Abstract
515
Objectives Accuracy of CT-guided fine needle aspiration (FNA) is subject to sampling errors. Current study assesses whether FDG-PET/CT can improve the accuracy and decrease the false negative (FN) rate of CT-guided FNA of lung lesions.
Methods Data of 311 consecutive patients with lung nodules who underwent FDG-PET/CT and CT-guided FNA within a month were retrospectively investigated. A dedicated software was developed in-house to register corresponding CT images of the PET/CT with those used to guide FNA (CT-FNA). The quality of registration was rated on a scale of 1(excellent) to 5 (mis-registration). Only cases scored 1-2 were further evaluated. The software also provided the Standard Uptake Value (SUV) at the location of the tip of the aspirating needle and in the whole lung lesion, as well as the distance (mm) between the needle tip and the area with the highest SUV within the lesion. The mean distance from the needle tip to highest SUV focus and the mean difference in SUVmax between the whole lesion and that at the needle tip were calculated and compared between true positive (TP) and FN FNA results using the unpaired t-test.
Results 267 patients (86%) with high registration score were evaluated, including 185 TP (69%) and 39 FN (15 %) FNA results. The distance between the location of the needle tip and the SUVmax point in the lesion was significantly greater in the FN group (18.35 ± 27.4mm) as compared to the TP group (7 ± 14.3, p<0.001). The difference between the SUV at the location of the tip and that of the whole lesion was significantly higher in the FN group (6.7 ± 15.6) as compared to the TP group (2.0 ± 4.1, p<0.001).
Conclusions Present results demonstrate a relationship between the degree of metabolism at the site of tissue sampling and the accuracy of FNA results. FDG-PET/CT guided FNA using metabolic information provided by FDG-PET in addition to CT can improve the accuracy of histological examinations, decrease the FN rate and thus increase the probability of achieving a definitive diagnosis.