Validation of three lung cancer risk prediction models in patients with pulmonary nodules

  Objective  To verify the diagnostic ability of three lung cancer risk prediction models by using clinical and imaging information of pulmonary nodules, and explore the clinical value of these prediction models.

  Methods  Clinical and imaging data of patients with pulmonary nodules admitted in the First Medical Center of Chinese PLA General Hospital from August 2015 to July 2016 were retrospectively analyzed, and then they were verified by Mayo model, VA model and Peking University model, respectively. The sensitivity, specificity, positive predictive value and negative predictive value of the models were compared.

  Results  A total of 407 patients with pulmonary nodules were included in the study, with average age of (56.09±9.56) years. There were 205 males and 202 females. Pathological results showed 84 (20.6%) nodules were benign, and 323(79.4%) nodules were malignant. Among the three models, Peking University model achieved the highest sensitivity (78.6%), and Mayo model achieved the highest specificity (76.2%). The area under ROC curve of Mayo model, VA model and Peking University model was 0.608, 0.550, and 0.615, respectively.

  Conclusion  Lung cancer risk prediction model of Peking University People’s Hospital is superior to Mayo model and VA model in the diagnosis of lung cancer.