Accuracy of F-18 fluorodeoxyglucose positron emission tomography for the evaluation of malignancy in patients presenting with new lung abnormalities: a retrospective review.

STUDY OBJECTIVE To evaluate the accuracy of positron emission tomography (PET) in determining the presence of malignancy in patients presenting with new lung findings, either as an incidental finding or after treatment of a primary carcinoma. DESIGN A retrospective review of the PET database of our hospital from April 29, 1997, to March 20, 1999, identified 196 patients referred for the evaluation of new lung findings, either as an incidental finding or following definitive treatment of a primary carcinoma. The diagnosis of either malignancy or a benign condition was established in 71 patients. This was determined by either histopathology from biopsy, or by subsequent imaging demonstrating disease progression, resolution, or stability of the initial lung findings. RESULTS In patients presenting with new lung findings without a history of carcinoma (n = 37), the sensitivity and specificity of PET was 95% and 82%, respectively. In this population, the negative predictive value was 93% and the positive predictive value was 86%. PET was less sensitive and specific for evaluating metastatic or recurrent disease in patients previously treated for carcinoma. In patients presenting with a previously treated primary lung cancer (n = 13), the sensitivity of PET was 70%, with a specificity of 67%. The negative predictive value was only 40% and the positive predictive value was 88% in this subset of patients. In patients with an extrapulmonary primary carcinoma presenting with new lung nodules (n = 21), the sensitivity and specificity of PET was 92% and 63%, respectively. In this population, the negative predictive value was 83% while the positive predictive value was 80%. Of the 71 total cases for which follow-up data were available, there were 5 false-negative cases and 7 false-positive cases, for an overall sensitivity of 88%, specificity of 75%, negative predictive value of 81%, and positive predictive value of 84%. CONCLUSIONS The sensitivity of PET is highest for the evaluation of new malignancy in patients without a known primary carcinoma. PET is less sensitive for evaluating metastatic or recurrent disease.