Skip to main content
. Author manuscript; available in PMC: 2021 Mar 22.
Published in final edited form as: Eur J Nucl Med Mol Imaging. 2017 Sep 11;45(1):4–11. doi: 10.1007/s00259-017-3818-x

Fig. 3.

Fig. 3

a. ROC analysis performed for surgical patients (n = 59) to assess the ability of pre-scan PSA in distinguishing between 18F–DCFBC PET/ CT positive and negative result, with an AUC (area under the operating characteristic curve) of 76.4%. b. The optimal cut-off pre-scan PSA, which maximizes the difference between the true-positive rate (TPR) and the false-positive-rate (FPR), was 0.78 ng/mL, i.e., at pre-scan PSA of 0.78 ng/mL, TPR was 87.5% (28/32) and FPR was 33.3% (9/27)