LAUSR

Objectives Small cell lung cancer (SCLC) is a lethal human malignancy, and previous studies support the contribution of microRNA to cancer progression. The prognostic value of miR-219-5p in patients with SCLC remains unclear. This study aimed to evaluate the predictive value of miR-219-5p with respect to mortality in patients with SCLC and to incorporate miR-219-5p level into a prediction model and nomogram for mortality. Design Retrospective observational cohort study. Setting and participants Our main cohort included data from 133 patients with SCLC between 1 March 2010 and 1 June 2015 from the Suzhou Xiangcheng People’s Hospital. Data from 86 patients with non-SCLC at Sichuan Cancer Hospital and the First Affiliated Hospital of Soochow University were used for external validation. Outcome measures Tissue samples were taken during admission and stored, and miR-219-5p levels were measured at a later date. A Cox proportional hazard model was used for survival analyses and for analysing risk factors to create a nomogram for mortality prediction. The accuracy of the model was evaluated by C-index and calibration curve. Results Mortality in patients with a high level of miR-219-5p (≥1.50) (n=67) was 74.6%, while mortality in the low-level group (n=66) was 100.0%. Based on univariate analysis, we included significant factors (p<0.05) in a multivariate regression model: patients with high level of miR-219-5p (HR 0.39, 95% CI 0.26–0.59, p<0.001), immunotherapy (HR 0.44, 95% CI 0.23–0.84, p<0.001) and prognostic nutritional index score >47.9 (HR=0.45, 95% CI 0.24–0.83, p=0.01) remained statistically significant factors for improved overall survival. The nomogram had good accuracy in estimating the risk, with a bootstrap-corrected C-index of 0.691. External validation indicated an area under the curve of 0.749 (0.709–0.788). Conclusions The miR-219-5p level was associated with a reduced risk of mortality in patients with SCLC. A nomogram incorporating MiR-219-5p level and clinical factors demonstrated good accuracy in estimating the risk of overall mortality. Prospective validation of the prognostic nomogram is needed.