Implementation and evaluation of a dynamic whole-body PET/CT protocol

Objectives Current whole-body (WB) PET protocols do not capture tracer kinetic data and are overlooking potentially valuable information. We describe an alternative protocol that measures dynamic data over the usual WB range, opening up an additional dimension of complementary kinetic information.

Methods The following dynamic WB protocol was implemented on a clinical PET/CT system (Discovery VCT) and its potential was investigated with a series of oncology patient studies (n=5). Low-dose CT from head to thighs; FDG administered with patient’s heart centered in the PET field-of-view; 0-6 min single bed-position dynamic PET over the heart; six sequential head-to-thigh multi-bed PET scans (45 sec / bed). A manually defined ROI in the left atrium was applied to the initial single bed-position and the multi-bed data for non-invasive input function determination.

Results Patient motion was minimal (mean shift 0.53 ± 0.45 mm, mean angle 0.16 ± 0.17 degrees), estimated using software registration of non-attenuation corrected images. Time-activity-curves (TACs) showing regional tracer kinetics were available throughout the WB volume. Normal organ (liver, lung, kidney, mediastinum, muscle) FDG TACs decreased between 6-45 minutes, with the exception of brain, salivary glands and myocardium (variable). Tumor TACs (n=4) consistently increased over this time interval. These temporal characteristics can be used to aid image assessment and may have advantages in terms of the partial volume effect. Principal component analysis produced a parametric image weighted towards decreasing TACs and a separate image that reflected tissues with increasing FDG uptake, including tumors. Absolute quantification of metabolism (Patlak) was possible for tumors at all locations, potentially offering advantages over SUV.

Conclusions Dynamic WB PET/CT provides kinetic information that is currently not available with the conventional WB protocol. These additional data facilitate an approach to tumor detection and characterization that benefits from an integration of both spatial and temporal information