TY - JOUR T1 - PET/CT as a predictor of malignancy in indeterminate lung nodules JF - Journal of Nuclear Medicine JO - J Nucl Med SP - 1381 LP - 1381 VL - 56 IS - supplement 3 AU - Jonathan Chung AU - David Lynch AU - James Hamill AU - Darrell Burckhardt Y1 - 2015/05/01 UR - http://jnm.snmjournals.org/content/56/supplement_3/1381.abstract N2 - 1381 Objectives 18F-FDG PET/CT has the potential to predict lung nodule malignancy. We ask which quantitative measures have the highest predictive power in the case of indeterminate pulmonary nodules (IPN).Methods 89 patients had a history of COPD and IPN were detected on CT (107 nodules total). Whole-body PET/CT with 15 mCi FDG injection was performed at 1 hour post injection (p.i.), then six minutes of list-mode PET data were acquired (LM PET). LM PET was again acquired at 2 hours p.i. Malignancy or benignity was determined by nodule biopsy or by two years of disease-free progression. Using receiver-operator characteristics analysis (ROC) we considered the following measurements as potential predictors of malignancy: nodule volume (V) on CT; SUV-max based on the first time point; SUV-mean; the relative and absolute SUV change between time points, d-SUV and d-SUV/SUV. In each case the predictive power was reflected by area under the ROC curve (AUC). We performed reconstructions with 1/3 and 1/10 the counts, corresponding to lower injected dose, then re-calculated AUC.Results 29 nodules were malignant and 78 were benign. When the full injected dose was used, AUC was 0.72, 0.82, 0.74 and 0.68 for V, SUV-max, d-SUV and d-SUV/SUV, respectively. SUV-mean analysis and count reduction are in progress.Conclusions All measures were valid predictors of malignancy. AUC based on SUV was higher than AUC based on nodule volume or change in SUV between 1 and 2 hours. ER -