Performance evaluation of the PennPET explorer with expanded axial coverage

Abstract

Objective.This work evaluated the updated PennPET Explorer total-body (TB) PET scanner, which was extended to 6 rings with updated readout firmware to achieve a 142 cm axial field of view (AFOV) without 7.6 cm inter-ring axial gaps.Approach.National Electrical Manufacturers Association (NEMA) NU 2-2018 measurements were performed with modifications including longer phantoms for sensitivity and count-rate measurements and additional positions for spatial resolution and image quality. A long uniform phantom and the clinical trials network (CTN) phantom were also used.Main results.The total sensitivity increased to 140 kcps MBq^-1for a 70 cm line, a gain of 1.8x compared to the same system with axial gaps; an additional 47% increase in total counts was observed with a 142 cm line at the same activity per cm. The noise equivalent count rate (NECR) increased by 1.8x without axial gaps. The peak NECR is 1550 kcps at 25 kBq cc^-1for a 140 cm phantom; due to increased randoms, the NECR is lower than with a 70 cm phantom, for which NECR is 2156 kcps cc^-1at 25 kBq cc^-1and continues increasing. The time-of-flight resolution is 250 ps, increasing by <10 ps at the highest activity. The axial spatial resolution degrades by 0.6 mm near the center of the AFOV, compared to 4 mm resolution near the end. The NEMA image quality phantom showed consistent contrast recovery throughout the AFOV. A long uniform phantom demonstrated axial uniformity of uptake and noise, and the CTN phantom demonstrated quantitative accuracy for both¹⁸F and⁸⁹Zr.Significance. The performance evaluation of the updated PennPET Explorer demonstrates significant gains compared to conventional scanners and shows where the current NEMA standard needs to be updated for TB-PET systems. The comparisons of systems with and without inter-ring gaps demonstrate the performance trade-offs of a more cost-effective TB-PET system with incomplete detector coverage.

Keywords: NEMA performance; axial detector gaps; total-body PET.

Figure 1.

Representative maximum intensity projection (MIP) images and axial count rate (trues + scatter) profiles on the PennPET Explorer with 142 cm AFOV at three time points in a dynamic [¹⁸F]-fluorodeoxyglucose (FDG) study (349 MBq injection) for a male patient of average height and weight (174 cm and 88 kg).

Figure 2.

(a) Data flow of a single detector module in the PennPET Explorer. (b) Schematic of multi-ring data acquisition.

Figure 3.

(a) Axial sensitivity profiles measured with a 142 cm line source on the 6-ring PennPET Explorer with and without axial gaps. 2 mm coronal image slice of the reconstructed 10 × 190 cm pipe phantom uniformly filled with 120 MBq (8.4 kBq cc⁻¹) of ¹⁸F and imaged in a single bed position for 15 min on the 6-ring scanner (b) with and (c) without axial gaps. (d) Axial uniformity (SUV) and (e) axial noise (image roughness) profile of the reconstructed pipe phantoms.

Figure 4.

Count rate performance measured with the 20 × 70 cm NEMA phantom on the PennPET Explorer in its current configuration, 6 rings without axial gaps. (a) Trues, scatter, randoms and noise-equivalent count (NEC) rates, (b) the corresponding scatter fraction and randoms fraction, and (c) accuracy as a function of activity concentration.

Figure 5.

Count rate performance measured with single 20 × 70 cm and double (20 × 140 cm) NEMA phantoms on the PennPET Explorer without axial gaps. (a) Trues rates, (b) NEC rates, (c) scatter fractions, and (d) randoms fractions as a function of activity concentration. Schematic depicting the placement of (e) the 20 × 70 cm and (f) the 20 × 140 cm count rate phantoms.

Figure 6.

Count rate performance measured with the double (20 × 140 cm) NEMA phantom on the PennPET Explorer with and without axial gaps. (a) Trues rates, (b) NEC rates, (c) scatter fractions, and (d) randoms fractions as a function of activity concentration.

Figure 7.

(a) Time-of-flight resolution measured with single 20 × 70 cm and double (20 × 140 cm) NEMA phantoms on the PennPET Explorer without axial gaps. (b) Timing histograms of single rings from daily QC measurements performed with a point source in the center of each ring. The timing resolution per ring averages 240 ps.

Figure 8.

Spatial resolution performance measured on the PennPET Explorer with 6 rings without axial gaps. Schematic depicting (a) the transversal and (b) the axial locations of the measurements (c) Radial and tangential resolution as a function of radial position. (d) Axial resolution as a function of axial position.

Figure 9.

(a) Transaxial image slice (2 mm thick) of the standard NEMA image quality phantom placed at the center of the AFOV and imaged for 30 min on the PennPET Explorer without axial gaps. (b) Schematic of 6 detector rings overlaid with maximum-intensity projection (MIP) of the IQ phantom depicting the center and off-center axial locations where the phantom was imaged. (c) Contrast recovery coefficient (CRC) and (d) background variability measured at two axial positions for 30 min and 3 min scans on the 6-ring system without axial gaps.

Figure 10.

Maximum intensity projection (MIP) images of the Clinical Trials Network (CTN) phantom filled with (a) ⁸⁹Zr and (b) ¹⁸F. The phantoms were placed at the center of AFOV and scanned for 60 min and 11.3 min respectively for equal activity-scan duration. Contrast recovery measured with the (c) ⁸⁹Zr and (d) ¹⁸F CTN phantom on the 6-ring PennPET Explorer without axial gaps.