A predictive model for distant metastasis of mediastinal neuroendocrine neoplasms

LU Jieyu; PAN Jialing; QIANG Xinyun; XU Ruitong

doi:10.3969/j.issn.1674-4136.2024.06.005

Abstract：Objective To evaluate the independent risk factors for distant metastasis of mediastinal neuroendocrine neoplasms and to establish a predictive model.Methods A total of 392 patients with mediastinal neuroendocrine neoplasms diagnosed between 2000 and 2021 were included in the SEER database,and the data of covatiates ( including gender,age at diagnosis,race,histological classification,tumor size,primary tumor invasion extent,and regional lymph node metastasis) was collected.Univariate and multivariate Logistic regression analysis were conducted to identify the predictors of distant metastasis and subsequently construct a predictive nomogram,and the AUC was used to evaluate the modle discriminatory,and the decision curve analysis was carried out to evaluate the clinical application value.Results A total of 114 patients ( 29.1％) were found to have distant metastasis.Logistic regression analysis showed that the primary tumor invasion extent and regional lymph node metastasis were independently correlated with the appearance of distant metastasis.Based on these findings,a predictive nomogram model was established,and the AUC was 0.763.The empirical evaluation verified the good discrimination ability and significant clinical application value of the model.Conclusions Primary tumor invasion extent and regional lymph node metastasis are independent risk factors of distant metastasis for mediastinal neuroendocrine neoplasms.The nomogram model established based on this result can serve as a valuable tool for providing some clues for clinical prognosis evaluation and the formulation of examination and treatment strategies.