{"id":"ce5210c4-480f-4902-ad71-c60680f30632","arxiv_id":"2511.13310","paper_version":2,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":5.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"PyPeT is an open Python framework for unified CTP and MRP processing that outputs CBF, CBV, MTT, TTP, and Tmax maps and reports mean SSIM of approximately 0.8 against three FDA-approved commercial tools.","lead":"PyPeT is an open-source Python tool that takes raw four-dimensional CT and MR perfusion scans and automatically generates standard maps of cerebral blood flow, blood volume, and timing parameters. It provides a free, modifiable alternative to costly closed commercial software for stroke and cerebrovascular research.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.3","headline":"SSIM ~0.8 with commercial tools does not establish numerical accuracy of CBF/CBV/Tmax values","rationale":"The reader's weakest assumption correctly isolates the validation gap. Because the manuscript positions PyPeT as producing quantitative maps and the only external check is SSIM against other black-box tools, the numerical fidelity of the output parameters remains the single most load-bearing untested condition. No other internal inconsistency or missing prerequisite was identified from the available description.","tokens_in":1774,"tokens_out":389,"duration_ms":31827,"concrete_test":"On the same 10-patient CTP/MRP cohort used in the paper, compute ROI-averaged CBF, CBV and Tmax inside manually delineated gray-matter, white-matter and infarct regions for PyPeT and each commercial tool; report Pearson correlation, bias and 95% limits of agreement. If any correlation falls below 0.85 or bias exceeds 15%, the quantitative-accuracy claim is not supported by the current evidence.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that PyPeT produces quantitative CBF, CBV, MTT, TTP and Tmax maps whose mean SSIM with three FDA-approved commercial outputs is ~0.8. For this to support the tool as a research or clinical alternative, the numerical hemodynamic values (not merely their spatial patterns) must be reliable. SSIM is insensitive to global scaling, offset, or absolute calibration differences; commercial packages are known to differ by 20-40% in ROI-averaged CBF even on identical data. The paper reports only aggregate SSIM and visual comparison; it does not supply ROI-level Pearson r, mean absolute percentage error, or Bland-Altman limits for the actual parameter values, nor any phantom or simulated ground-truth test. Without these, structural similarity cannot be taken as evidence that the derived quantities are quantitatively correct.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.3","summary":"The manuscript introduces PyPeT, an open-source Python tool for processing raw 4D CT and MR perfusion data to generate quantitative maps of cerebral blood flow (CBF), cerebral blood volume (CBV), mean transit time (MTT), time-to-peak (TTP), and time-to-maximum (Tmax). It emphasizes modularity, low computational cost, inline documentation, and a debug visualization mode. Validation consists of visual inspection plus quantitative comparison against outputs from three FDA-approved commercial perfusion packages, with a reported mean SSIM of approximately 0.8 across the generated maps.","tokens_in":1948,"tokens_out":471,"duration_ms":32998,"significance":"An accessible, customizable, and fully open perfusion analysis package would address a clear practical gap in stroke imaging research. The unified CTP/MRP framework and emphasis on reproducibility are genuine strengths that could facilitate wider adoption if the quantitative fidelity of the derived hemodynamic values is convincingly demonstrated.","major_comments":[{"comment":"Validation / Results section: The central claim that PyPeT produces reliable quantitative CBF, CBV, MTT, TTP, and Tmax maps rests on a mean SSIM of ~0.8 versus commercial outputs. SSIM is insensitive to global scaling, offset, and absolute calibration; commercial packages are known to differ by 20–40 % in ROI-averaged CBF on identical data. The manuscript supplies only aggregate SSIM and visual comparison; it does not report ROI-level Pearson r, mean absolute percentage error, or Bland–Altman limits for the actual parameter values, nor any phantom or simulated ground-truth test. This gap directly undermines the assertion of quantitative accuracy.","section":"Validation / Results"}],"minor_comments":[{"comment":"Abstract and Methods: The patient cohort size, exact preprocessing pipeline, arterial-input-function selection criteria, and any regularization or deconvolution parameters are not stated. These details are required for independent reproduction.","section":"Abstract / Methods"},{"comment":"Figure captions and text: Ensure that all quantitative comparison figures include the number of slices or volumes averaged and the precise SSIM formula or implementation used.","section":"Figures"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the constructive feedback on the validation of PyPeT. We agree that stronger quantitative metrics would improve the manuscript and have revised accordingly where data permit.","responses":[{"response":"We acknowledge that SSIM measures structural similarity but is limited for absolute quantification and does not address scaling or calibration differences, especially given inter-vendor variability in commercial tools. In the revised manuscript we will add ROI-level Pearson correlation coefficients and mean absolute percentage error for CBF, CBV, MTT, TTP and Tmax. We will also include Bland–Altman analysis to assess bias and limits of agreement. These metrics will be computed on the existing comparison datasets and reported in an expanded Validation section. Phantom or simulated ground-truth experiments lie outside the present study scope, which centered on direct comparison with FDA-approved clinical tools; we will state this limitation explicitly.","revision_made":"partial","referee_comment":"The central claim that PyPeT produces reliable quantitative CBF, CBV, MTT, TTP, and Tmax maps rests on a mean SSIM of ~0.8 versus commercial outputs. SSIM is insensitive to global scaling, offset, and absolute calibration; commercial packages are known to differ by 20–40 % in ROI-averaged CBF on identical data. The manuscript supplies only aggregate SSIM and visual comparison; it does not report ROI-level Pearson r, mean absolute percentage error, or Bland–Altman limits for the actual parameter values, nor any phantom or simulated ground-truth test. This gap directly undermines the assertion of quantitative accuracy."}],"tokens_in":1433,"tokens_out":369,"duration_ms":90141,"standing_objections":["New phantom or simulated ground-truth experiments, which were not performed in the original work and would require separate data acquisition and validation infrastructure."]},"desk_editor":{"model":"grok-4.3","letter":"The paper's main contribution is a practical open-source Python tool that handles both CT and MR perfusion data in one framework to output CBF, CBV, MTT, TTP, and Tmax maps. It emphasizes modularity, low compute cost, inline documentation, and a debug mode that visualizes every processing step. The GitHub release and direct side-by-side comparisons to three FDA-approved commercial packages are the concrete deliverables here.","headline":"PyPeT gives researchers a free, modular Python pipeline for CTP and MRP maps with debug visuals and SSIM ~0.8 against commercial tools, but the numbers need tighter checks.","tokens_in":2438,"tokens_out":169,"would_cite":false,"duration_ms":20680,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":{"model":"grok-4.3","evidence":[{"relation":"unclear","rs_module":"IndisputableMonolith/Cost/FunctionalEquation.lean","rs_theorem":"washburn_uniqueness_aczel","paper_passage":"PyPeT employs an SVD-based deconvolution method to generate perfusion maps... CBF = max R(t), CBV = ∫R(t)dt, MTT = CBV/CBF"},{"relation":"unclear","rs_module":"IndisputableMonolith/Foundation/RealityFromDistinction.lean","rs_theorem":"reality_from_one_distinction","paper_passage":"Validation... mean SSIM around 0.8... compared to three FDA-approved commercial perfusion tools"}],"headline":"Perfusion imaging software tool with SVD deconvolution and SSIM validation lies outside RS scope","alignment":"orthogonal","rationale":"The paper presents a modular Python pipeline for CTP/MRP processing that extracts hemodynamic maps via SVD-based deconvolution of contrast time curves against an arterial input function. Its central machinery (bolus detection, gamma-variate AIF fitting, residue-function deconvolution, and SSIM-based comparison to commercial outputs) operates entirely within applied medical-image analysis and contains none of the RS primitives: no J-cost functional, no φ-ladder, no 8-tick periodicity, no parameter-free derivation of constants, and no recognition-theoretic forcing chain. RS modules such as Cost.FunctionalEquation, Foundation.RealityFromDistinction, and Foundation.DimensionForcing therefore have no bearing on the work.","tokens_in":46259,"confidence":"high","tokens_out":349,"duration_ms":13576,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"PyPeT produces CBF, CBV, MTT, TTP and Tmax maps from raw four-dimensional CT and MR perfusion data with mean SSIM of 0.8 against FDA-approved tools.","keywords":["perfusion imaging","CT perfusion","MR perfusion","open source software","brain hemodynamics","stroke assessment","quantitative mapping","Python framework"],"falsifier":"A side-by-side clinical study in which PyPeT-derived infarct core or penumbra volumes differ from commercial-tool volumes by more than 20 percent on the same patient scans.","tokens_in":2661,"feed_emoji":"🧠","tokens_out":653,"duration_ms":36738,"temperature":0.7,"pith_summary":"The paper introduces PyPeT as an open-source Python tool designed to process both CT and MR head perfusion scans into standard quantitative maps. It processes raw four-dimensional data to output cerebral blood flow, cerebral blood volume, mean transit time, time-to-peak, and time-to-maximum. Validation relies on visual and quantitative comparisons to maps from three commercial FDA-approved tools plus one research tool. The reported mean SSIM value near 0.8 is presented as evidence of stable correlation with existing solutions. The design emphasizes modularity, documentation, a debug visualization mode, and low computational demands to support research customization.","feed_headline":"Open Python tool generates brain perfusion maps with SSIM 0.8 to commercial standards","feed_subtitle":"PyPeT turns raw four-dimensional CT and MR data into CBF, CBV, MTT, TTP and Tmax maps using one customizable pipeline.","key_machinery":"The PyPeT processing pipeline, a unified modular framework for CTP and MRP data that includes inline documentation and an extensive debug visualization mode for each processing step.","core_discovery":"PyPeT is an openly available Python framework that accepts raw four-dimensional CTP and MRP data and generates CBF, CBV, MTT, TTP, and Tmax maps through a unified modular pipeline, achieving a mean SSIM around 0.8 in direct comparisons with reference maps produced by three FDA-approved commercial perfusion tools.","pith_inferences":["Widespread adoption could let groups add custom denoising or deconvolution methods and share them as community extensions.","If quantitative values prove repeatable across scanner vendors, the tool might serve as a common reference for multi-center perfusion studies.","Embedding PyPeT inside larger image-analysis pipelines could enable automated stroke triage workflows that combine perfusion with diffusion or angiography data."],"forward_implications":["Researchers gain a free, modifiable alternative to costly closed-source perfusion software for both CT and MR modalities.","A single codebase can handle data from two different imaging modalities without separate programs.","Inline documentation and debug mode allow users to inspect and verify every step of map generation.","Low computational requirements support use on standard research workstations rather than specialized hardware."],"fun_headline_variants":["PyPeT open source Python tool matches commercial brain perfusion maps at SSIM 0.8","PyPeT generates CBF CBV MTT TTP Tmax maps from raw CTP and MRP data","Unified modular Python tool PyPeT analyzes brain perfusion with SSIM 0.8 validation","PyPeT provides accessible Python pipeline for automated CT and MR perfusion analysis"],"cache_read_input_tokens":2112,"weakest_assumption_plain":"Agreement measured by SSIM of approximately 0.8 with commercial outputs constitutes sufficient validation of quantitative accuracy for research or potential clinical use.","fun_headline_variants_meta":{"raw":{"variants":["PyPeT open source Python tool matches commercial brain perfusion maps at SSIM 0.8","PyPeT generates CBF CBV MTT TTP Tmax maps from raw CTP and MRP data","Unified modular Python tool PyPeT analyzes brain perfusion with SSIM 0.8 validation","PyPeT provides accessible Python pipeline for automated CT and MR perfusion analysis"]},"model":"grok-4.3","cost_usd":0.007727,"raw_usage":{"total_tokens":3558,"prompt_tokens":719,"num_sources_used":0,"completion_tokens":92,"cost_in_usd_ticks":77274500,"prompt_tokens_details":{"text_tokens":719,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":2747,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":719,"tokens_out":92,"duration_ms":33447,"temperature":1.0,"reasoning_tokens":2747,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-05-21T18:28:01.274712+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"A side-by-side clinical study in which PyPeT-derived infarct core or penumbra volumes differ from commercial-tool volumes by more than 20 percent on the same patient scans.","supporting_citations":[],"review_version":1}