Objective: The objective of this work was to assess the Standardized Uptake Value (SUV) in the differential diagnosis of radiologically indeterminate lung lesions by means of ROC curves.
Material and method: Forty seven patients were studied by Positron Emission Tomography with 18-fluorine-2-desoxy-D-glucose (FDG PET) analyzing the value of maximum SUV. The patients were classified into three groups. Group 1 = patients without previous neoplasia (WPN) + patients with previous neoplasia (PN). Group 2 = WPN. Group 3 = PN.
Results: The ROC curves showed a high diagnostic accuracy in the three groups, with area under the curve (AUC) values of 0.96, 0.98 and 0.91 respectively. The typical error was 0.03, 0.02 and 0.08. The maximum SUV cutoffs with the best diagnostic accuracy for the three groups were: 2.6; 3 and 2.4, with an accuracy (A) of 93.6%, 97% and 92.3%, respectively. Analyzing all the patients globally (group 1), we obtained one false positive result in a patient with hamartoma (max SUV = 2.8) and two false negative results in one patient with lung metastases from malignant fibrohistiocytoma (max SUV = 0.7) and in another patient with lung metastases from unknown origin adenocarcinoma (max SUV = 1.9).
Conclusions: FDG PET permits differentiation with a very high diagnostic accuracy of benign and malignant lung lesions using the maximum SUV. The differences observed between the different groups are due to the different disease prevalence, obtaining a lower negative predictive value of max SUV in patients with previous neoplasia.