PennPET Explorer: Design and Preliminary Performance of a Whole-Body Imager

Abstract

We report on the development of the PennPET Explorer whole-body imager. Methods: The PennPET Explorer is a multiring system designed with a long axial field of view. The imager is scalable and comprises multiple 22.9-cm-long ring segments, each with 18 detector modules based on a commercial digital silicon photomultiplier. A prototype 3-segment imager has been completed and tested with an active 64-cm axial field of view. Results: The instrument design is described, and its physical performance measurements are presented. These include sensitivity of 55 kcps/MBq, spatial resolution of 4.0 mm, energy resolution of 12%, timing resolution of 256 ps, and a noise-equivalent count rate above 1,000 kcps beyond 30 kBq/mL. After an evaluation of lesion torso phantoms to characterize quantitative accuracy, human studies were performed on healthy volunteers. Conclusion: The physical performance measurements validated the system design and led to high-quality human studies.

Keywords: NEMA performance; PET; whole-body imager.

FIGURE 1.

PennPET Explorer in its prototype configuration with 3 ring segments, housed in dry, cool enclosure. View of back of gantry shows modular detector and electronic bays. Also shown is couch with flat pallet installed for human studies.

FIGURE 2.

NEMA NU-2 count-rate performance with 70-cm line source inside 20-cm-diameter polyethylene scatter cylinder. Count-rate results were acquired up to 40 kBq/mL, although clinical ¹⁸F-FDG studies are typically performed with activity concentrations of less than 5 kBq/mL.

FIGURE 3.

(A) NEMA image-quality phantom shown with standard spheres and half-sized spheres. Body activity concentration is 2 kBq/mL, sphere contrast is 9.7:1, and scan duration is 7.5 min. (B) CRCs as per NEMA guidelines for both standard and half-sized spheres. (C) Background variability as per NEMA guidelines for half-sized spheres.

FIGURE 4.

(A) Clinical Trials Network torso phantom with activity concentration of 5.9 kBq/mL and lesion contrast of 4.2:1. (B) CRC of representative lesions as function of scan duration. (C) SD of CRC, determined from replicates of data.

FIGURE 5.

(A) Representative sagittal (left), axial (middle), and coronal (right) views of 20-min PennPET Explorer image of subject 1 at 1.5 h after injection of 555-MBq dose of ¹⁸F-FDG. All are 2-mm sections. (B) PennPET Explorer image, subsampled (⅛ data) to represent 2.5-min scan. (C) Clinical scan acquired with Philips Ingenuity TF PET/CT scanner at 1 h after injection, using clinical protocol with 10 bed positions for total of 20 min. These data were reconstructed off-line with same reconstruction method as for PennPET Explorer data.

FIGURE 6.

Sagittal (left), axial (middle), and coronal (right) views of PennPET Explorer images of subject 2 positioned with head near edge of AFOV (A) and at center of AFOV (B). These scans were acquired starting at 1.5 h after injection of ¹⁸F-FDG for 10 min each. All are 1-mm sections.

FIGURE 7.

Dynamic ¹⁸F-FDG study of subject 7 acquired for 60 min after 555-MBq dose of ¹⁸F-FDG. Representative maximum-intensity projections are shown from 0 to 60 min after injection.