Pith. sign in

REVIEW 5 major objections 6 minor 44 references

Three-Photon Bayesian Imaging of Ortho-Positronium

T0 review · 5 major / 6 minor · reviewed 2026-08-01 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read This paper claims that a Bayesian combination of photon timing, energy, and quantum-electrodynamics decay-law correlations can localize three-photon ortho-positronium annihilations in a clinical PET scanner to about 1.6 cm, roughly twice as

desk verdict A useful Bayesian fusion of timing and energy for three-gamma PET, but the claimed QED prior is unproven and its derivation skips a Jacobian. read the letter →

arxiv 2607.27741 v1 pith:YTHLEXJN submitted 2026-07-30 physics.med-ph cs.CVphysics.comp-ph

L. Raczynski , W. Krzemien , A. Coussat , M. Bala , B.C. Hiesmayr , K. Klimaszewski , M. Obara , R. Y. Shopa This is my paper · ORCID
classification physics.med-phcs.CVphysics.comp-ph
keywords positronemissiontomographyortho-positroniumthree-photonannihilationBayesianimagereconstructiontime-of-flightPETmaximumaposterioriQEDpriorvertex
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

TRIO treats each three-photon ortho-positronium decay as a Bayesian inference problem: the vertex position is estimated by maximizing a posterior that multiplies a time-based likelihood, an energy-based likelihood, and a prior derived from the QED amplitude for three-photon decay. The paper's central claim is that this fusion works at the modest energy resolution of a current clinical PET scanner and that the QED prior is a genuinely new ingredient, not just a repackaging of data. In simulations of a long-axial-field-of-view scanner, TRIO achieves a mean position error of 1.62 cm, versus 3.05 cm for timing trilateration and about 18 cm for energy-only reconstruction. If correct, this would make three-photon imaging practical with standard radionuclides such as 18F, because no prompt photon is required.

What carries the argument

The load-bearing object is the Bayesian posterior of Eq. (9), built from three densities: a Gaussian time likelihood P(xt|x) comparing measured arrival times with times predicted from a candidate vertex; a Gaussian energy likelihood P(xE|x) comparing measured energies with energies predicted from the candidate vertex via the opening angles among the three hits; and the QED prior P(x) = κ Σ_i [(m_e c² − Ẽ_i)/(Ẽ_j Ẽ_k)]², evaluated at the vertex-predicted energies. Maximizing this product with a derivative-free simplex search explores the two-dimensional decay plane—the plane containing the three hits and the vertex, enforced by momentum conservation—for the most probable position. The QED pri

What would settle it

Run TRIO twice on the same simulated events, once with the QED prior P(x) as printed and once with P(x) uniform; if mean position error stays at ~1.62 cm in both, the QED prior is not responsible for the gain. Separately, numerically integrate κ Σ_i [(m_e c²−Ẽ_i)/(Ẽ_j Ẽ_k)]² over the feasible decay plane with the change-of-variables Jacobian; a divergent integral would show the printed prior is not a normalizable density.

Watch

Extended reading notes

Core claim

On its own terms, the discovery is that the three measured ingredients of a triple-coincidence event—hit positions, arrival times, and deposited energies—are best combined multiplicatively rather than by choosing one reconstruction. The authors write the posterior P(x|xt∩xE) ∝ P(xt|x) P(xE|x) P(x), where P(xt|x) scores how well a candidate vertex explains the measured times, P(xE|x) scores how well it explains the measured energies, and P(x) assigns prior weight according to the QED prediction for the relative probability of the photon momentum configuration implied by the candidate vertex. They report that this product sharpens the localization in directions where the individual likelihoods

Load-bearing premise

The central claim depends on the assumption that plugging the geometry-predicted photon energies into the QED energy-space distribution yields a valid probability density over decay positions, which requires a Jacobian and a finite normalization that the paper does not provide.

Editorial extensions

If this is right

  • Three-photon o-Ps imaging no longer needs a prompt photon, so standard radiotracers like 18F-FDG can in principle supply this channel.
  • Combining time and energy information in the posterior gives about a twofold accuracy gain over timing-only trilateration at current clinical energy resolution.
  • Even when energy-only reconstruction is inaccurate (~18 cm error), the energy likelihood still constrains the posterior in a direction complementary to timing.
  • The framework automatically adapts to the relative quality of timing versus energy, so it should transfer to scanners with different resolution trade-offs.
  • Per-event reconstruction is independent and parallelizable; the observed ~61 ms single-CPU time is not a fundamental limit.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A correct Bayesian transfer of the QED energy-space density to position space requires the Jacobian of the map x→Ẽ(x); the paper does not include it, so the printed prior is not strictly the QED density pulled back to the plane. Without the Jacobian (or a normalization check), the 'physics-informed prior' label is stronger than the math shown.
  • The reported improvement could in principle come entirely from the timing×energy likelihood product; the paper does not run a control with the prior removed. A uniform-prior ablation would settle whether the QED term, not just the fusion, earns the gain.
  • For real scanners, hit-position uncertainty is not negligible as assumed; the conditional-independence claim weakens with depth-of-interaction blur. A joint likelihood over positions, times, and energies would be the next step.
  • If TRIO's vertex accuracy holds on experimental data, the same event-by-event localization could serve as a veto or selection tool in fundamental o-Ps symmetry tests, a direction the paper only sketches.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 6 minor

Summary. The paper proposes TRIO, a Bayesian maximum-a-posteriori reconstruction algorithm for the vertex position of ortho-positronium three-photon decays in PET. The posterior combines three ingredients: a time-based trilateration likelihood P(x_t|x), an energy-based reconstruction likelihood P(x_E|x), and a prior P(x) derived from the QED energy-angle correlations in o-Ps decay. The authors derive closed-form likelihood models under Gaussian measurement noise, then use Nelder-Mead optimization to find the MAP estimate. In Monte Carlo simulations modelled on the Siemens Biograph Quadra, TRIO reports a mean position error of 1.62 cm, versus 3.05 cm for time-only trilateration and about 18 cm for energy-only reconstruction. The paper claims this is the first unified probabilistic framework to include a physics-informed QED prior for three-photon imaging.

Significance. If the central claim were fully supported, the work would have practical relevance for event-by-event localization of three-photon o-Ps decays in conventional TOF-PET scanners. The manuscript also has clear strengths: the likelihood derivations in Secs. 2.1 and 2.2 are mostly explicit and checkable, the simulation setup includes realistic detector geometry, DOI and resolution parameters, and comparisons against TReco and EReco baselines are provided. However, as discussed below, the contribution of the QED prior—the claimed novelty—is not established by the reported experiments, and the prior as formulated in Sec. 2.3 has a mathematically non-normalizable form and an unjustified change of variables. The paper is therefore of moderate significance at best until these load-bearing points are addressed.

major comments (5)
  1. [Sec. 2.3, Eqs. (25)–(26)] The prior transformation x -> E(x) omits the Jacobian determinant |∂E/∂x|. Even if P(E) is the QED energy-space density, the correct spatial density on the decay plane is P(x) = P(E(x)) · |∂E/∂x|. Without this factor, the printed P(x) is not the push-forward of the Ore–Powell density, so the claim that the prior 'encodes' the QED correlations is not mathematically justified. Additionally, Eq. (25) as written has no explicit normalization constant and the summand diverges like 1/E_j^2 at the simplex boundary E_j -> 0; thus P(x) is not normalizable over the domain of feasible positions. This non-integrability may push the MAP estimate toward the triangle edges in an uncontrolled way. The authors should present the correct Jacobian-transformed density and demonstrate that it is normalizable (or explicitly state that they use an unnormalized heuristic).
  2. [Sec. 5 and Fig. 2 (ablation missing)] The paper's central novelty is the physics-informed QED prior, but the reported factor-of-two improvement is explicitly attributed in Sec. 5 to 'the complementary nature of the two likelihood terms', not to the prior. No experiment is reported that runs the same posterior with P(x) = constant and compares it to the full TRIO posterior. Without this ablation, the claim that the QED prior contributes any localization improvement is unsupported; the observed gain could come entirely from fusing P(x_t|x) and P(x_E|x), an ingredient that is not claimed as new. This is a decisive gap for the abstract's 'for the first time' claim and should be addressed with a quantitative comparison in the results.
  3. [Sec. 2.3, Eq. (22) and usage of data in the prior] The prior P(x) is evaluated from the geometry of the event: the candidate position x and the measured hit positions x_i enter through the distances r_i and angles α_ij in Eqs. (15) and (22). Thus P(x) is not data-independent; it depends on the measured interaction positions of the same event. The text says the prior is 'independent of the measured times and energies', but that is not the same as being data-independent. If the hit positions are treated as fixed conditioning variables, then the posterior should be written as P(x | x_t, x_E, x_1, x_2, x_3) and the factor P(x | x_1, x_2, x_3) should be separately specified. As written, the notation P(x) and the claim of a 'pure physical constraint' conflate a data-conditioned geometry factor with a Bayesian prior, which risks double-counting the position information and weakens the statistical interpretation of Eq. (9).
  4. [Sec. 2.3, Eq. (25) normalization] Even aside from the Jacobian issue, the functional form in Eq. (25) is not self-consistent: κ is introduced as a normalization constant but never defined or computed. The expression Σ [(m_ec^2 - E_i)/(E_j E_k)]^2 is not integrable over the full simplex because of the 1/E_j^2 singularities. If the authors intend to use a truncated or regularized version, that must be stated explicitly and its effect on the reconstruction must be assessed. As printed, the discussion of a 'normalization constant κ' is not supported by the formula.
  5. [Sec. 4 and Sec. 5, statistical assessment] The reported errors are mean Euclidean distances over 500,000 simulated events per source point. The paper does not report standard errors, confidence intervals, or a statistical comparison across source positions. Given that the TRIO improvement versus TReco is roughly a factor of two and the simulation has very large N, the difference is probably significant, but the reader cannot check whether the improvement is uniform or driven by particular source locations. Reporting the distribution of errors (e.g., median, quartiles) or per-position values in tabular form would strengthen the claim, especially since the prior ablation requested above will require such statistics.
minor comments (6)
  1. [Abstract] 'for the first time' appears in the abstract and Sec. 1. The phrase is strong and should be conditional on the corrected prior derivation and ablation; otherwise it is vulnerable to the issues in Sec. 2.3.
  2. [Eq. (17)] There is a formatting typo: the formula has an open parenthesis in the text but the mathematical expression is clear. Please check the equation typesetting.
  3. [Sec. 2.3, first paragraph] There is a minor spacing typo in '˜Ein' and the sentence ending in 'decay pointx.' should be rephrased for clarity.
  4. [Fig. 4 caption] The top-left panel is labelled 'Prior distribution P(x)', but as noted in the major comments this is actually a data-conditioned geometry factor. The caption should be changed to reflect the correct interpretation.
  5. [Sec. 3] The phrase 'with the NM optimization performed using the MATLAB fminsearch function from the Optimization Toolbox with the default parameter settings' is useful, but no convergence criteria or initialization strategy for the NM search is described. Adding this would improve reproducibility.
  6. [References] Ref. [38] is described as a preprint; if this is the source of the closed-form energy solution, its status and availability should be clarified.

Circularity Check

1 steps flagged · score 2.0 of 10

No central circularity; the reported gain is a simulation result with external parameters. A self-cited preprint supplies the energy map used by the prior/likelihood, which is a minor load-bearing self-reference.

  1. self citation load bearing [Section 2, Eq. (24); Section 2.3, Eqs. (25)-(26); Ref. [38]]
    "The authors of the original work do not describe how the system of equations was solved. The detailed derivation of the exact solution for the energy-based model can be found in [38]. ... The angles in the triangle defined by the momentum vectors are given by [38]: θij = π − αij."

    The geometry-to-energy map x → E~(x), used both in the energy likelihood P(x_E|x) (Eq. 20) and in the QED prior P(x) = P(E~(x)) (Eqs. 25-26), is not derived in this paper; the paper instead defers to the authors' own unreviewed preprint [38]. Thus a load-bearing component of the prior's spatial form rests on a self-citation rather than on an independent derivation. This is not a full reduction of the reported error to its inputs, since the QED spectrum itself is traced to external sources and the headline simulation uses externally characterized scanner parameters, but it is a genuine self-reference burden in the derivation chain.

full rationale

The central numerical result (1.62 cm vs 3.05 cm vs 18 cm) is a Monte Carlo simulation output, not a fitted prediction: the energy-resolution parameter η and timing σ_t are taken from the published Quadra characterization, and the TReco and EReco baselines are external algorithms. The 'physics-informed prior' in Eq. (25) is explicitly attributed to Ore-Powell [5] and Berestetskii [37], i.e., external QED sources, not to the authors' own prior claims. No equation in the paper reduces the posterior MAP to one of its inputs by construction: P(x_t|x) is a Gaussian in time residuals, P(x_E|x) is a Gaussian in energy residuals, and P(x) is a QED density evaluated at geometrically predicted energies; none equals the measured data or the final error. The main methodological concerns identified in the text—omitted Jacobian in the change of variables P(x)=P(E~(x)), possible non-normalizability of Eq. (25), data-dependence of the 'prior' via hit positions, and lack of an ablation with P(x)=const—are correctness and validation issues, not circularity under the rubric. The one self-citation that carries part of the load is Ref. [38] for the closed-form energy solution used in Eq. (24); this is a minor self-reference, so the score is 2 rather than 0.

Assumptions & free parameters 3 free parameters · 7 assumptions · 0 invented entities

The reconstruction rests on standard QED and measurement-noise assumptions plus the paper's own construction of the spatial prior. No physical constant is fitted to the reported outcome; the scanner-noise parameters (eta, sigma_t) are inputs from the published Quadra characterization. The main structural assumptions are coplanarity, conditional independence of the two likelihoods, and neglect of interaction-position uncertainty, the latter two flagged by the authors. The prior construction in Sec. 2.3 adds a paper-specific assumption (Jacobian-free transfer of the QED density) that is not justified. TRIO introduces an algorithm and a 'physics-informed prior' (a function of standard QED), not a new physical entity, so the invented-entities list is empty.

free parameters (3)
  • eta (energy-resolution coefficient) = 2.25 (keV)^(1/2), i.e. ~10% FWHM at 511 keV
    Sets the width of the energy likelihood P(xE|x) via sigma_i = eta*sqrt(E_i) (Eq. 19, Sec. 3). Taken from the Quadra scanner's published ~10% at 511 keV energy resolution, not fitted to the TRIO outcome; the 1.62 cm figure is directly sensitive to it.
  • sigma_t (timing jitter) = CRT/(sqrt(2)*2.35) ~ 64.4 ps for CRT = 214 ps
    Sets the width of the time likelihood P(xt|x) (Eq. 13, Sec. 3). Taken from the Quadra's published 214 ps coincidence resolving time; not fitted to outcomes, but the improvement factor vs TReco depends on it.
  • kappa (prior normalization) = not given
    Eq. (25) defines the QED prior up to a constant kappa that is never evaluated or reported. As printed the density is non-normalizable on the energy simplex (diverges ~1/E^2 at the boundary), so kappa may not exist; any implementation must have made an undocumented choice here.
assumptions (7)
  • domain assumption The QED differential rate for o-Ps -> 3-gamma (Ore-Powell) is given by Eq. (25) and is normalizable with a finite kappa
    Sec. 2.3 invokes [5,37] by reference. As printed the density diverges on the energy-simplex boundary, so the normalization assumption is doubtful; the paper does not show the mapping from the cited sources to Eq. (25).
  • domain assumption Momentum conservation implies the decay vertex and the three hit positions are coplanar, justifying the 3D-to-2D reduction
    Sec. 2 opening; standard kinematics. Degrades with DOI and position-measurement errors, which the paper partially acknowledges in Sec. 5.
  • domain assumption Conditional independence of the time and energy likelihoods given x (Eq. 7)
    Sec. 2 (comment after Eq. 9) and Sec. 5. Approximate because both estimators share the same measured hit positions; the paper defends it via small crystal size and reports no systematic violation in its own simulation.
  • domain assumption Interaction-position uncertainty is negligible: x_hat_i = x_i exactly (used throughout Eqs. 12-26)
    Sec. 2.1-2.3. The paper itself flags (Sec. 5) that worse position resolution or large DOI would require revisiting; with 20 mm crystals and DOI smearing the assumption is only partially satisfied.
  • ad hoc to paper The prior construction P(x) = P_QED(E(x)) is a valid spatial prior without a Jacobian factor
    Sec. 2.3 claims a 'transform' of the QED distribution to spatial positions but omits the measure change |partial E/partial x|, and the 'prior' is data-dependent because E(x) is computed from measured hit positions.
  • domain assumption Gaussian measurement noise for times and energies (Eqs. 13, 20)
    Standard phenomenological detector model; sigma_i = eta*sqrt(E) and constant sigma_t per event. Physically motivated but not derived in the paper.
  • domain assumption A true-events-only, point-source simulation transfers to clinical performance
    Sec. 3 and Sec. 5: no scatter, no random coincidences, photoelectric-only interactions. The authors acknowledge that these dominate clinical background and that 'dedicated event selection strategies ... would be required.'

how reviews work

0 comments
Cite this review

Pith. "Pith review of Three-Photon Bayesian Imaging of Ortho-Positronium." pith.science (2026). https://pith.science/paper/YTHLEXJN

@misc{pith2026260727741,
  author       = {Pith},
  title        = {Pith review of: Three-Photon Bayesian Imaging of Ortho-Positronium},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YTHLEXJN}},
  note         = {Machine review of arXiv:2607.27741}
}
read the original abstract

PET provides functional images relying on two-photon coincidences from positron-electron annihilation. In human tissue, about 40\% of annihilations are preceded by Ps formation, of which o-Ps component partially decays into three photons, with the remainder annihilating via pick-off or spin-exchange into two photons. This three-photon channel carries additional information about the surrounding micro-environment, including the three-to-two-photon yield ratio as a potential diagnostic marker. We propose the TRIO algorithm, a novel three-photon event-by-event image reconstruction algorithm formulated as a Bayesian maximum a posteriori inference problem. TRIO unifies time-based trilateration, energy-based reconstruction and, for the first time, a physics-informed prior derived from the QED description of Ps decay within a single probabilistic framework. In contrast to positronium lifetime imaging, which requires a prompt photon and is therefore restricted to specific radionuclides, TRIO relies solely on the three photons and is fully compatible with standard radionuclides such as 18F. Monte Carlo simulation modelled after the Siemens Biograph Quadra scanner demonstrates a mean position error of 1.62~cm, improving by approximately a factor of two over the time-based trilateration (3.05 cm) and by about an order of magnitude over energy-based reconstruction alone (18 cm). More importantly, the proposed Bayesian approach is compatible with existing TOF-PET scanners that can register three-photon annihilation coincidences.

Figures

Figures reproduced from arXiv: 2607.27741 by the authors.

Figure 1
Figure 1. Example of a three-photon o-Ps annihilation. (a) Schematic 3D representation in the [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Mean position errors (ε) for three reconstruction algorithms for point sources along radial (a) and axial (b) direction. (a) (b) [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. Biases for three reconstruction algorithms for point sources along radial (a) and axial [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: An example of the position reconstruction of a single o-Ps event with additional [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 4
Figure 4. Figure 4: Three positions of the interactions of the photons with the PET scanner are located at [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references

  1. [1]

    Positron Annihilation Spectroscopy

    R. W. Siegel. “Positron Annihilation Spectroscopy”. In:Annual Review of Materials Re- search10 (1980), pp. 393–425

    1980

  2. [2]

    Three-Gamma Annihilation Imaging in Positron Emission Tomogra- phy

    K. Kacperski et al. “Three-Gamma Annihilation Imaging in Positron Emission Tomogra- phy”. In:IEEE Transactions on Medical Imaging23.4 (Apr. 2004), pp. 525–529. 14

    2004

  3. [3]

    Experimental Uses of Positronium and Potential for Biological Applica- tions

    A. Hourlier et al. “Experimental Uses of Positronium and Potential for Biological Applica- tions”. In:IEEE Transactions on Radiation and Plasma Medical Sciences8.6 (July 2024), pp. 581–594

    2024

  4. [4]

    Positronium: Review of Symmetry, Conserved Quantities and Decay for the Radiological Physicist

    M. D. Harpen. “Positronium: Review of Symmetry, Conserved Quantities and Decay for the Radiological Physicist”. In:Medical Physics31.1 (Dec. 2003), pp. 57–61

    2003

  5. [5]

    Three-Photon Annihilation of an Electron-Positron Pair

    A. Ore and J. L. Powell. “Three-Photon Annihilation of an Electron-Positron Pair”. In: Physical Review75.11 (June 1949), pp. 1696–1699

    1949

  6. [6]

    Positronium Decay in Molecular Substances

    W. Brandt et al. “Positronium Decay in Molecular Substances”. In:Physical Review120.4 (Nov. 1960), pp. 1289–1295

    1960

  7. [7]

    Investigation of Ortho-Positronium Annihilation for Porous Materials with Different Geometries and Topologies

    N. T. Trung et al. “Investigation of Ortho-Positronium Annihilation for Porous Materials with Different Geometries and Topologies”. In:Scientific Reports13.1 (2023), p. 13707

    2023

  8. [8]

    Ortho-Parapositronium Quenching by Paramagnetic Molecules and Ions

    R. A. Ferrell. “Ortho-Parapositronium Quenching by Paramagnetic Molecules and Ions”. In:Phys. Rev.110 (6 1958), pp. 1355–1357

    1958

Show all 44 references
  1. [9]

    Ortho-para spin conversion of Ps by paramagnetic O2 dissolved in organic compounds

    B. Zgardzińska et al. “Ortho-para spin conversion of Ps by paramagnetic O2 dissolved in organic compounds”. In:Nukleonika60.4 (2015), pp. 801–804

    2015

  2. [10]

    Feasibility Study of the Positronium Imaging with the J-PET Tomo- graph

    P. Moskal et al. “Feasibility Study of the Positronium Imaging with the J-PET Tomo- graph”. In:Physics in Medicine and Biology64.5 (Mar. 2019), p. 055017

    2019

  3. [11]

    Positronium Image of the Human Brain in Vivo

    P. Moskal et al. “Positronium Image of the Human Brain in Vivo”. In:Science Advances 10.37 (Sept. 2024), eadp2840

    2024

  4. [12]

    Fast High-Resolution Lifetime Image Reconstruction for Positron Lifetime Tomography

    B. Huang et al. “Fast High-Resolution Lifetime Image Reconstruction for Positron Lifetime Tomography”. In:Communications Physics8.1 (2025), p. 181

    2025

  5. [13]

    First positronium lifetime imaging with scandium-44 on a long axial field-of-view PET/CT

    L. Mercolli et al. “First positronium lifetime imaging with scandium-44 on a long axial field-of-view PET/CT”. In:Frontiers in Nuclear Medicine5 (2025)

    2025

  6. [14]

    Phantom Imaging Demonstration of Positronium Lifetime with a Long Axial Field-of-View PET/CT and 124I

    L. Mercolli et al. “Phantom Imaging Demonstration of Positronium Lifetime with a Long Axial Field-of-View PET/CT and 124I”. In:EJNMMI Physics12.1 (Aug. 2025), p. 80

    2025

  7. [15]

    Positronium Lifetime Measurement Using a Clinical PET System for Tumor Hypoxia Identification

    S. Takyu et al. “Positronium Lifetime Measurement Using a Clinical PET System for Tumor Hypoxia Identification”. In:NIMA1065 (Aug. 2024), p. 169514

    2024

  8. [16]

    High-Resolution Positronium Lifetime Tomography at Clinical Activity Levels on the PennPET Explorer

    B. Huang et al. “High-Resolution Positronium Lifetime Tomography at Clinical Activity Levels on the PennPET Explorer”. In:Journal of Nuclear Medicine66.9 (Sept. 2025), pp. 1464–1470

    2025

  9. [17]

    Positronium Lifetime Image Reconstruction for TOF PET

    J. Qi and B. Huang. “Positronium Lifetime Image Reconstruction for TOF PET”. In:IEEE Transactions on Medical Imaging41.10 (Oct. 2022), pp. 2848–2855

    2022

  10. [18]

    Positronium imaging in J-PET with an iterative activity reconstruction and a multistage fitting algorithm

    R. Y. Shopa and K. Dulski. “Positronium imaging in J-PET with an iterative activity reconstruction and a multistage fitting algorithm”. In:Bio-Algorithms and Med-Systems 19.1 (2023), pp. 54–63

    2023

  11. [19]

    SPLIT: Statistical Positronium Lifetime Image Reconstruction via Time- Thresholding

    B. Huang et al. “SPLIT: Statistical Positronium Lifetime Image Reconstruction via Time- Thresholding”. In:IEEE Transactions on Medical Imaging43.6 (June 2024), pp. 2148– 2158

    2024

  12. [20]

    High-Resolution Positronium Lifetime Tomography by the Method of Moments

    B. Huang and J. Qi. “High-Resolution Positronium Lifetime Tomography by the Method of Moments”. In:Physics in Medicine and Biology69.24 (Dec. 2024)

    2024

  13. [21]

    A Statistical Reconstruction Algorithm for Positronium Lifetime Imag- ing Using Time-of-Flight Positron Emission Tomography

    H.-H. Huang et al. “A Statistical Reconstruction Algorithm for Positronium Lifetime Imag- ing Using Time-of-Flight Positron Emission Tomography”. In:IEEE Transactions on Ra- diation and Plasma Medical Sciences9.4 (2025), pp. 478–486

    2025

  14. [22]

    A New PET Diagnostic Indicator Based on the Ratio of 3-γ/2-γPositron Annihilation

    B. Jasińska and P. Moskal. “A New PET Diagnostic Indicator Based on the Ratio of 3-γ/2-γPositron Annihilation”. In:Acta Physica Polonica B48.10 (2017), p. 1577. 15

    2017

  15. [23]

    Interaction of Positronium with Dissolved Oxygen in Liquids

    P. S. Stepanov et al. “Interaction of Positronium with Dissolved Oxygen in Liquids”. In: Physical Chemistry Chemical Physics22.9 (2020), pp. 5123–5131

    2020

  16. [24]

    Ratio of Positron Annihilation into Three Photons versus Two

    S. C. Pevovar et al. “Ratio of Positron Annihilation into Three Photons versus Two”. In: physica status solidi c4.10 (2007), pp. 3447–3450

    2007

  17. [25]

    Investigations of Positronium Formation and Destruction Using 3- γ/2-γAnnihilation-Ratio Measurements

    W. E. Kauppila et al. “Investigations of Positronium Formation and Destruction Using 3- γ/2-γAnnihilation-Ratio Measurements”. In:Physical Review Letters93.11 (Sept. 2004), p. 113401

    2004

  18. [26]

    Performance of Three-Photon PET Imaging: Monte Carlo Simulations

    K. Kacperski and N. M. Spyrou. “Performance of Three-Photon PET Imaging: Monte Carlo Simulations”. In:Physics in Medicine and Biology50.23 (Dec. 2005), pp. 5679–5695

    2005

  19. [27]

    Performance Evaluation of the uEXPLORER Total-Body PET/CT Scanner Based on NEMA NU 2-2018 with Additional Tests to Characterize PET Scanners with a Long Axial Field of View

    B. A. Spencer et al. “Performance Evaluation of the uEXPLORER Total-Body PET/CT Scanner Based on NEMA NU 2-2018 with Additional Tests to Characterize PET Scanners with a Long Axial Field of View”. In:J Nucl Med62.6 (June 2021), pp. 861–870

    2018

  20. [28]

    First Human Imaging Studies with the EXPLORER Total-Body PET Scanner*

    R. D. Badawi et al. “First Human Imaging Studies with the EXPLORER Total-Body PET Scanner*”. In:Journal of Nuclear Medicine60.3 (Mar. 2019), pp. 299–303

    2019

  21. [29]

    Extending the Clinical Capabilities of Short- and Long-Lived Positron- EmittingRadionuclidesthroughHighSensitivityPET/CT

    J. Van Sluis et al. “Extending the Clinical Capabilities of Short- and Long-Lived Positron- EmittingRadionuclidesthroughHighSensitivityPET/CT”.In:Cancer Imaging22.1(Dec. 2022), p. 69

    2022

  22. [30]

    Long Axial Field of View PET

    B. Mehadji et al. “Long Axial Field of View PET”. In:PET Clinics20.4 (Oct. 2025), pp. 499–510

    2025

  23. [31]

    LongAxialField-of-View(LAFOV)PETintheEraofMulti-Parametric Imaging and Theranostics

    N.A.Karakatsanis.“LongAxialField-of-View(LAFOV)PETintheEraofMulti-Parametric Imaging and Theranostics”. In:British Journal of Radiology(Mar. 2026)

    2026

  24. [32]

    The Impact of Long Axial Field of View (LAFOV) PET on Oncologic Imaging

    G. J. Cook et al. “The Impact of Long Axial Field of View (LAFOV) PET on Oncologic Imaging”. In:European Journal of Radiology183 (Feb. 2025), p. 111873

    2025

  25. [33]

    Three-Photon Annihilation in PET: 2D Imaging Experiments

    E. Abuelhia et al. “Three-Photon Annihilation in PET: 2D Imaging Experiments”. In: Journal of Radioanalytical and Nuclear Chemistry271.2 (Feb. 2007), pp. 489–495

    2007

  26. [34]

    Fujimoto et al.Advancing PET through Direct Imaging of Three-Photon Decay Using Pure Positron Emitters

    M. Fujimoto et al.Advancing PET through Direct Imaging of Three-Photon Decay Using Pure Positron Emitters. Dec. 2025

    2025

  27. [35]

    Trilateration-based reconstruction of ortho-positronium decays into three photons with the J-PET detector

    A. Gajos et al. “Trilateration-based reconstruction of ortho-positronium decays into three photons with the J-PET detector”. In:NIMA819 (2016), pp. 54–59

    2016

  28. [36]

    Testing CPT Symmetry in Ortho-Positronium Decays with Positronium Annihilation Tomography

    P. Moskal et al. “Testing CPT Symmetry in Ortho-Positronium Decays with Positronium Annihilation Tomography”. In:Nature Communications12.1 (Sept. 2021)

    2021

  29. [37]

    V. B. Beresteckij et al.Quantum Electrodynamics. 2. ed., reprint. Course of Theoretical Physics / L. D. Landau and E. M. Lifshitz 4. Oxford: Butterworth-Heinemann, 2008

    2008

  30. [38]

    Ortho-Positronium Three-Photon Decays: Physics Constraints and a Closed-Form Energy Method for Annihilation Vertex Reconstruction

    L. Raczyński et al. “Ortho-Positronium Three-Photon Decays: Physics Constraints and a Closed-Form Energy Method for Annihilation Vertex Reconstruction”. In:pre-print(2026)

    2026

  31. [39]

    A Simplex Method for Function Minimization

    J. A. Nelder and R. Mead. “A Simplex Method for Function Minimization”. In:The Com- puter Journal7.4 (Jan. 1965), pp. 308–313

    1965

  32. [40]

    Convergence Properties of the Nelder-Mead Simplex Method in Low Dimensions

    J. C. Lagarias et al. “Convergence Properties of the Nelder-Mead Simplex Method in Low Dimensions”. In:SIAM Journal on Optimization9.1 (Jan. 1998), pp. 112–147

    1998

  33. [41]

    Effect of Dimensionality on the Nelder-Mead Simplex Method

    L. Han and M. Neumann. “Effect of Dimensionality on the Nelder-Mead Simplex Method”. In:Optimization Methods and Software21.1 (Feb. 2006), pp. 1–16

    2006

  34. [42]

    Optimisation of the event-based TOF filtered back-projection for online imaging in total-body J-PET

    R. Shopa et al. “Optimisation of the event-based TOF filtered back-projection for online imaging in total-body J-PET”. In:Medical Image Analysis73 (2021), p. 102199

    2021

  35. [43]

    Performance Characteristics of the Biograph Vision Quadra PET/CT System with a Long Axial Field of View Using the NEMA NU 2-2018 Standard

    G. A. Prenosil et al. “Performance Characteristics of the Biograph Vision Quadra PET/CT System with a Long Axial Field of View Using the NEMA NU 2-2018 Standard”. In: Journal of Nuclear Medicine63.3 (2022), pp. 476–484. 16

    2018

  36. [44]

    Measurement of Positronium Decays at 7T with NeuroSphere PET Modules

    M. S. Allen et al. “Measurement of Positronium Decays at 7T with NeuroSphere PET Modules”. In:2025 IEEE Nuclear Science Symposium (NSS), Medical Imaging Conference (MIC) and Room Temperature Semiconductor Detector Conference (RTSD). Yokohama, Japan: IEEE, Nov. 2025. 17

    2025

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.