{"id":"ee759ba4-68c7-4bd6-affc-5f94d243c00c","arxiv_id":"2501.00630","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Terahertz polarization imaging of tissue phantoms and a porcine burn shows scattering-related contrast that matches a Monte Carlo Mie-scattering model, suggesting a path toward structural disease markers.","lead":"Experiments with plastic beads in gelatin and a burned pig skin sample show that terahertz light scattering patterns, measured as polarized intensity and polarization retention, differ with particle size. The results back a simulation-based argument that this contrast could help spot tumor budding or burn injuries.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Phantom data contradict the claimed Monte Carlo verification: the two smallest particle phantoms (A, B) show the opposite DOP trend from simulation in the 0.2–0.8 THz band, so the model-based burn interpretation is not established.","rationale":"The reader identifies surface roughness as the weakest assumption; that is a real and explicitly acknowledged limitation (Discussion, final paragraph). However, the phantom surfaces are cast gelatin and likely smooth, and the surface-roughness issue mostly affects the single ex vivo burn, which is already presented as preliminary. A more decisive problem is internal to the verification claim: the phantom experiment that is supposed to validate the model fails in half the cases, and the failure has the wrong sign to be explained by SNR or bandwidth alone. This matters because the burn inference is model-mediated: the paper interprets the burn DOP contrast as a decrease in scatterer size by matching the shape of the simulated small-particle DOP curves. If those curves are not reliable for the relevant sizes, the central mechanism is unsupported. I therefore agree with the reader's CONDITIONAL verdict but not with their choice of the single weakest assumption. The concrete check—re-computing the comparison with proper geometry and error bars, plus an independent Monte Carlo cross-check—would settle whether the model failure is real or an artifact of the analysis. This check should be required before the paper's 'verification' language is accepted; the central concept may still be promising.","tokens_in":11731,"tokens_out":8562,"duration_ms":91552,"concrete_test":"Recompute the experimental DOP for phantoms A and B with propagated uncertainty from the 100 averaged time-domain traces and the air-reference deconvolution, and compare only to the simulated DOP evaluated over the same 0.2–0.8 THz band and same detection geometry (20° cone at 140°) rather than the full angular integral. If A and B still deviate beyond the propagated error, the model is not valid in the small-particle regime; an independent polarized-light Monte Carlo implementation with identical inputs should be run as a cross-check. If the reversal disappears under correct geometry and error analysis, the verification claim can be restored.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim of 'Experimental and Monte Carlo verification' is undermined by the paper's own Fig. 8 comparison. Phantoms A (115 µm) and B (130 µm) are reported to show 'an opposite trend than expected' from the Monte Carlo prediction over the usable 0.2–0.8 THz band. This is not a benign bandwidth limitation: insufficient SNR or bandwidth would add noise or flatten the DOP curve, not reverse its spectral slope. A sign reversal in exactly the small-particle regime that corresponds to tumor budding and early structural lesions means the simulation has not been verified for the application it is meant to support. Section 3.4 then uses that same simulation to translate the porcine-burn DOP contrast into 'a decrease in the size of scatterers within the burn tissue.' If the small-particle prediction is wrong, that inference has no quantitative basis. The paper's conditional language ('potential') prevents a full rejection, but the abstract's statement that experimental phantom results are 'in agreement with Monte Carlo simulation results' is not supported by the A and B data. The load-bearing assumption—that the simulation correctly predicts the size-dependent DOP contrast that the burn is interpreted through—is exactly where the evidence fails.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a Monte Carlo model, based on the authors' earlier polarized-light transport code, for Mie scattering of broadband terahertz light by spherical particles embedded in an absorbing medium, and compares the simulated diffuse reflectance and degree of polarization (DOP) with experiments on gelatin phantoms containing polypropylene particles of four sizes. It then applies the same framework to interpret a single ex vivo porcine burn measurement, claiming that the observed DOP and intensity contrasts between burned and healthy skin are consistent with a burn-induced decrease in scatterer size. The central claim is that polarimetric terahertz imaging can detect structural changes due to disease processes such as tumor budding and burn injury.","tokens_in":11949,"tokens_out":2477,"duration_ms":25496,"significance":"If the claimed verification were established, the paper would provide a useful mechanistic basis for interpreting THz polarimetric contrast in terms of scatterer size and density, which is currently lacking in the THz biophotonics literature. The study is commendable for using a forward Monte Carlo model with independently measured dielectric properties, for explicitly measuring the phantom particle distributions after fabrication, and for acknowledging several limitations of the experimental system. The Mueller-matrix decomposition argument that a single polarization measurement may suffice under the assumed symmetry is also of interest. However, the experimental verification is only partial: the two smallest particle phantoms behave opposite to the model prediction in the usable band, and the burn experiment is a single sample. The significance of the work therefore rests more on its conceptual demonstration than on a validated quantitative model.","major_comments":[{"comment":"The claimed 'agreement with Monte Carlo simulation results' (Abstract) is not supported for Phantoms A and B: over the usable 0.2–0.8 THz band, the measured DOP for these two smallest-particle phantoms shows the opposite trend to the simulation. The paper attributes this to 'SNR and bandwidth limitations,' but insufficient bandwidth or SNR would add noise or flatten the DOP curve, not reverse its spectral slope. This is a load-bearing discrepancy because the small-particle regime (115–130 µm) is exactly the regime relevant to tumor budding and early structural lesions, which the burn interpretation in §3.4 relies on.","section":"§3.3, Fig. 8"},{"comment":"The interpretation of the porcine burn data is built on the same simulation that is contradicted by Phantoms A and B. The paper states that 'healthy tissue DOP drops much faster than the burn DOP from 0.2-0.8 THz' and concludes from the simulation that this indicates 'a decrease in the size of scatterers within the burn tissue compared to healthy tissue.' Since the small-particle DOP prediction is not experimentally verified, this inference has no quantitative basis. Moreover, the burn measurement is a single sample with no replicate, no error bars, and no independent measure of water content or surface roughness, both of which are acknowledged confounds in the Discussion.","section":"§3.4"},{"comment":"The diffuse-intensity comparison between simulation and experiment is not quantitative: the simulation collects photons over all backscattered angles, whereas the experimental detection cone is only 20 degrees, as the paper itself states in §3.3. Consequently, the claimed agreement for intensity is limited to a shared qualitative trend (higher intensity at lower frequencies and with larger particles). The paper should either restrict the verification claims to the DOP, or provide an angularly integrated experimental measure that can be compared with the simulation.","section":"§3.3, §2.1"},{"comment":"The conversion from surface density (particles/mm²) to volume density (mm⁻³) is described as dividing by 'the summation of the depth of field of the microscope and twice the particle diameter size,' but no explicit formula or justification is given. This conversion directly determines the simulated µs and µa through Eqs. (1)–(2), and a mis-specified conversion would change the simulation output. The authors should state the depth-of-field value and the exact formula, and examine how sensitive the DOP predictions are to this conversion.","section":"§2.2, Table 1"}],"minor_comments":[{"comment":"The abstract states that experimental phantom results 'showed contrast ... in agreement with Monte Carlo simulation results,' which is contradicted by the later admission in §3.3 that Phantoms A and B have 'an opposite trend than expected.' The abstract should be qualified to reflect the partial agreement.","section":"Abstract"},{"comment":"The DOP curves are stated to be averaged over 100 pixels, but no error bars, standard deviations, or confidence intervals are shown. Given the large spatial variability visible in the burn maps, the reader cannot assess whether the phantom DOP differences are statistically significant.","section":"§3.3, Fig. 8"},{"comment":"The Stokes parameters are defined using ensemble averages, but in practice the averages are taken over a finite number of pixels. For low-intensity pixels, the DOP computed from Eq. (5) will be biased upward by noise in I, Q, and U; the paper does not address this noise bias in the DOP estimation.","section":"§2.1, Eq. (10)"},{"comment":"The assumption that 'the surface of the phantom and the ex vivo skin is optically flat and smooth' is a strong simplification that the paper only mentions in the last paragraph of the Discussion. Given that surface roughness depolarization is a known effect (Ref. 27), this assumption should be elevated to a stated limitation in the Abstract or Results, and its potential impact on the DOP contrast should be discussed more prominently.","section":"§4, Discussion"},{"comment":"There is a typo: 'Muleler matrices' should read 'Mueller matrices.'","section":"§5, Conclusion"},{"comment":"The Monte Carlo code (Ref. 14) originates from the same research group, and the phantom inputs are measured from the same phantoms used for verification. While this is not circularity (the model is a forward calculation), the paper should explicitly acknowledge that the simulation has not been independently validated by another group, to avoid the impression of self-verification.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper's central claim of 'Experimental and Monte Carlo verification' is weakened by the opposite DOP trend for the two smallest phantoms, which are precisely the particle sizes most relevant to the burn and tumor-budding applications. The burn experiment is a single sample, and the interpretation rests on the unverified small-particle part of the simulation. I do not think this is a reject: the work is exploratory and the authors are candid about limitations, but the manuscript needs substantial revision to bring the claims in line with the evidence, and ideally additional measurements or a re-analysis that quantifies the uncertainty in the DOP comparison. The same-group origin of the simulation code is another reason to request a clearer statement of what the experiment independently establishes."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a reasonable experimental extension of the group's earlier Monte Carlo work on polarized THz scattering, with a useful new result—the claim that a full Mueller matrix may not be needed—plus phantom and burn data. But the headline 'verification' overstates what the data show. Phantoms A and B, the smallest particles, show the opposite DOP-frequency trend from the simulation in the usable band, and the paper's own explanation (insufficient bandwidth) doesn't really account for a sign reversal. That leaves the model unverified in exactly the small-particle regime that matters for tumor budding, so the porcine burn interpretation is not quantitatively grounded.\n\nWhat's new: the phantom measurements themselves, the porcine burn example, and the Lu-Chipman decomposition that suggests one linear or circular polarization measurement suffices. If that simplification holds up, it would make clinical measurements much easier. The paper is also honest in the discussion about the surface-roughness assumption and the water-loss confound in the burn.\n\nSoft spots: (1) The A and B discrepancy is more than a footnote; it should drive the conclusions. The abstract says the phantom results are 'in agreement' with simulation, but that's only true for the larger particles. (2) The simulation inputs (particle size and density) were measured from the same phantoms used for comparison. That's not circular in the fitting sense, but it means the comparison is not an independent test. (3) The burn experiment is n=1, and the DOP/intensity trade-off the simulation predicts is reversed in the burn, which the authors attribute to water loss. That's plausible but unmeasured. (4) The surface-roughness assumption is stated, not tested; if the skin surface contributes depolarization, the contrast may not be from internal Mie scatterers.\n\nWho this is for: THz biophotonics researchers and people working on polarized light scattering in tissue. It deserves a serious referee because it presents new empirical data and a falsifiable model, but it needs major revision: the A/B reversal needs either an instrument explanation or a model update, the abstract needs to be reined in, and the burn inference should be labeled as illustrative, not confirmatory.","headline":"New phantom and burn data extend the group's Monte Carlo work, but two of four phantoms show the opposite DOP trend from the model, so the verification claim is not yet solid.","tokens_in":12487,"tokens_out":2682,"would_cite":false,"duration_ms":25614,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Polarimetric terahertz imaging can detect tissue structure changes through Mie-scattering contrast in diffuse backscattered light, as shown by Monte Carlo simulation, phantom experiments, and an ex vivo porcine burn.","keywords":["terahertz imaging","polarimetry","Mie scattering","Monte Carlo simulation","degree of polarization","tissue phantoms","burn injury","diffuse reflectance"],"falsifier":"Measure the diffuse DOP and intensity of a phantom with a fixed internal particle size while deliberately varying the surface roughness from smooth to a known RMS height; if the DOP spectrum shifts as much as it does when particle diameter changes from 115 to 280 µm, then surface roughness, not internal Mie scattering, could explain the observed contrast.","tokens_in":11524,"feed_emoji":"🔬","tokens_out":3579,"duration_ms":32470,"temperature":0.7,"pith_summary":"The paper argues that polarimetric terahertz imaging can detect disease-related structural changes in tissue by measuring how diffuse backscattered light is depolarized by Mie scattering from internal particles. It builds a Monte Carlo model of polarized THz transport in absorbing tissue, extracts Mueller matrices and their Lu-Chipman decomposition, and shows that the diffuse reflectance and degree of polarization (DOP) vary systematically with scatterer size. Phantom experiments with polypropylene spheres in gelatin reproduce the simulated trends, and an ex vivo porcine burn shows DOP and intensity contrast between burned and healthy skin. The authors conclude that these signals could become diagnostic markers for conditions such as tumor budding and burn injuries.","feed_headline":"Terahertz polarimetry sees tissue structure via Mie scattering","feed_subtitle":"Diffuse intensity and degree of polarization shift with scatterer size, confirmed in phantoms and a porcine burn.","key_machinery":"The central object is a polarized-light Monte Carlo transport model of terahertz radiation in an absorbing medium containing Mie-scattering spheres, built on the Meridian Plane Polarized Light Monte Carlo code with a Mie calculator that computes scattering and absorption coefficients, the asymmetry parameter, and the scattering matrix from the measured dielectric properties of the medium and particles. From simulated spatial Stokes maps for four incident polarization states, the full Mueller matrix is constructed and decomposed with the Lu-Chipman polar decomposition, showing that each phantom acts as a uniform attenuator times a depolarizer. Because the depolarizer is characterized by a single linear degree of polarization parameter, the model predicts that one polarization measurement (linear or circular) suffices to capture the polarimetric response, and the experimental DOP follows from the spatially averaged Stokes parameters.","core_discovery":"The central claim is that the diffuse backscattered intensity and the degree of polarization of broadband terahertz light carry a quantifiable signature of the size of Mie-scattering particles embedded in a highly attenuating medium, and that this signature is observable in tissue-like phantoms and in ex vivo burned skin. The paper demonstrates through Monte Carlo simulation that as particle size increases, diffuse reflectance increases and the first minimum in the frequency-dependent DOP shifts to lower frequencies, and it verifies the intensity and DOP trends experimentally for four gelatin phantoms containing polypropylene particles of 115–280 µm diameter. It further shows that burned porcine skin exhibits higher DOP and higher diffuse intensity than healthy skin, which the authors interpret as a decrease in average scatterer size from destruction of hair follicles and sweat glands, together with reduced water absorption.","pith_inferences":["The surface-flatness assumption is likely the main obstacle to clinical translation: real skin is rough, and the paper's own earlier speckle work shows surface roughness depolarizes THz light, so separating surface and bulk depolarization will be necessary before the Mie-scattering interpretation can be applied in vivo.","The burn data show DOP and intensity both increasing, whereas the simulation ties larger DOP to smaller scatterers with lower reflectance; the authors' explanation is reduced water absorption in the burn, which suggests that joint estimation of absorption and scattering is required to interpret clinical signals unambiguously.","The same approach could be tested on other conditions that change sub-millimeter tissue structure, such as fibrosis, edema, or tumor margins, provided the suspected scatterer sizes fall in the Mie regime for the THz bandwidth used.","A direct experimental falsification would be to fabricate phantoms with fixed particle size and deliberately varied surface roughness; if DOP changes as much from roughness as from a 115-to-280 µm particle size change, the diagnostic interpretation would need revision."],"forward_implications":["If the central claim holds, a single measurement of degree of polarization at a diffuse backscattering angle can recover scatterer size and concentration information from the spectral location and depth of the DOP minimum.","The Lu-Chipman decomposition implies that for spherical scatterers in an absorbing medium, full Mueller matrix characterization is redundant; one linear or circular polarization measurement suffices.","DOP is far less sensitive to system misalignment than absolute diffuse intensity, so it is the more reliable candidate diagnostic metric.","The porcine burn result suggests the technique can distinguish burned from healthy tissue ex vivo, with the DOP contrast attributed to destruction of large skin structures.","Extending bandwidth from roughly 1.5 THz to 8 THz would allow the DOP minimum to be resolved for smaller particle sizes, improving sensitivity to early disease changes."],"supporting_citations":[{"why":"Supplies the Monte Carlo model of Mie scattering of polarized THz light used for all simulations in the paper.","marker":"[14]"},{"why":"Supplies the Meridian Plane Polarized Light Monte Carlo code that the simulation is based on.","marker":"[39]"},{"why":"Supplies the Mie scattering coefficient and absorption formulas for spheres embedded in an absorbing host medium.","marker":"[40]"},{"why":"Supplies the Lu-Chipman polar decomposition used to interpret the simulated Mueller matrices.","marker":"[41]"},{"why":"Documents previous work on surface-roughness depolarization that motivates the paper's flat-surface assumption.","marker":"[27]"},{"why":"Demonstrates a broader-bandwidth THz system that the paper proposes for resolving DOP minima for smaller particles.","marker":"[42]"}],"fun_headline_variants":["Terahertz polarimetry maps scatterer size via Mie scattering","Polarized terahertz light reads tissue structure from Mie scattering","Scatterer size seen in terahertz diffuse intensity and polarization","Terahertz Mie scattering reveals burn damage in porcine skin","Diffuse terahertz light distinguishes tissue scatterer sizes in phantoms and burns"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The interpretation of measured DOP and intensity contrast as due to internal Mie scattering assumes that the phantom and ex vivo skin surfaces are optically flat and smooth, so that depolarization from surface roughness is negligible.","fun_headline_variants_meta":{"raw":{"variants":["Terahertz polarimetry maps scatterer size via Mie scattering","Polarized terahertz light reads tissue structure from Mie scattering","Scatterer size seen in terahertz diffuse intensity and polarization","Terahertz Mie scattering reveals burn damage in porcine skin","Diffuse terahertz light distinguishes tissue scatterer sizes in phantoms and burns"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000632,"raw_usage":{"total_tokens":2955,"prompt_tokens":1020,"completion_tokens":1935,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":636,"completion_tokens_details":{"reasoning_tokens":1839}},"tokens_in":636,"tokens_out":1935,"duration_ms":12279,"temperature":1.0,"reasoning_tokens":1839,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T22:45:50.180086+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the diffuse DOP and intensity of a phantom with a fixed internal particle size while deliberately varying the surface roughness from smooth to a known RMS height; if the DOP spectrum shifts as much as it does when particle diameter changes from 115 to 280 µm, then surface roughness, not internal Mie scattering, could explain the observed contrast.","supporting_citations":[{"cited_title":"Lu and R","cited_arxiv_id":null,"evidence_quote":"Supplies the Lu-Chipman polar decomposition used to interpret the simulated Mueller matrices."},{"cited_title":"Xu and M","cited_arxiv_id":null,"evidence_quote":"Supplies the Monte Carlo model of Mie scattering of polarized THz light used for all simulations in the paper."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Meridian Plane Polarized Light Monte Carlo code that the simulation is based on."},{"cited_title":"Yang, B.-C","cited_arxiv_id":null,"evidence_quote":"Supplies the Mie scattering coefficient and absorption formulas for spheres embedded in an absorbing host medium."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Documents previous work on surface-roughness depolarization that motivates the paper's flat-surface assumption."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates a broader-bandwidth THz system that the paper proposes for resolving DOP minima for smaller particles."}],"review_version":1}