Reid M, et al. - In this retrospective analysis, researchers proposed a malignancy probability model to estimate the likelihood of malignancy for pulmonary nodules considered high enough risk to recommend biopsy. They generated and internally validate eight models by using 200 malignant nodules and 101 benign nodules. Age, smoking history, upper lobe location, solid and irregular/spiculated nodule edges, emphysema, fluorodeoxyglucose-PET avidity, and history of cancer other than lung were the factors predictive of lung cancer used in the final models. The range of concordance index (C-index) of the models was from 0.75 to 0.81. For the proposed model vs the Mayo Clinic model, the C-index was 0.67 vs 0.63, respectively, in an independent sample used for validation. Compared with the Mayo Clinic model, a greater potential to impact clinical decisions was displayed by the ratio of malignant to benign nodules within each probability decile.