical external validation showed that patients in the medium- and high-risk groups had significant benefits from the surgical procedure (all P < 0.05), while there was no statistically significant difference in the low-risk group (P > 0.05). The predictive model constructed has good discriminative ability and calibration degree and can effectively identify the population that benefits from tunnel-type mediastinal lymph node dissection. Tunnel-type mediastinal lymph node dissection can significantly reduce the risk of lung cancer-specific death in patients with lung cancer, and the benefits are dependent on risk stratification; patients in the medium- and high-risk groups benefit the most. This model provides a quantitative tool for the selection of surgical procedures for patients with lung cancer and promotes the development of lung cancer surgery for precision and individualization.