{"id":"847e8a97-56f9-414c-829e-2facdb19db0a","arxiv_id":"2411.16735","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A closed-form one-dimensional proton-transport model, validated against Monte Carlo for a 62 MeV beam, provides fast dose and LET profiles that can drive LET-weighted treatment-plan optimization.","lead":"This paper builds a one-dimensional analytical model of proton beams in tissue, solving the energy-loss equation in closed form to compute dose and energy deposition quickly. It compares the model with Monte Carlo simulations and shows how the fast results can be used to test treatment-planning objectives that include biological effect.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Equation (7) has the wrong sign for energy-loss transport, so the claimed closed-form solution (15) does not actually solve the stated PDE.","rationale":"The reader identified the continuous-slowing-down / no-straggling approximation as the weakest assumption, which is a legitimate physical limitation. My stress-test found a sharper, more immediate problem: the printed PDE (7) has the wrong sign for the stated physics, and the closed-form solution (15) is not a solution of it. This is not merely a stylistic point; it undermines the derivation of the central fluence and dose formulas. The claim of high-fidelity agreement with MCsquare and TOPAS may still hold, because the code presumably implements the physically correct minus-sign equation or directly uses Eq (15), which conserves the fluence along the correct decreasing-energy characteristics. I therefore do not move the verdict: CONDITIONAL remains appropriate, with the sign inconsistency explicitly corrected and the MC comparison preferably repeated or confirmed against the intended equation. The paper otherwise has independent value as a clean analytical framework with released code, and the sensitivity analysis is a useful addition, but the sign error must be fixed before the derivation can be considered sound.","tokens_in":20019,"tokens_out":11910,"duration_ms":113221,"concrete_test":"Substitute u(z,E) from Eq (15) and S(E) from Eq (13) into Eq (7) at representative points, e.g. p = 1.77, α = 2.2 × 10^−3, E = 5 MeV, z = 1 cm, and compute the residual ∂_z u + ∂_E(S u). The residual will be nonzero; repeating with ∂_z u − ∂_E(S u) gives zero. This single analytic check decides whether Eq (15) is a solution of the stated equation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central transport equation is stated as ∂_z u + ∂_E(S u) = 0 in Eq (7), with S(E) positive as defined in Eq (13). For protons losing energy along the track, the phase-space continuity equation must be ∂_z u − ∂_E(S u) = 0, because dE/dz = −S(E). This is exactly the characteristic relation used in Eq (14), E^p = E_max^p − z/α, which implies dE/dz = −E^{1−p}/(αp) = −S(E). Substituting the proposed fluence (15) into Eq (7) leaves a nonzero residual; it satisfies the minus-sign equation instead. Consequently, Eq (15) and the dose formula (16) are not derived from the PDE as printed, and an independent implementation of Eq (7) would not reproduce the Monte Carlo agreement in Fig 6. The error is fixable as a sign typo if the code actually integrates the physical minus-sign equation, but as written the mathematical backbone of the paper is internally inconsistent.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a one-dimensional analytical model for proton transport in water-like geometries under continuous-slowing-down assumptions, yielding closed-form expressions for fluence, absorbed dose, track- and dose-averaged LET, and LET-dependent biological quantities such as survival fraction, RBE, and biological dose. The model is validated against MCsquare and TOPAS for a 62 MeV pristine Bragg peak, an uncertainty analysis of the Bragg-Kleeman parameters is carried out, and the framework is applied to treatment-planning optimisation with absorbed-dose and biological-dose objectives. The code used to generate the figures is made available via Zenodo.","tokens_in":20210,"tokens_out":12743,"duration_ms":119941,"significance":"If the internal sign inconsistency is corrected and the Monte Carlo comparison is made quantitative, the closed-form formulas would provide a genuinely fast and interpretable tool for exploring LET-weighted optimisation and variable RBE in one-dimensional geometries. The availability of code, the clear treatment of parameter uncertainty, and the explicit falsifiable predictions for dose and LET are strengths. However, as printed the central transport equation is inconsistent with the claimed solution, and the validation is single-case and qualitative, so the significance cannot be fully assessed without revision.","major_comments":[{"comment":"The sign in Eq (7) is inconsistent with the characteristic solution (14) and the closed-form fluence (15). Since Eq (13) defines S(E) > 0, physical energy loss requires dE/dz = -S(E), and the phase-space continuity equation is ∂_z u - ∂_E(S u) = 0. Equation (14), E^p = E_max^p - z/α, is exactly the characteristic of the minus-sign equation, and direct substitution of Eq (15) into Eq (7) leaves a nonzero residual; the proposed solution solves the minus-sign equation instead. Equation (1) carries the same plus-sign error. If the implementation behind Fig 6 uses the minus-sign equation, Eqs (1) and (7) should be corrected and the derivation revisited; if the implementation follows the printed plus-sign equation, the agreement in Fig 6 cannot be reproduced. This must be resolved before the transport model can be accepted.","section":"§2.2, §2.5, Eqs (7), (14)–(15)"},{"comment":"The Monte Carlo validation is a single-energy, qualitative comparison. Only the 62 MeV pristine Bragg peak is shown, no quantitative agreement metric (gamma index, range difference, dose difference) or Monte Carlo statistical uncertainty is reported, and LET—which the biological sections depend on—is not validated at all. The analytical model uses α = 2.2 × 10^-3 and p = 1.77 from [19], while Table 1 lists α = 0.00246 ± 0.00025 and p = 1.75 ± 0.02 for water; this discrepancy is not explained. The claim of high fidelity needs multi-energy comparisons with explicitly stated parameter choices and a quantitative agreement metric.","section":"§2.9, Fig 6"},{"comment":"The energy interval I = [E_min, E_max] used in Eq (16) and in Figs 5–6 is never specified. At depth z > 0, characteristics with initial energies near E_min produce protons with energies below E_min, so the dose integral is truncated at E_min; depending on the chosen E_min this can remove a non-negligible low-energy contribution near the Bragg peak. The authors should report E_min and E_max and demonstrate that the dose and LET results are insensitive to E_min, for example by a convergence study in E_min.","section":"§2.5–2.6, Eq (16)"}],"minor_comments":[{"comment":"The caption states σα = 0.01µα = 0.0175 and σp = 0.01µp = 0.0000246, which swaps the two parameters; the correct values should be σα = 0.0000246 and σp = 0.0175.","section":"§4.1, Fig 11 caption"},{"comment":"The text for Examples 1 and 2 sets beam ranges equally spaced in [z_prox, z_dist + 1/4], while the Fig 17 caption states that ranges extend to z_dist + 0.15; the two choices should be reconciled.","section":"§5.1, Fig 17"},{"comment":"Equation (21) is valid only for z ≤ α E_0^p; for larger depths the argument (E_0^p - z/α)^{1/p} is not real. The domain of validity should be stated.","section":"§2.3, Eq (21)"},{"comment":"Equation (12) defines S(z) := -dE/dx as a positive quantity, while Eq (17) writes L_Δ = dE_Δ/dz without a sign; since LET is a positive energy-deposition quantity, the sign convention should be clarified.","section":"§2.3, Eq (17)"}],"recommendation":"major_revision","confidential_remarks":"The sign error in Eq (7) is the most serious issue; it appears to be typographical because Eq (15) is computed from the physical minus-sign equation. The revision should require the authors to state explicitly which equation the code solves and to correct the printed equations accordingly. The paper fits the journal's scope, the code availability is a positive feature, and there are no citation-pattern concerns."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's the short version: this is a tidy repackaging of the classic continuous-slowing-down/Bragg-Kleeman model into a one-dimensional analytic dose and LET engine, plus an LET-weighted biological-dose optimisation demo and a parameter sensitivity study. Code is released on Zenodo. It is not a new physics result—Bortfeld (1997) already gives an analytical Bragg curve and Unkelbach et al. already did LET-guided optimisation—but as a compact, readable worked framework it has real value for teaching, prototyping, and adaptive-plan exploration.\n\nThe main internal issue is a sign error in the transport equation. Eq (7), and Eq (1), is written as ∂_z u + ∂_E(S u)=0 with S>0. Since dE/dz = −S for protons slowing down, the correct equation is ∂_z u − ∂_E(S u)=0. Eq (15) solves the minus-sign equation, not the printed one; the characteristic (14) also implies dE/dz=−S. So the mathematical backbone is internally inconsistent as printed. This looks like a fixable sign typo rather than a conceptual failure, because the closed-form fluence is the standard correct solution for the physical equation. Still, Eqs (1) and (7) must be corrected, and a referee should confirm the code implements the physical sign.\n\nThe other soft spots are milder but real. Validation is one 62 MeV pristine beam, with no quantitative error metric and no comparison for the SOBP or LET profiles. The model deliberately ignores lateral scatter and straggling, so agreement on integrated depth-dose does not test what the model cannot do. The Bragg-Kleeman parameters switch between α=2.2×10⁻³, p=1.77 in Sec 2.9 and α=0.00246, p=1.75 in Table 1 without explanation; the Fig 11 caption also swaps σ_α and σ_p. Minor, but careless. The claim of high fidelity when nuclear interactions are included is overstated: the analytic model cannot reproduce the nuclear tail, and the figure is only qualitative.\n\nWhat I liked: the sensitivity analysis is honest and genuinely useful—the ±1 cm range uncertainty from α and p alone is a clinically relevant finding even in 1D. The optimisation examples are simple but clearly show how LET-weighted biological dose changes the SOBP shape.\n\nWho is it for? Medical physicists and applied mathematicians who want a fast, transparent 1D engine for prototyping or teaching, not for clinical commissioning. It deserves a serious referee after the sign and validation issues are addressed. I would send it to review, asking for a corrected equation, more than one beam energy, and at least some quantitative error metrics.","headline":"A clean, reproducible 1D dose/LET framework with a fixable but central sign error in the transport equation; worth refereeing after correction.","tokens_in":20793,"tokens_out":5036,"would_cite":false,"duration_ms":45977,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that a one-dimensional continuous-slowing-down transport model with a closed-form fluence reproduces Monte Carlo depth-dose curves for proton beams and makes LET-weighted biological-dose optimization practical.","keywords":["proton beam therapy","proton transport","Bragg-Kleeman rule","closed-form fluence","linear energy transfer","relative biological effectiveness","biological dose optimization","Monte Carlo validation"],"falsifier":"Compare the analytical dose from Equation (16) with a TOPAS or MCsquare simulation of the same 62 MeV beam with nuclear interactions turned on, measuring the 80%-to-20% distal falloff distance and the dose tail beyond the Bragg peak: if the analytical curve is visibly narrower or ends more sharply than the Monte Carlo curve beyond statistical uncertainty, the no-straggling assumption is falsified. A second decisive check is the dose-averaged LET peak, which weights the square of stopping power and therefore magnifies any difference in the low-energy tail.","tokens_in":19794,"feed_emoji":"⚛️","tokens_out":6772,"duration_ms":59637,"temperature":0.7,"pith_summary":"The paper proposes that proton transport relevant to beam therapy can be described by a one-dimensional deterministic balance equation for particle fluence, and that under the Bragg-Kleeman stopping-power law this equation has an explicit closed-form solution. From that solution the authors obtain depth-dose, track- and dose-averaged LET, cell survival fraction, spatially variable RBE, and a linear biological-dose metric, and they validate the depth-dose curve against MCsquare and TOPAS for a 62 MeV beam with reported high fidelity. Treatment planning is then posed as a weighted least-squares problem over Gaussian basis beams, with examples covering uniform target dose, organ-at-risk sparing, and LET-weighted biological dose. If the model holds, it would give near-instant evaluations of physical and biological quantities that currently require Monte Carlo simulation.","feed_headline":"Closed-form proton model matches Monte Carlo dose curves","feed_subtitle":"A 1D transport equation yields dose, LET, and biological dose for proton planning in near-instant time.","key_machinery":"The load-bearing object is the one-dimensional transport equation $\\partial u/\\partial z + \\partial(S u)/\\partial E = 0$ on a track-length coordinate $z$, with inflow spectrum $g(E)$ and the Bragg-Kleeman stopping power $S(E) = (\\alpha p)^{-1} E^{1-p}$. The equation is hyperbolic, and along characteristic curves $E^p = E_0^p - z/\\alpha$ the fluence is transported with an explicit density factor, yielding Equation (15). This closed-form fluence is the engine of the paper: Equation (16) integrates it against stopping power for dose, Equations (18)-(19) take weighted moments for track- and dose-averaged LET, Equation (25) evaluates cell survival, and Equation (29) defines the linear biological dose that makes LET-aware optimization a least-squares problem.","core_discovery":"The central claim is that the continuous-slowing-down approximation plus the Bragg-Kleeman rule $S(E) = (\\alpha p)^{-1} E^{1-p}$ turns proton transport into a hyperbolic PDE whose method-of-characteristics solution is the closed-form fluence $u(z,E) = (E^p + z/\\alpha)^{(1-p)/p}\\, g\\big((E^p + z/\\alpha)^{1/p}\\big)\\, E^{p-1}$. Dose is the stopping-power-weighted integral of this fluence (Equation 16), and LET averages are its first and second moments in stopping power (Equations 18 and 19). The authors compare the resulting depth-dose curve to MCsquare and TOPAS for a 62 MeV beam, with and without nuclear interactions, and state that the analytical model captures the depth-dose behaviour with high fidelity. The same fluence feeds a LET-dependent linear-quadratic survival model, a spatially variable RBE defined by matching the equivalent photon dose, and a linear biological dose $BD(z) = D(z)(1 + (\\lambda/c_{\\rm X-ray}) L_D(z))$ used for optimization. On this basis the paper presents the model as a complement to Monte Carlo for rapid, biologically informed treatment planning.","pith_inferences":["A natural next step beyond the paper is to concatenate one-dimensional slabs with different material parameters, which would extend the closed-form fluence to layered heterogeneous geometries while keeping the speed advantage.","Because the model omits energy-loss straggling, it likely overpredicts the sharpness of the distal falloff; convolving Equation (16) with a Gaussian straggling kernel is a testable extension that would quantify the bias in the no-straggling assumption.","The same characteristic solution generalizes to heavier ions by changing the exponent $p$, which could connect this framework to carbon-ion LET-weighted planning, though the linear-quadratic survival model would need the heavy-ion corrections the paper explicitly sets aside.","The convex linearity of the biological dose in beam weights suggests the optimization could be embedded directly in existing inverse-planning solvers, a clinical integration the paper demonstrates in one dimension but does not develop for three-dimensional pencil-beam scanning."],"forward_implications":["Depth-dose and LET profiles for any energy spectrum can be evaluated by direct quadrature of closed forms, removing Monte Carlo cost for one-dimensional water-like geometries.","Because dose is linear in beam weights, LET-weighted biological dose optimization becomes a weighted least-squares problem that can be solved rapidly, allowing fast exploration of organ-at-risk trade-offs.","The framework yields spatially resolved RBE and cell survival predictions showing RBE near 1.1 before the Bragg peak and larger values in the distal falloff, matching the clinical picture.","Uncertainty in the Bragg-Kleeman parameters $\\alpha$ and $p$ produces roughly $\\pm 1$ cm 95% confidence intervals for the Bragg peak depth at the studied energy, a level the paper flags as clinically relevant for plan robustness.","The biological-dose optimization example produces a tapered distal dose with reduced healthy-tissue dose while keeping the biological effect uniform in the target, illustrating a concrete planning consequence."],"supporting_citations":[{"why":"Makes the case that clinically useful LET visualization and LET-inclusive optimization are the priority, which motivates the paper's biological-metric framework.","marker":"[1]"},{"why":"Introduces LET-weighted reoptimization of intensity-modulated proton therapy plans, the concept the paper's biological dose and Example 3 build on.","marker":"[4]"},{"why":"Supplies measured linear-quadratic parameters ($c_{\\rm X-ray}$, $\\lambda$, $\\beta$) for AG01522 and U87 cells used for survival fraction, RBE, and biological dose.","marker":"[8]"},{"why":"Provides the analytical Bragg-curve approximation and the water parameters ($\\alpha = 2.2 \\times 10^{-3}$, $p = 1.77$) used in the validation.","marker":"[19]"},{"why":"Defines track-averaged and dose-averaged LET, whose formulas Equations (18)-(19) adopt.","marker":"[23]"},{"why":"Supplies MCsquare, one of the two Monte Carlo benchmarks for the 62 MeV depth-dose comparison.","marker":"[24]"},{"why":"Supplies TOPAS, the Geant4-based Monte Carlo benchmark used alongside MCsquare.","marker":"[25]"}],"fun_headline_variants":["Closed-form proton transport model captures dose and LET","Analytical proton fluence model agrees with Monte Carlo","Fast proton transport model yields biological dose","Exact proton transport equation enables rapid RBE assessment","Proton beam therapy modeling from a closed-form equation"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The model assumes protons slow down continuously and deterministically, with no random energy jitter (straggling), almost no sideways scattering, and very few nuclear collisions; if energy straggling meaningfully broadens the distal dose falloff at clinical energies, the closed-form match to Monte Carlo will not hold beyond the single 62 MeV water-slab benchmark.","fun_headline_variants_meta":{"raw":{"variants":["Closed-form proton transport model captures dose and LET","Analytical proton fluence model agrees with Monte Carlo","Fast proton transport model yields biological dose","Exact proton transport equation enables rapid RBE assessment","Proton beam therapy modeling from a closed-form equation"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000169,"raw_usage":{"total_tokens":1223,"prompt_tokens":860,"completion_tokens":363,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":476,"completion_tokens_details":{"reasoning_tokens":290}},"tokens_in":476,"tokens_out":363,"duration_ms":3787,"temperature":1.0,"reasoning_tokens":290,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T14:11:37.538748+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compare the analytical dose from Equation (16) with a TOPAS or MCsquare simulation of the same 62 MeV beam with nuclear interactions turned on, measuring the 80%-to-20% distal falloff distance and the dose tail beyond the Bragg peak: if the analytical curve is visibly narrower or ends more sharply than the Monte Carlo curve beyond statistical uncertainty, the no-straggling assumption is falsified. A second decisive check is the dose-averaged LET peak, which weights the square of stopping power and therefore magnifies any difference in the low-energy tail.","supporting_citations":[{"cited_title":"Treatment planning for proton therapy: what is needed in the next 10 years?","cited_arxiv_id":null,"evidence_quote":"Makes the case that clinically useful LET visualization and LET-inclusive optimization are the priority, which motivates the paper's biological-metric framework."},{"cited_title":"Reoptimization of intensity modulated proton therapy plans based on linear energy transfer","cited_arxiv_id":null,"evidence_quote":"Introduces LET-weighted reoptimization of intensity-modulated proton therapy plans, the concept the paper's biological dose and Example 3 build on."},{"cited_title":"Relative biological effectiveness variation along monoenergetic and mod- ulated Bragg peaks of a 62-MeV therapeutic proton beam: a preclinical assessment","cited_arxiv_id":null,"evidence_quote":"Supplies measured linear-quadratic parameters ($c_{\\rm X-ray}$, $\\lambda$, $\\beta$) for AG01522 and U87 cells used for survival fraction, RBE, and biological dose."},{"cited_title":"An analytical approximation of the Bragg curve for therapeutic proton beams","cited_arxiv_id":null,"evidence_quote":"Provides the analytical Bragg-curve approximation and the water parameters ($\\alpha = 2.2 \\times 10^{-3}$, $p = 1.77$) used in the validation."},{"cited_title":"A systematic review on the usage of averaged LET in radiation biology for particle therapy","cited_arxiv_id":null,"evidence_quote":"Defines track-averaged and dose-averaged LET, whose formulas Equations (18)-(19) adopt."},{"cited_title":"Fast multipurpose Monte Carlo simulation for proton therapy using multi-and many-core CPU architectures","cited_arxiv_id":null,"evidence_quote":"Supplies MCsquare, one of the two Monte Carlo benchmarks for the 62 MeV depth-dose comparison."},{"cited_title":"The TOPAS tool for particle simulation, a Monte Carlo simulation tool for physics, biology and clinical research","cited_arxiv_id":null,"evidence_quote":"Supplies TOPAS, the Geant4-based Monte Carlo benchmark used alongside MCsquare."}],"review_version":1}