Interobserver variability among measurements of the maximum and mean standardized uptake values on (18)F-FDG PET/CT and measurements of tumor size on diagnostic CT in patients with pulmonary tumors.

BACKGROUND: (18)F-fluoro-2-deoxyglucose positron emission tomography ((18)F-FDG PET) imaging has been shown to be an accurate method for diagnosing pulmonary lesions, and the standardized uptake value (SUV) has been shown to be useful in differentiating benign from malignant lesions.

PURPOSE: To survey the interobserver variability of SUV(max) and SUV(mean) measurements on (18)F-FDG PET/CT scans and compare them with tumor size measurements on diagnostic CT scans in the same group of patients with focal pulmonary lesions.

MATERIAL AND METHODS: Forty-three pulmonary nodules were measured on both (18)F-FDG PET/CT and diagnostic chest CT examinations. Four independent readers measured the SUV(max) and SUV(mean) of the (18)F-FDG PET images, and the unidimensional nodule size of the diagnostic CT scans (UD(CT)) in all nodules. The region of interest (ROI) for the SUV measurements was drawn manually around each tumor on all consecutive slices that contained the nodule. The interobserver reliability and variability, represented by the intraclass correlation coefficient (ICC) and coefficient of variation (COV), respectively, were compared among the three parameters. The correlation between the SUV(max) and SUV(mean) was also analyzed.

RESULTS: There was 100% agreement in the SUV(max) measurements among the 4 readers in the 43 pulmonary tumors. The ICCs for the SUV(max), SUV(mean), and UD(CT) by the four readers were 1.00, 0.97, and 0.97, respectively. The root-mean-square values of the COVs for the SUV(max), SUV(mean), and UD(CT) by the four readers were 0%, 13.56%, and 11.03%, respectively. There was a high correlation observed between the SUV(max) and SUV(mean) (Pearson's r=0.958; P <0.01).

CONCLUSION: This study has shown that the SUV(max) of lung nodules can be calculated without any interobserver variation. These findings indicate that SUV(max) is a more valuable parameter than the SUV(mean) or UD(CT) for the evaluation of therapeutic effects of chemotherapy or radiation therapy on serial studies.