The relative value of 18F-FDG PET interpretation criteria in generating a probabilistic approach for characterization of pulmonary nodules.

PURPOSE: The evaluation of pulmonary nodules constitutes a large part of PET-CT studies. In this study, we aimed to evaluate the different interpretation criteria in F-fluorodeoxyglucose PET attenuation-corrected and non-attenuation-corrected studies as individual predictors of malignancy in order to propose a useful combination of criteria that can be used in daily practice to classify nodules appropriately.

PATIENTS AND METHODS: We performed a historical prospective survey of all consecutive patients referred to our service for the initial assessment of pulmonary nodules and sought the final characterization of these nodules either from tissue sampling or from radiological and clinical follow-up.

RESULTS: A total of 104 nodules from 82 patients were included, with a prevalence of malignancy of 53%. Absence of uptake on non-attenuation-corrected studies was found to be the best predictive criterion for benignancy, with a negative predictive value of 97%, and the highest relative risk for malignancy, with a value of 20.9. Uptake higher than that of the mediastinum on attenuation-corrected images was found to be the best criterion for predicting malignancy, with a positive predictive value of 89% and a sensitivity of 73%, which is slightly better than the use of a maximal standardized uptake value cutoff of 3.0. By combining our best negative and positive criteria, we were able to classify 71% (74/104) of the lung nodules with a high level of confidence. More specifically, these two criteria allowed the correct classification of 72% (40/55) of malignant nodules and 57% (28/49) of benign nodules. The 30 remaining nodules were equally distributed in terms of malignancy and had similar characteristics on both PET and CT images.

CONCLUSION: A probabilistic approach to pulmonary nodule characterization may help the reading physician to appropriately classify lung nodules into useful categories for the treating physician, moving away from nonstandardized reporting terms.