A mathematical model for predicting malignancy of solitary pulmonary nodules

World J Surg. 2012 Apr;36(4):830-5. doi: 10.1007/s00268-012-1449-8.

Abstract

Background: The goal of the present study was to differentiate between benign and malignant solitary pulmonary nodules (SPN) by developing a mathematical prediction model.

Methods: Records from 371 patients (197 male, 174 female) with SPN between January 2000 and September 2009 were reviewed (group A). Clinical data were collected to estimate the independent predictors of malignancy of SPN with multivariate logistic regression analysis. A clinical prediction model was subsequently developed. Between October 2009 and May 2011, data from an additional 145 patients with SPN were used to validate this new clinical prediction model (group B). The same data were also estimated with two previously published models for comparison with our new model.

Results: The median patient age was 57.1 years in group A; 54% of the nodules were malignant and 46% were benign. Logistic regression analysis identified six clinical characteristics (age, diameter, border, calcification, spiculation, and family history of tumor) as independent predictors of malignancy in patients with SPN. The area under the receiver operator characteristic (ROC) curve for our model (0.874 ± 0.028) was higher than those generated using the other two reported models. In our model, sensitivity = 94.5%, specificity = 70.0%, positive predictive value = 87.8%, and negative predictive value = 84.8%).

Conclusions: Age, diameter, border, calcification, spiculation, and family history of tumor were independent predictors of malignancy in patients with SPN. Our prediction model was sufficient to estimate malignancy in patients with SPN and proved to be more accurate than the two existing models.

MeSH terms

  • Area Under Curve
  • Female
  • Humans
  • Logistic Models
  • Lung Neoplasms / pathology*
  • Male
  • Middle Aged
  • Models, Biological*
  • Predictive Value of Tests
  • Solitary Pulmonary Nodule / pathology*