Validation of existing clinical prediction models for patients with solitary pulmonary nodules (SPN) managed by a lung multi-disciplinary team (MDT)

Background: Management of patients with SPNs depends critically on the pre-test probability of malignancy. There are currently two clinical predictions models for SPNs based on data from North America. However, these models have not been validated in UK patients, in particular those managed by a Lung MDT. Objective: To validate two existing clinical prediction models in patients with SPNs managed by the Lung MDT at a large teaching hospital. Methods: 175 patients with SPNs measuring 8–30 mm managed by the Lung MDT over 3 years (2007-2009) were identified retrospectively through the institutional Lung Cancer database. Data on age, smoking, cancer history, nodule size, location, spiculation, and final diagnosis was collected. Each case9s final diagnosis was compared with the probability of malignancy predicted by two models: the Mayo Clinic model and the Veteran Affairs (VA) one. The accuracy of each model was assessed by calculating areas under the receiver operating characteristic (ROC) curve and the models were calibrated by comparing predicted and observed rates of malignancy. Results: The area under the ROC curve for the Mayo model (0.832; 95% CI 0.753-0.911) was higher than that of the VA model (0.739; 95% CI 0.641-0.838). Calibration curves showed that both models slightly underestimated the probability of malignancy for patients across all deciles of predicted probabilities, except for those with highest probability of malignancy, where the VA model slightly overestimated probability. Conclusions: The two existing prediction models are sufficiently accurate to guide management of patients with SPNs managed by a Lung MDT.