A prognostic nomogram based on LASSO Cox regression in patients with pulmonary large cell neuroendocrine carcinoma

Abstract

Abstract Background This study was designed to construct and validate a new nomogram that was available for predicting overall survival (OS)in patients with pulmonary large-cell neuroendocrine carcinoma (LCNEC). Methods 1864 LCNEC patients in total, of which 556 patients served as the validation group and 1308 patients as the training cohort, were drawn from the SEER database. We created a new nomogram for the training cohort using independent factors for overall survival (OS) via LASSO Cox regression, ultimately selecting five independent factors by a stepwise forward process. The nomogram contains all of the Cox regression results. Every result of the Cox regression is contained in the nomogram. Analyses of the calibration curve, decision curve, area under the curve (AUC), and concordance index (C-index) values were performed to assess the effectiveness and discriminative capability of the nomogram (DCA). NRI and IDI index were also used to assess the nomogram's performance in predicting OS against the 7th edition AJCC TNM staging system.Finally, we used the time-dependent ROC curves to assess the specificity and sensitivity of the newly constructed nomogram against the 7th edition AJCC TNM staging system at 3 years and 5years. Results Five optimal variablesfor predictive factors for OS were selected and merged to construct a 3- and 5 years OS nomogram.In the training cohort, the nomogram's C-index was 0.716, and it was 0.708 in the validation cohort. The actual survival rate and the calibration curves showing the nomogram forecasts were in good agreement. Results from the DCA in two cohorts demonstrated the clinical validity and applicability of the nomogram. The training cohort's NRI for 3 and 5 years OS was 0.296 and 0.161, while the IDI was 0.054 and 0.055 for predicting OS in comparison to the 7th edition AJCC TNM staging scheme. In the validation cohort, the NRI for 3 and 5 years OS was 0.289 and − 0.009, while the IDI was 0.053 and 0.041. Comparing the new nomogram andthe 7th edition AJCC TNM staging system in predicting 3-years OS were 0.811and 0.820,and 5-years OS were 0.821 and 0.828 in the training cohort. Moreover,Survival curves for the low-risk and high-risk groups were plotted using Kaplan-Meier analysis. Conclusions The prognostic nomogram may be very helpful in estimating the overall survival of patients with pulmonary LCNEC.