ABSTRACT: Delayed imaging that coincides with the highest uptake of fluorine-18 fluorodeoxyglucose (FDG) by tumour may be advantageous in oncological positron emission tomography (PET), where delineation of metastasis from normal tissue background is important. In order to identify the better imaging protocol for tumour detection, whole-body FDG-PET images acquired at 1 h and 2 h after injection were evaluated in 22 subjects, with a post-injection transmission scan at 90 min for attenuation correction. After visual interpretation, tumour uptake [tumour standardised uptake ratio (SUR)], normal tissue uptake (normal SUR) and tumour to background contrast (tumour SUR/normal tissue SUR) were evaluated in the images acquired at 1 h and at 2 h. Most malignant lesions, including primary lung cancer, metastatic mediastinal lymph nodes and lymphoma lesions, showed higher FDG uptake at 2 h than at 1 h. By contrast, benign lesions, with the exception of sarcoidosis, showed lower uptake of FDG at 2 h than at 1 h. Among normal tissues, the kidney, liver, mediastinum, lung, upper abdomen and left abdomen showed significant falls in FDG uptake from 1 h to 2 h. The lower abdomen, right abdomen and muscles (shoulder and thigh) showed no significant changes. Consequently, malignant lesions of the lung, mediastinum and upper abdomen showed significant increases in tumour to background contrast from 1 to 2 h. Three lesions (two lung cancers and a malignant lymphoma) that were equivocal on 1-h images became evident on 2-h images, changing the results of interpretation. All other malignant lesions were detected on 1-h images, but were clearer, with higher contrast, on 2-h images. Lesion-based sensitivity was improved from 92% (49/53) to 98% (52/53), and patient-based sensitivity from 78% (14/18) to 94% (17/18). It is concluded that delayed whole-body FDG-PET imaging is a better and more reliable imaging protocol for tumour detection.
European journal of nuclear medicine and molecular imaging 01/2001; 28(6):696-703. · 4.99 Impact Factor