Novel Asymmetrical High-Resolution and High-Sensitivity Brain Dedicated PET system: Design optimization and performance evaluation

Yuemeng Feng, Lisa Bläckberg, Arkadiusz Sitek, Hamid SabetarXiv 2026 · DOI · arXiv

TL;DR

Researchers have designed a novel, asymmetrical brain-dedicated PET system that achieves sub-millimeter spatial resolution and record-breaking sensitivity (up to 35%). By combining an elliptical geometry with advanced depth-of-interaction (DOI) and time-of-flight (TOF) capabilities, this design promises to significantly improve our ability to perform dynamic brain imaging and detect minute neurological lesions.

What Problem Does This Paper Address?

Positron Emission Tomography (PET) is a vital tool for neuroimaging, but standard whole-body scanners are "jacks of all trades and masters of none." Their large fields of view (FOV) lead to wasted sensitivity and limited spatial resolution when imaging a small organ like the brain.

While brain-dedicated PET systems exist, they often face a trade-off: they either lack coverage of the neck (missing vital carotid artery data) or use symmetric cylindrical designs that don't match the human head's anatomy, leading to "dead space" and lost photons. This paper tackles the challenge of designing a practical, high-performance scanner that maximizes solid angle coverage—essentially catching as many gamma rays as possible—while maintaining a patient-friendly design suitable for dynamic studies.

Methods

The authors used GATE v9.3, a gold-standard Monte Carlo simulation toolkit, to model a unique asymmetrical scanner.

Key Design Features: * Asymmetrical Geometry: An elliptical cylinder featuring a "neck cut-out" to allow the scanner to sit closer to the shoulders. * Front/Back Panels: Unlike open-ended cylinders, this design adds panels to "plug the holes," significantly increasing the number of detected coincidences. * Advanced Detectors: The system uses 20 mm thick LYSO:Ce crystals. To combat the "parallax error" (blurring at the edges of the FOV), they simulated up to 10 layers of Depth of Interaction (DOI). * Reconstruction: Images were produced using list-mode Maximum Likelihood Expectation Maximization (MLEM), incorporating Time-of-Flight (TOF) information ranging from 200 ps down to a cutting-edge 50 ps.

Datasets

The study relied on simulated phantoms to benchmark performance:

  • Derenzo-type Phantom: A standard tool consisting of "hot rods" of varying diameters (0.8 mm to 5.0 mm) used to measure spatial resolution.
  • XCAT Digital Brain Phantom: A voxelated model of the human brain used to evaluate contrast recovery and image quality in realistic anatomical structures.
  • Point-Source Map: A point-like 511 keV source was simulated at 16 different locations to create a comprehensive sensitivity map of the FOV.

Key Insights and Ideas

  • The Synergy of DOI and TOF: One of the most important "aha moments" was that high-speed timing (TOF) and depth-sensing (DOI) are not independent upgrades—they are synergistic. The study found that better TOF information yields little improvement if DOI resolution is poor. To see the real benefits of 50 ps timing, you need sufficient DOI layers to localize the event accurately in 3D space.
  • Asymmetry is an Advantage: By moving away from a perfect cylinder to an elliptical shape with a neck cut-out, the researchers achieved higher sensitivity in the cerebral cortex and parietal lobes (up to 35%) compared to conventional designs.
  • Practicality First: The design includes a 5x10 cm opening in the front panel specifically for Virtual Reality (VR) cameras and motion trackers, acknowledging that real-world brain studies often require patient interaction or stimulus.

Key Results

  • Sensitivity: Reached a peak of 35.04% in the parietal lobe and 23.72% at the center of the FOV. For context, the widely used HRRT system sits at roughly 4.3%.
  • Spatial Resolution: The system achieved a resolution of 0.8 mm using 10 DOI layers. Even with a more "practical" 2-layer DOI setup, resolution remained at approximately 1.0 mm in the temporal lobes.
  • Crystal Thickness: Increasing crystal thickness from 15 mm to 20 mm provided a massive boost in sensitivity without degrading resolution, provided DOI was utilized to manage the increased thickness.
  • Contrast Recovery: Using 50 ps TOF and 2 DOI layers improved contrast recovery in the brain phantom to 44.5%, compared to just 27.2% for a standard 1 DOI, 200 ps setup.

Limitations and Future Work

The authors acknowledge that these results are "optimistic" because the simulations did not include:

  1. Physical Degradation Factors: Photon scattering, attenuation, and positron range (the distance a positron travels before annihilating) were omitted to isolate the impact of the geometry and DOI/TOF.
  2. System Matrix Complexity: The reconstruction used a simplified model. Future work will involve implementing more accurate Point Spread Function (PSF) modeling to further sharpen the images.
  3. Hardware Realization: Achieving 50 ps coincidence timing resolution (CTR) and 10 DOI layers in a physical system remains a significant engineering challenge, requiring the development of specialized ASICs and SiPM arrays.

Summary updated

Abstract

Objective. This study investigates the best achievable performance of a brain-dedicated PET system with high resolution and sensitivity by evaluating different detector configurations, while maintaining a practical system design suitable for dynamic brain Imaging. Approach. Monte Carlo simulations were performed to evaluate system sensitivity and image quality under various timing resolutions (200 ps, 100 ps and 50 ps). The PET scanner geometry was optimized for human head imaging, featuring an elliptical cylindrical configuration with a neck cut-out, and front/back panels to enhance sensitivity and line of response (LOR) sampling. Detector configurations using LYSO:Ce crystals of varying thicknesses (15 mm and 20 mm) and depth of interaction (DOI) levels were simulated. Sensitivity was calculated using a point-like 511 keV back-to-back gamma source simulated at multiple locations within the field of view (FOV). Image reconstruction was conducted using list mode Maximum Likelihood Expectation Maximization (MLEM), assessing both a Derenzo-type phantom and a voxelated digital brain phantom. Main results. A location-dependent sensitivity ranging from 35.04% to 13.59% was achieved using a 20 mm thick LYSO:Ce crystal. Spatial resolution ranged from 0.8 mm to 1.5 mm within the FOV without applying resolution recovery techniques, measured using the FWHM of reconstructed hot rods. Previous results using 15 mm crystals with three DOI levels showed sensitivity between 23.42% and 15.99%, confirming the performance benefits of increased detector thickness and DOI capability. Significance. This study demonstrates the potential of a practical, brain-optimized PET system to achieve superior resolution and sensitivity for brain imaging. The findings offer valuable insights into optimal detector configurations, supporting the development of next-generation high-performance brain PET technologies.
BibTeX
@misc{feng2026novel,
  author = {Yuemeng Feng and Lisa Bläckberg and Arkadiusz Sitek and Hamid Sabet},
  title = {{Novel Asymmetrical High-Resolution and High-Sensitivity Brain Dedicated PET system: Design optimization and performance evaluation}},
  year = {2026},
  eprint = {2602.21354},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2602.21354},
}

Loading related content...

Figure