Limited Utility of Pulmonary Nodule Risk Calculators for Managing Large Nodules

Publication date: Available online 8 April 2017
Source:Current Problems in Diagnostic Radiology
Author(s): Mark M. Hammer, Arun C. Nachiappan, Eduardo Mortani Barbosa
Rationale and ObjectivesThe optimal management of large pulmonary nodules, at higher risk for lung cancer, has not been determined, and it remains unclear which patients should undergo follow-up imaging versus invasive tissue diagnosis via biopsy or surgical resection.Materials and MethodsThrough search of radiology reports, 86 nodules from our institution were identified using the inclusion criterion of solid nodules measuring greater than 8 mm. We evaluated these nodules with a number of risk prediction calculators, including the Brock University model, and compared these against the proven diagnosis.Results59 (69%) of the nodules were malignant. The most accurate predictive model, the Brock University calculator, underestimated the risk for this group at 33%. At its optimal threshold, this model had a positive predictive value of 81% and negative predictive value of 53%. Notwithstanding the low negative predictive value, the positive predictive value was no better than patients clinically selected for biopsy (86% of biopsies were malignant).ConclusionExisting nodule risk prediction calculators are of limited utility in guiding the management of large pulmonary nodules. At present, the accuracy of these models in this setting is inferior to expert clinical judgment, and future work is needed to develop management algorithms for higher-risk nodules.



http://ift.tt/2oPza20