{"id":"311b8e73-5e81-45a2-a48c-8824283ad034","arxiv_id":"2607.00370","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"PaRC-mix applies coherent feature mixup across parallel routes in residual networks to reduce domain shift in prostate cancer segmentation from ADC MRI across five external scanner datasets.","lead":"The paper introduces PaRC-mix, a training method that mixes features from different samples at multiple layers inside segmentation networks to improve prostate cancer lesion detection on MRI. It tests whether this helps models trained on one scanner type perform better on images from other scanner manufacturers and field strengths.","discovery_kind":"unclear","skeptic_critique":{"model":"grok-4.3","headline":"Feature-level linear combinations may create non-physical activations that improve metrics on the 5 tested external scanners without ensuring true domain-invariant lesion features.","rationale":"The reader's weakest_assumption directly identifies the same load-bearing modeling assumption. Full-text results remain purely empirical on the chosen external cohorts; the proposed internal-partition check would falsify or corroborate whether the feature-mixup mechanism itself drives scanner-agnostic performance.","tokens_in":1908,"tokens_out":329,"duration_ms":15562,"concrete_test":"Re-train both MRRN and UNet++ with PaRC-mix, then evaluate the normalized composite score on a 20% held-out partition of the original 3T GE training distribution versus the five external scanner sets; if the external advantage disappears or reverses on the internal partition, the reported generalization benefit is likely an artifact of the particular test domains.","verdict_should_be":"UNCHANGED","load_bearing_attack":"PaRC-mix performs linear interpolation of intermediate activations drawn from different training samples (all from the single 3T GE source domain) at multiple layers. For the reported gains in composite DSC/HD95/MSD and the narrowed aggressive/non-aggressive gap to reflect genuine multi-domain generalization rather than dataset-specific regularization artifacts, those interpolated activations must preserve spatially coherent lesion semantics while suppressing scanner-specific cues. The paper provides no ablation that isolates whether the observed external-set improvements survive when the same models are evaluated on an internal held-out GE partition or on synthetically shifted data that does not match the test scanners; without that check the central empirical claim rests on an unverified invariance assumption.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.3","summary":"The manuscript proposes a novel Parallel-Route Coherent Mixup (PaRC-mix) regularization for single-source to multi-domain generalization in prostate cancer lesion segmentation on ADC MRI. PaRC-mix performs linear interpolation of intermediate activations from different training samples (all from one 3T GE source domain) at multiple layers of MRRN and UNet++ networks. Models trained with PaRC-mix are reported to outperform no-mixup and input-mixup baselines on 1547 samples from five external scanners (3T Siemens, 3T Philips, 1.5T Elekta), with improved DSC/HD95/MSD and a narrowed accuracy gap between aggressive and non-aggressive lesions (composite score gap reduced from 21.1/19.5 to 5.2/7.9).","tokens_in":2036,"tokens_out":549,"duration_ms":15556,"significance":"If the reported gains reflect genuine domain-invariant lesion features rather than test-set-specific artifacts, the method offers a practical, network-agnostic way to improve generalization for MR-guided adaptive radiotherapy without multi-source training data, which would address a clinically relevant barrier to automated segmentation adoption.","major_comments":[{"comment":"Abstract and results: claims of 'significantly more accurate' tumor detection and segmentation, plus the composite-score gap reductions, are presented without error bars, p-values, confidence intervals, or any statistical test details, undermining the ability to assess whether the improvements exceed experimental variability.","section":null},{"comment":"Methods: the mixup interpolation strength (lambda) and the precise definition of 'parallel routes' and affected layers are not specified, nor are ablation results isolating PaRC-mix from other regularization effects or from input-mixup.","section":null},{"comment":"Evaluation design: the central multi-domain generalization claim rests on external-scanner performance, yet no control experiments are described that evaluate the same models on a held-out partition of the source 3T GE domain or on synthetic domain shifts; without these, it is impossible to distinguish true invariance from scanner-specific regularization artifacts.","section":null},{"comment":"Results: the normalized composite DSC/HD95/MSD score used to quantify the aggressive/non-aggressive gap is introduced without definition of the normalization, weighting scheme, or exact formula, making the reported reductions (21.1/19.5 to 5.2/7.9) difficult to interpret or reproduce.","section":null}],"minor_comments":[],"recommendation":"major_revision","confidential_remarks":"The soundness rating is low primarily because of missing statistical reporting and controls; a revision should be required to supply these before the generalization claim can be properly evaluated."},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the detailed and constructive comments on our manuscript. We address each of the major comments below and will make the necessary revisions to improve the clarity and rigor of the paper.","responses":[{"response":"We agree that statistical details are essential. In the revised manuscript, we will report standard deviations across multiple training runs, include p-values from paired statistical tests comparing PaRC-mix to baselines, and provide confidence intervals for the key metrics including the composite scores.","revision_made":"yes","referee_comment":"Abstract and results: claims of 'significantly more accurate' tumor detection and segmentation, plus the composite-score gap reductions, are presented without error bars, p-values, confidence intervals, or any statistical test details, undermining the ability to assess whether the improvements exceed experimental variability."},{"response":"The full methods section in the manuscript provides some details, but we acknowledge the need for more precision. We will explicitly define the lambda distribution used for interpolation, clarify that parallel routes refer to the multiple resolution paths in MRRN and the nested skip connections in UNet++, specify the layers where mixup is applied, and include additional ablation experiments to isolate the effect of PaRC-mix.","revision_made":"yes","referee_comment":"Methods: the mixup interpolation strength (lambda) and the precise definition of 'parallel routes' and affected layers are not specified, nor are ablation results isolating PaRC-mix from other regularization effects or from input-mixup."},{"response":"This point is well-taken. While the primary goal is multi-domain generalization, we will add results evaluating the models on a held-out subset of the source domain to confirm that performance is maintained or improved. We will also discuss the rationale for focusing on real external scanner data rather than synthetic shifts.","revision_made":"yes","referee_comment":"Evaluation design: the central multi-domain generalization claim rests on external-scanner performance, yet no control experiments are described that evaluate the same models on a held-out partition of the source 3T GE domain or on synthetic domain shifts; without these, it is impossible to distinguish true invariance from scanner-specific regularization artifacts."},{"response":"We will revise the manuscript to include a clear definition of the composite score, including the normalization method (e.g., min-max scaling per metric), the weighting scheme (equal weights or specified), and the exact formula used to compute the gap reductions.","revision_made":"yes","referee_comment":"Results: the normalized composite DSC/HD95/MSD score used to quantify the aggressive/non-aggressive gap is introduced without definition of the normalization, weighting scheme, or exact formula, making the reported reductions (21.1/19.5 to 5.2/7.9) difficult to interpret or reproduce."}],"tokens_in":1643,"tokens_out":617,"duration_ms":23492,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The paper's core contribution is applying linear interpolation of intermediate features from different samples across parallel routes at multiple layers inside residual networks, then showing this helps MRRN and UNet++ models trained only on 3T GE data perform better on five external datasets from Siemens, Philips, and Elekta scanners. The reported drop in the aggressive versus non-aggressive lesion gap on a normalized composite of DSC, HD95, and MSD is the practical result worth noting.\n\nIt does a clean job of comparing the new scheme against no-mixup and input-only mixup baselines on the same two architectures, and the method itself is network-agnostic and simple to add. That matches the kind of incremental regularization work that can actually get used in clinics.\n\nThe soft spots sit in the missing details. The abstract states statistically significant gains but gives no error bars, p-values, exact interpolation coefficients, or ablation tables that isolate the parallel coherent placement from other mixup choices. The stress-test concern lands: without results on a held-out partition from the original GE domain or on synthetic shifts that do not match the test scanners, it remains unclear whether the interpolated features truly suppress scanner cues or simply produce activations that score well on these particular external sets. The assumption that linear combinations preserve lesion semantics while removing domain signals is stated but not directly tested.\n\nThis work is aimed at researchers doing multi-vendor MRI segmentation who need a lightweight way to improve generalization. A reader already familiar with mixup variants will see the specific implementation and the external-scanner numbers, but will still want the full methods and statistics before adopting it.\n\nThe paper deserves peer review because the task is clinically relevant and the approach is reproducible enough to check, even though the current evidence is thin on controls.","headline":"PaRC-mix is a straightforward multi-layer feature mixup variant that narrows the scanner gap in prostate lesion segmentation on external data, but the abstract supplies too few controls and statistics to confirm the gains reflect genuine domain invariance rather than regularization artifacts.","tokens_in":2548,"tokens_out":449,"would_cite":false,"duration_ms":16195,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"PaRC-mix applies linear feature combinations at multiple network layers to improve prostate lesion segmentation across MRI scanners and lesion types.","keywords":["prostate cancer segmentation","MRI","multi-domain generalization","mixup regularization","deep learning","residual networks"],"falsifier":"A new external test set from an unseen scanner vendor or field strength where PaRC-mix models show no improvement in composite DSC-HD95-MSD score over non-mixup baselines.","tokens_in":2812,"feed_emoji":"","tokens_out":683,"duration_ms":12809,"temperature":0.7,"pith_summary":"The paper introduces PaRC-mix, a training method that augments features inside residual networks by linearly mixing intermediate activations from different samples in a batch. It tests this on MRRN and UNet++ models trained on one 3T GE dataset and evaluated on five external datasets from Siemens, Philips, and Elekta scanners. The approach yields higher Dice scores, lower surface distances, and better detection of both aggressive and non-aggressive lesions than standard training or input-level mixup. The accuracy difference between aggressive and non-aggressive lesions shrinks markedly when PaRC-mix is used. The method is presented as network-agnostic and easy to add to multi-stream architectures for single-source to multi-domain generalization.","feed_headline":"Layer-wise feature mixup narrows prostate lesion accuracy gaps across scanners","feed_subtitle":"PaRC-mix training on two residual networks shrinks the aggressive vs non-aggressive performance difference from ~20 to under 8 normalized po","key_machinery":"parallel-route coherent mixup (PaRC-mix), which performs linear combinations of intermediate features drawn from different samples at several depths inside a residual network.","core_discovery":"PaRC-mix training creates coherent feature augmentations at multiple layers by linear combination of features from different training samples; when applied to MRRN and UNet++ networks for prostate cancer lesion segmentation, it produces significantly higher composite accuracy on external multi-scanner data and narrows the performance gap between aggressive and non-aggressive lesions from roughly 20 points to under 8 points.","pith_inferences":["The same layer-wise mixing strategy could be tried on other medical image segmentation tasks that suffer from scanner-domain shift.","Because the mixing occurs inside the network, it may interact differently with residual connections than with purely convolutional backbones.","If the linear combinations are applied only on selected layers, the benefit might be retained with lower computational cost during training."],"forward_implications":["Models trained with PaRC-mix achieve higher tumor detection recall at fixed precision on data from different manufacturers.","The performance difference between aggressive and non-aggressive lesions shrinks when the same architecture is trained with PaRC-mix instead of standard or input-mixup methods.","The method can be inserted into any multi-stream residual network without changing the architecture itself.","PaRC-mix outperforms both no-mixup training and mixup applied only at the input or backbone layers on the reported external datasets."],"fun_headline_variants":["PaRC-mix reduces prostate lesion accuracy gaps on multi-scanner datasets","Coherent mixup shrinks aggressive nonaggressive performance difference","Feature mixup at multiple routes narrows PCa segmentation gaps across vendors","PaRC-mix training lowers the accuracy gap between aggressive and nonaggressive PCa"],"cache_read_input_tokens":2112,"weakest_assumption_plain":"Linear combinations of intermediate features from different samples preserve the lesion-specific signal without creating new correlations that only look useful on the tested external datasets.","fun_headline_variants_meta":{"raw":{"variants":["PaRC-mix reduces prostate lesion accuracy gaps on multi-scanner datasets","Coherent mixup shrinks aggressive nonaggressive performance difference","Feature mixup at multiple routes narrows PCa segmentation gaps across vendors","PaRC-mix training lowers the accuracy gap between aggressive and nonaggressive PCa"]},"model":"grok-4.3","cost_usd":0.008171,"raw_usage":{"total_tokens":3779,"prompt_tokens":806,"num_sources_used":0,"completion_tokens":75,"cost_in_usd_ticks":81712000,"prompt_tokens_details":{"text_tokens":806,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":2898,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":806,"tokens_out":75,"duration_ms":19803,"temperature":1.0,"reasoning_tokens":2898,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-02T04:51:27.593191+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"A new external test set from an unseen scanner vendor or field strength where PaRC-mix models show no improvement in composite DSC-HD95-MSD score over non-mixup baselines.","supporting_citations":[],"review_version":1}