# Development and validation of nomograms to predict survival of neuroendocrine carcinoma in genitourinary system: A population-based retrospective study

## TL;DR

This study creates and validates models to predict survival for patients with neuroendocrine carcinoma in the genitourinary system.

## Contribution

The study introduces new nomograms using LASSO and FGM models to improve survival prediction for NEC patients.

## Key findings

- The FGS model showed higher net benefits for disease-specific survival analysis.
- Nomograms were validated using bootstrap self-sampling and ROC curves.
- The FGS model excluded additional occurrences to better benefit patients.

## Abstract

Neuroendocrine carcinoma (NEC) is a rare yet potentially perilous neoplasm. The objective of this study was to develop prognostic models for the survival of NEC patients in the genitourinary system and subsequently validate these models. A total of 7125 neuroendocrine neoplasm (NEN) patients were extracted. Comparison of survival in patients with different types of NEN before and after propensity score-matching (PSM). A total of 3057 patients with NEC, whose information was complete, were extracted. The NEC influencing factors were chosen through the utilization of the least absolute shrinkage and selection operator regression model (LASSO) and the Fine & Gary model (FGM). Furthermore, nomograms were built. To validate the accuracy of the prediction, the efficiency was verified using bootstrap self-sampling techniques and receiver operating characteristic curves. LASSO and FGM were utilized to construct three models. Confirmation of validation was achieved by conducting analyses of the area under the curve and decision curve. Moreover, the FGS (DSS analysis using FGM) model produced higher net benefits. To maximize the advantages for patients, the FGS model disregarded the influence of additional occurrences. Patients are expected to experience advantages in terms of treatment options and survival assessment through the utilization of these models.