Positron emission tomography depends on catching pairs of 511 keV annihilation photons in coincidence, and the material that absorbs those photons sets the ceiling on how sensitive and how fast a scanner can be. A new simulation study published in EJNMMI Physics examines an unusual candidate for that role: a plastic scintillator doped with 5 wt% lead. Using Geant4 Monte Carlo simulations, Makoto Yamazaki and Masahiro Sato evaluated the commercially available scintillator EJ-256 against a conventional polyvinyltoluene (PVT) plastic, quantifying how efficiently each material absorbed monoenergetic 511 keV gamma rays and how sharply each resolved their energy.
The motivation is a long-standing trade-off in detector design. Inorganic crystals such as bismuth germanate (BGO), lutetium oxyorthosilicate (LSO) and lutetium yttrium oxyorthosilicate (LYSO) offer high stopping power and good energy resolution, but their fluorescence decay times of roughly 40 to 44 ns limit coincidence timing performance and add to system cost. Plastic scintillators, by contrast, decay in a few nanoseconds, are mechanically robust and cheap to produce, yet their low effective atomic number leaves them poor at stopping 511 keV photons. Loading heavy elements into the polymer is the proposed fix, and EJ-256, which contains 5 wt% Pb and a decay time of approximately 2.1 ns, is one such commercially available option.
The simulation setup was deliberately simple and tightly controlled. A scintillator block with a 4 × 4 mm² cross section, optically coupled to a silicon photomultiplier, was irradiated with a monoenergetic 511 keV gamma-ray pencil beam entering normal to the surface. Thickness was swept from 20 mm to 200 mm in 5 mm increments, with 100 000 primary events per run and varied random-number seeds to confirm reproducibility. The SiPM itself was not modelled as an active readout; a simplified 0.3-mm silicon layer stood in for the sensor material so that Compton scattering and backscattering at the boundary could be accounted for, while energy deposited in that layer was excluded from the photopeak analysis. Optical photon transport, photon-detection efficiency, gain, dark noise and electronic response were all left out of the model.
Fully absorbed events were defined as those depositing between 505 and 515 keV, and detection efficiency was the ratio of such events to the total number of incident photons. Absorbed-energy spectra were analysed in ROOT using Gaussian fitting to extract peak centroids and full widths at half maximum. The physics list, G4EmStandardPhysics_option4, includes photoelectric absorption, Compton scattering and pair production with optimised low-energy transport, and material parameters came from Eljen Technology datasheets, NIST photon cross-section tables and the Geant4 material-properties database.
The headline result is the emergence of a photopeak where none exists otherwise. EJ-256 produced a distinct 511 keV peak, corresponding to full-energy-absorption events arising mainly from photoelectric interactions in the Pb-doped polymer, whereas the reference PVT scintillator showed only a broad Compton continuum with no discernible peak. Detection efficiency for EJ-256 rose monotonically from 0.9% ± 0.1% at 20 mm to 2.9% ± 0.2% at 200 mm, while PVT remained negligible at 0.01% or less across all thicknesses. The authors attribute the EJ-256 trend to successive Compton scatterings followed by final photoelectric absorption, a sequence made far more likely by the presence of lead atoms.
The efficiency curve also revealed a natural stopping point. Applying a Kruskal–Wallis test to consecutive thickness groups, the authors found statistically significant differences between groups below approximately 135 mm but no significant differences beyond that range, marking the onset of a saturation plateau. Past roughly 135 mm of lead-loaded plastic, adding more material no longer meaningfully changes the balance between Compton scattering and photoelectric absorption. The authors note that the photopeak fraction of the whole spectrum remained small, but was clearly enhanced by the lead loading, meaning EJ-256 does not merely increase total interaction probability but enables a measurable fraction of complete energy-deposition events.
Energy resolution, at least in the simulated form, was stable. The photopeak FWHM stayed essentially constant at approximately 10% ± 0.3% across the entire 20 to 200 mm range, a value that numerically falls within the 8% to 12% typically reported for LYSO and BGO. The authors are careful, however, to flag what this number is not. Because optical photon transport, SiPM photon-detection efficiency, gain fluctuation, dark noise and electronic response were not simulated, the 10% figure should not be compared directly with experimentally measured detector energy resolution. It indicates only that the simulated deposited-energy spectra and intrinsic energy-transfer processes were stable across thickness, and that experimental measurements of light collection will be needed to establish real resolution.
The authors also draw explicit boundaries around how the efficiency numbers may be interpreted. The single-block pencil-beam simulation measures full-energy absorption probability in a small scintillator element, not the system sensitivity of whole-body scanners such as uEXPLORER or Biograph Vision Quadra, which achieve high sensitivity through extended axial coverage and high-stopping-power crystals. Nor is the setup comparable to J-PET systems, which use long plastic strips and exploit Compton interactions; the 505 to 515 keV full-energy window used here is a stricter criterion, so the reported efficiencies should be read as conservative, photopeak-based values rather than estimates of Compton-threshold detection performance.
Where the material genuinely shines is timing. With a decay time of roughly 2 ns, about twenty times shorter than LYSO’s approximately 40 ns, EJ-256 coupled to modern silicon photomultipliers could in principle support coincidence-timing resolutions below 200 ps, a regime that matters enormously for time-of-flight PET, where localising the annihilation point along the line of response sharpens image quality. The combination of moderate simulated energy resolution, ultrafast temporal response, high machinability and low production cost is what makes lead-loaded plastics attractive candidates for next-generation TOF detector modules and large-area geometries where reduced mass and fabrication flexibility are advantageous.
Substantial hurdles remain before any of this reaches a scanner. The light yield of EJ-256, approximately 5 200 photons per MeV, is only about 14% that of NaI(Tl), and its 425 nm emission peak may not match standard SiPM spectral sensitivity well, requiring careful optical coupling and wavelength-matching work. Polymer scintillators also degrade under radiation faster than oxide crystals, with reported light-output loss at doses above 10 kGy, and maintaining optical uniformity in thick monolithic or multilayer configurations is unresolved. Lead loading itself may attenuate scintillation photons within the plastic, reducing detected photoelectrons and degrading measured energy and timing resolution, an effect the simulation could not capture. The authors conclude that direct measurements of light output, optical attenuation length, coincidence-timing resolution and radiation durability are essential next steps before lead-loaded plastics can be considered viable alternatives to inorganic crystals in clinical and preclinical PET.
Subject of Research: Geant4 simulation of gamma-ray detection performance of a lead-loaded plastic scintillator for PET detectors
Article Title: Evaluation of a lead-loaded plastic scintillator for next-generation positron emission tomography detectors using Geant4 simulation
Article References: Yamazaki, M., & Sato, M. (2026). Evaluation of a lead-loaded plastic scintillator for next-generation positron emission tomography detectors using Geant4 simulation. EJNMMI Physics, 13(1), Article 99. https://doi.org/10.1186/s40658-026-00941-w
Image Credits: AI Generated
DOI: 10.1186/s40658-026-00941-w
Keywords: PET, scintillator, EJ-256, Geant4, Monte Carlo simulation, lead-loaded plastic, time-of-flight, detection efficiency, photopeak, silicon photomultiplier, medical imaging, Evaluation
News Source: Ophelia Keating. (October 29, 2026). Lead-Loaded Plastic Scintillator Shows Promise for Faster, Cheaper PET Detectors in Simulation. Scienmag.
