Establishment and validation of a prediction model for the probability of malignancy in solid solitary pulmonary nodules in northwest China

Background and Objectives To construct a prediction model of solitary pulmonary nodules (SPNs), to predict the possibility of malignant SPNs in patients aged 15-85 years in northwest China for clinical diagnostic and therapeutic decision-making. Methods The features of SPNs were assessed by multivariate logistic regression, followed by visualization using a nomogram. Hosmer lemeshow was applied to evaluate the fitting degree of the model. The area under the receiver operating characteristic (ROC) curve was identified to determine the discriminative ability of the model. Results Lobulation, spiculation, pleural-tag, carcinoembryonic antigen, neuron-specific enolase, and total serum protein were independent predictors of malignant pulmonary nodules (p < .05). Lobulation (100 points) scored the highest in the nomogram, and the Hosmer-Lemeshow goodness-of-fit statistic was 0.805 (p > .05). The area under curve (AUC) of the modeling and validation groups using logistic regression were 0.859 (95% CI, 0.805-0.903) and 0.823 (95% CI, 0.738-0.890), respectively. Moreover, the AUC of our model was higher than that of the Mayo model, VA model, and Peking University (AUC 0.823 vs. 0.655 vs. 0.603 vs. 0.521). Conclusion Our prediction model is more suitable for predicting the possibility of malignant SPNs in northwest China, and can be calculated using a nomogram to determine further treatments.