Pith. sign in

REVIEW 4 major objections 6 minor 58 references

Histopathological Spectrum-Guided Prostate Stratification via Segmentation-Assisted Diagnostic Transformer

T0 review · 4 major / 6 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read This paper claims that adding text-prompted prostate-region masking and slice-sequence fusion to MRI classifiers yields expert-comparable four-class risk stratification, with the best model surpassing a 15-year radiologist in overall accura

desk verdict A useful four-class prostate MRI benchmark and a sensible segmentation-assisted fusion method, but the evidence behind the headline claims is weaker than the prose suggests—especially the cohort selection and the radiologist comparison. read the letter →

arxiv 2607.22703 v1 pith:ZATLDPUA submitted 2026-07-19 cs.CV

Leyang Li , Lihua Chen , Huangang Hu , Tianhang Hao , Hao Cheng , Xin Zhang , Qianru Sun , Bingxu Lu
show 2 more authors
Wenlong Yu Feng Duan
This is my paper · ORCID
classification cs.CV
keywords prostatecancermultiparametricMRIfour-classclassificationhistopathologylabelszero-shotsegmentationslice-wiseattentionriskstratificationpathology-groundeddataset
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

The paper tries to establish that prostate MRI diagnosis can move beyond binary benign/malignant or PI-RADS scoring to a four-class, pathology-grounded stratification—normal, benign prostatic hyperplasia, non-significant cancer, and clinically significant cancer—and that a model combining automatic text-prompted prostate-region segmentation with cross-slice attention achieves expert-comparable accuracy. To test this, the authors construct a small biopsy-confirmed dataset and propose the LSDT pipeline: a zero-shot segmenter masks the prostate in each T2WI/ADC/DWI slice, a pre-trained backbone extracts per-slice features, and a transformer fuses slice features before classification. Across five-fold cross-validation on 344 patients, the best variant reaches 0.633 mean accuracy and 0.768 JointRecall, beating the average of three radiologists and the senior reader in overall accuracy. If correct, this would support using anatomy-grounded, pathology-supervised models for more clinically meaningful MRI risk stratification and potentially fewer unnecessary biopsies.

What carries the argument

The load-bearing mechanism is the text-prompted zero-shot segmentation of the prostate, performed by a promptable open-vocabulary segmentation model with the single word 'prostate' as prompt. The resulting binary mask is multiplied element-wise into each slice of all three MRI modalities, suppressing background and focusing the encoder on the gland. Around this, slice embeddings are computed by a pre-trained vision backbone from the concatenated masked modalities, positionally encoded, and passed through a two-layer transformer encoder whose self-attention weights adaptively combine informative slices; a mean-pooled representation feeds a four-class head. The same mask can be added as a weak

What would settle it

Run LSDT on an unselected consecutive cohort (or a held-out multi-center four-class dataset) with pre-registered inclusion/exclusion counts; if the accuracy gap over the senior radiologist shrinks or reverses, or if class-wise confusion shifts with case mix, the central claim of expert-level fine-grained stratification is not generalizable.

Watch

Extended reading notes

Core claim

The central claim is that pathology-grounded supervision plus anatomical priors materially improves fine-grained prostate MRI classification. The authors introduce PCa-HSD, a four-class prostate mpMRI dataset with biopsy-confirmed labels, and LSDT, which uses a text-prompted zero-shot segmentation model (prompt: 'prostate') to generate per-slice prostate masks, applies those masks to T2WI, ADC, and DWI volumes, extracts slice embeddings with pre-trained vision models, and aggregates them with a transformer that learns slice order and inter-slice attention. On 344 patients, LSDT-Large achieves mean accuracy 0.633±0.050 and JointRecall 0.768±0.040, consistently improving over its backbone with

Load-bearing premise

The 344-patient cohort is assumed to be a representative, unbiased sample of the clinical population, but the paper does not report how or why patients were excluded from the 3,810 screened cases.

Editorial extensions

If this is right

  • The four-class task (normal/BPH/nsPCa/sPCa) is learnable from mpMRI with pathology labels, yielding accuracy meaningfully above chance and above generic foundation models.
  • Anatomy-guided masking via text-prompted segmentation improves accuracy across multiple backbone architectures, not just one model.
  • Slice-wise attention fusion outperforms mean, sum, concat, and CNN-based fusion on this task, indicating that adaptive cross-slice weighting matters.
  • On a public multi-center binary dataset, the same framework reaches AUC 0.840, suggesting cross-institutional transfer of the masking/fusion strategy.
  • The best model matches or exceeds radiologist accuracy on the internal cohort, supporting potential clinical utility for pre-biopsy triage.

Reading between the lines

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

  • If the cohort selection is biased toward clear-cut cases, the radiologist comparison may overstate the model's practical edge; a consecutive or externally validated four-class cohort would be the real test.
  • The 'normal' class being defined as no clinical evidence despite biopsy in all patients raises labeling questions; a truly normal group without biopsy would behave differently, and the model might be learning scan-level artifacts rather than tissue status.
  • The method's reliance on a frozen zero-shot segmenter means failures in low-contrast peripheral-zone lesions (which the paper itself notes) would propagate into classification; quantifying segmentation error's effect on accuracy is a natural next step.
  • Combining PSA or other clinical variables with the imaging pipeline is an obvious extension, as the paper notes, and might close the remaining gap on nsPCa.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper introduces PCa-HSD, a retrospectively collected mpMRI dataset of 344 patients labeled into four pathology-derived classes (Normal, BPH, nsPCa, sPCa), and proposes LSDT, a pipeline that uses text-prompted SAM3 zero-shot segmentation to mask the prostate region, extracts slice-wise features with pretrained vision backbones, and aggregates them with a slice-level Transformer for four-class classification. The authors report internal five-fold cross-validation ACC and JointRecall for several backbones, with LSDT-Large reaching 0.633±0.050 ACC and 0.768±0.040 JointRecall, claim consistent gains over baselines and over a senior radiologist, and add an external binary experiment on the public PI-CAI cohort. The paper's central claims are that the new four-class benchmark is clinically meaningful and that pathology-grounded supervision plus anatomical priors improves fine-grained prostate MRI classification.

Significance. If the dataset and results are fully documented and the comparisons are statistically sound, the paper would be a useful contribution: a four-class, pathology-grounded prostate MRI benchmark is genuinely needed, and the LSDT design is a reasonable, reproducible template combining zero-shot segmentation, pretrained encoders, and slice-sequence attention. The external PI-CAI evaluation is a commendable attempt at generalization testing. However, the current evidence is not yet strong enough to support the headline claims: the cohort-selection process is not documented in a way that rules out selection bias, and the reported gains are mostly within one standard deviation of baseline performance, with no paired significance testing. The method itself is not circular and the main equations are straightforward masking/feature-extraction/fusion steps, but the empirical foundation needs substantial strengthening.

major comments (4)
  1. [§4.1.1, Dataset construction] The cohort-selection procedure is load-bearing but underdocumented. The text says 3,810 patients were retrospectively collected, inclusion criteria are listed, and then "Based on imaging findings and pathology, 344 patients were included in the final analysis" — with no exclusion counts, no criteria for the imaging-based part of the selection, and no order of filtering. If imaging appearance or technical quality beyond PI-QUAL ≥4 was used to exclude cases, all internal accuracies and the radiologist comparison are estimates on an enriched sample and may not transfer to a screened clinical population. In addition, the Normal class is defined as "no clinical evidence of prostate disease," yet every included patient had biopsy or radical prostatectomy within 4 weeks; this needs operationalization (e.g., negative biopsy, no cancer in radical prostatectomy, PI-RADS 1–2 with negative biopsy).
  2. [§5.1, Table 1] The evidence for "consistent performance gains" and "surpassing even the 15-year senior radiologist" is not statistically supported as presented. The comparison is between means and standard deviations across folds; no paired significance tests, confidence intervals for differences, or effect sizes are reported. For the headline comparison, LSDT-Large ACC 0.633±0.050 versus ViT-Large 0.596±0.035 and versus the senior radiologist 0.622 are differences well within one standard deviation. JointRecall is also not consistently improved: LSDT-ResNet50 (0.674±0.134) is essentially unchanged versus ResNet-50 (0.672±0.047), and LSDT-Omnirad (0.765±0.029) is numerically below Omnirad-Vision (0.770±0.067). Please report per-fold paired results (e.g., Wilcoxon signed-rank or paired bootstrap), difference intervals, and be precise about which comparisons actually reach significance. The phrase "surpa
  3. [§5.3.1, Table 3] The ablation conclusion as written is contradicted by Table 3 for the ResNet-50 backbone. The text states "For both ResNet-50 and ViT-Large, SAM3 achieves the best performance, followed by MedSAM..." but in Table 3, LSDT-ResNet50 with SAM3 has ACC 0.567±0.061, which is lower than LSDT-ResNet50 with MedSAM(1/2) at 0.570±0.040, and its JointRecall 0.674±0.134 is substantially worse than both MedSAM variants (0.735±0.084 and 0.753±0.060). The claim that mask quality is a key factor and that SAM3 consistently outperforms MedSAM is therefore not supported across backbones. Please correct the claim or explain why the ResNet-50 case is consistent with the stated conclusion.
  4. [§5.2, Fig. 5] The external validation is interpreted too strongly. The PI-CAI experiment changes the task to binary classification, adapts LSDT by using the SAM3 mask as an additional input channel rather than hard masking, and uses learnable slice fusion; this is not the same model or task as the internal four-class claim. The baseline comparisons (UNet, ResNet3D, EfficientNet, ViT) may not reproduce the current PI-CAI state of the art, and no leaderboard comparison is given. The statement that the same model shows "consistent superiority across the internal four-class cohort and external binary cohort" therefore overstates what the experiment shows. Please either narrow the claim or add results from the original four-class pipeline on PI-CAI, if feasible, plus published reference points.
minor comments (6)
  1. [Table 1] The LSDT-Large row is corrupted in the text (e.g., "0 00...666333333±±±000...000555000"). Please ensure the table renders correctly.
  2. [Eq. (1)] S is defined as the total number of slices for a patient, but the preprocessing section says the central 16 slices are selected and used as input. Please reconcile the notation or define S consistently as the number of retained slices.
  3. [§3.4, Eq. (2)] The description says T2WI, ADC, and DWI are "stacked along the channel dimension." Please clarify the channel count and how this aligns with pretrained backbones expecting a specific number of input channels.
  4. [§4.1.4, MLLM evaluation] The MLLM prompt construction is described only briefly; details such as the composite image resolution, ordering of modalities, and whether the model can see all 16 slices are needed for reproducibility. Since MLLM performance is much lower, a sensitivity analysis of prompting style would also help.
  5. [Eq. (8)] JointRecall, as defined, is simply the overall accuracy after merging the four classes into benign (Normal+BPH) and cancer (nsPCa+sPCa). The name "JointRecall" is nonstandard and could mislead readers into thinking it is a recall-like metric. Please rename or explicitly describe it as the merged binary accuracy.
  6. [§5.4, Limitations] The limitations paragraph is candid and useful, but it could acknowledge more directly that the internal cohort is from two institutions in one city and that the Normal class definition limits conclusions about screening populations.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the core derivation is empirical and self-contained, with only peripheral non-load-bearing self-citations.

full rationale

The paper's central claim is that LSDT improves four-class prostate MRI classification by masking with SAM3 and fusing slice features. The derivation chain is explicit: Eq. (1) is an element-wise mask multiplication; Eq. (2) is a standard encoder over channel-concatenated masked slices; Eqs. (3)-(7) describe positional encoding and multi-head self-attention; Eq. (8) defines JointRecall as a merged-binary accuracy measure; Eq. (9) is the standard AUC integral. None of these equations define the predicted class in terms of the ground-truth label or fit a parameter that is then reported as a prediction. The class labels come from histopathology and the ROI mask comes from frozen SAM3 with a text prompt, not from the classification target. Comparisons are empirical five-fold cross-validation on an internal cohort plus an external PI-CAI AUC evaluation, so the central result is not forced by construction. The paper contains a few self-citations ([25], [27], [58]), but they appear in background or future-work contexts and do not carry the load of the main claim; there is no imported uniqueness theorem or ansatz smuggled in via self-citation. The cohort-selection documentation concern raised by the skeptic is a representativeness/validation issue, not a circularity issue, because it does not make any equation or fitted parameter equivalent to its own input. Thus the paper is self-contained with respect to circularity.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The central claim rests on clinical labeling assumptions, on the quality of SAM3 masks, and on the representativeness of a heavily filtered retrospective cohort. No new physical or biological entity is introduced; the four-class labels reuse standard pathology categories.

free parameters (5)
  • Number of central slices S = 16
    Center-cropped to the central 16 slices covering the prostate; chosen by hand, may omit informative slices for large prostates.
  • Input resolution = 224x224
    All slices resized to 224x224 for pretrained backbones; standard but discards native resolution.
  • Training hyperparameters = lr=1e-5, epochs=100, batch=16, Adam
    Fixed across variants; chosen without reported tuning or sensitivity analysis.
  • Fusion Transformer depth L = 2
    Two-layer encoder selected without an ablation of depth or width.
  • Augmentation ranges = rotation ±15°, flip 0.5, translate ±10%, scale 0.9-1.1
    Hand-chosen augmentation parameters; no sensitivity analysis.
assumptions (5)
  • domain assumption Histopathology from biopsy or radical prostatectomy is an error-free gold standard.
    Labels come from clinical pathology reports; no central pathology review and no biopsy-to-MRI lesion registration are described, so sampling errors could mislabel cases.
  • domain assumption The four-class partition (Normal, BPH, nsPCa=Gleason 6, sPCa=Gleason 7+) is clinically meaningful and MRI-separable.
    The partition follows ISUP/Gleason thresholds and clinical management categories, but the paper provides no outcome-based validation of the grouping.
  • domain assumption SAM3 zero-shot text-prompted masks are accurate enough on prostate MRI to improve downstream classification.
    Only qualitative examples are shown in Figure 6; the paper itself acknowledges in Section 5.4 that SAM3 may underperform in low-contrast or peripheral-zone regions.
  • domain assumption Preprocessing, co-registration, and intensity standardization preserve the diagnostic signal in all modalities.
    N4ITK bias correction, intra-examination registration, resampling, and min-max normalization are applied to all volumes; the paper assumes these steps do not remove lesion contrast or introduce misalignment artifacts.
  • ad hoc to paper The PCa-HSD cohort of 344 patients is representative of the target population despite screening 3,810 patients.
    Inclusion was based on imaging findings and pathology, with no exclusion counts; the Normal class definition conflicts with the biopsy/RP inclusion criterion. This assumption is not independently grounded.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Histopathological Spectrum-Guided Prostate Stratification via Segmentation-Assisted Diagnostic Transformer." pith.science (2026). https://pith.science/paper/ZATLDPUA

@misc{pith2026260722703,
  author       = {Pith},
  title        = {Pith review of: Histopathological Spectrum-Guided Prostate Stratification via Segmentation-Assisted Diagnostic Transformer},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZATLDPUA}},
  note         = {Machine review of arXiv:2607.22703}
}
read the original abstract

Prostate cancer diagnosis with multiparametric MRI (mpMRI) is commonly based on PI-RADS assessment or binary classification, which suffer from subjectivity and fail to capture clinically relevant pathological heterogeneity. To address this limitation, we construct a Prostate Cancer Histopathology Spectrum Dataset (PCa-HSD) and formulate a clinically meaningful four-class classification task, addressing the underrepresentation of benign lesions that are easily confounded with prostate cancer in existing datasets. We propose Language-guided Segmentation-assisted Diagnostic Transformer model (LSDT), which leverages zero-shot segmentation to provide anatomical priors and performs effective multi-modal slice fusion for classification. Our proposed method consistently improves accuracy across backbones, achieving the best average accuracy of 0.633 and JointRecall of 0.768 in five-fold cross-validation on a cohort of 344 patients. These results demonstrate that integrating pathology supervision and anatomical priors significantly enhances fine-grained prostate MRI classification and provides a more clinically relevant paradigm for risk stratification. Code will be made publicly available in a future revision.

Figures

Figures reproduced from arXiv: 2607.22703 by the authors.

Figure 1
Figure 1. Performance comparison on the internal PCa-HSD co [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Overview of the proposed LSDT framework. Multi-modal MRI volumes (T2WI, ADC, DWI) are used as [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Representative T2WI examples of the four [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Confusion matrices on the internal PCa-HSD [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Public binary-cohort comparison (AUC) on the [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Prostate segmentation masks of different strate [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

58 extracted references · 3 canonical work pages

  1. [1]

    Epidemiology of prostate cancer

    Prashanth Rawla. Epidemiology of prostate cancer. World Journal of Oncology, 10(2):63–89, 2019

    2019

  2. [2]

    Clinical variability and molecular heterogeneity in prostate cancer.Asian Journal of Andrology, 18(4):543–548, 2016

    Jonathan Shoag and Christopher E Barbieri. Clinical variability and molecular heterogeneity in prostate cancer.Asian Journal of Andrology, 18(4):543–548, 2016

    2016

  3. [3]

    Pi- rads prostate imaging–reporting and data system: 2015, version 2.European Urology, 69(1):16–40, 2016

    Jeffrey C Weinreb, Jelle O Barentsz, Peter L Choyke, Francois Cornud, Masoom A Haider, Katarzyna J Macura, Daniel Margolis, Mitchell D Schnall, Faina Shtern, Clare M Tempany, et al. Pi- rads prostate imaging–reporting and data system: 2015, version 2.European Urology, 69(1):16–40, 2016

    2015

  4. [4]

    Xueqing Cheng, Yuntian Chen, Hui Xu, Lei Ye, Shun Tong, Haixia Li, Tianjing Zhang, Song Tian, Jin Qi, Hao Zeng, et al. Avoiding unnecessary systematic biopsy in clinically significant prostate cancer: Comparison between mri-based radiomics model and pi-rads category.Journal of Magnetic Resonance Imaging, 57(2):578–586, 2023

    2023

  5. [5]

    Singer, Kaynaat Syed, Bruce Jacobs, Andres Correa, Marc Smaldone, and Serge Ginzburg

    Johnathan Drevik, Zafardjan Dalimov, Robert Uzzo, John Danella, Thomas Guzzo, Lawrence Belkoff, Jay Raman, Jeffrey Tomaszewski, Edouard Trabulsi, Adam Reese, Eric A. Singer, Kaynaat Syed, Bruce Jacobs, Andres Correa, Marc Smaldone, and Serge Ginzburg. Utility of psa density in patients with pi- rads 3 lesions across a large multi-institutional col- labora...

    work page doi:10.1016/j.urolonc.2022.08.003 2022

  6. [6]

    Multimodal approach to optimize biopsy decision-making for pi-rads 3 lesions on multiparametric mri.Clinical Imaging, 117:110363, 2025

    Omer Tarik Esengur, Enis C Yilmaz, Kutsev B Ozy- oruk, Alex Chen, Nathan S Lay, David G Gelik- man, Maria J Merino, Sandeep Gurram, Bradford J Wood, Peter L Choyke, et al. Multimodal approach to optimize biopsy decision-making for pi-rads 3 lesions on multiparametric mri.Clinical Imaging, 117:110363, 2025

    2025

  7. [7]

    Franco Palmisano, Vito Lorusso, Rebecca Leg- nani, Vincenzo Martorello, Carlotta Nedbal, Pietro Tramanzoli, Federica Marchesotti, Simona Ferraro, Michele Talso, Antonio Maria Granata, et al. Anal- ysis of the performance and accuracy of a psa and psa ratio-based nomogram to predict the probabil- ity of prostate cancer in a cohort of patients with pirads 3...

    2024

  8. [8]

    To biopsy or not biopsy, that is the question-pi-rads 3 prostate lesions– validation of clinical and radiological parameters for biopsy decision-making.BMC Urology, 25(1):274, 2025

    Toni Franz, Tom Sicker, Julian Lueke, Benny Dinh, Thi Phuc Ho, Theodoros Spinos, Lars-Christian Horn, Alexander Schaudinn, Evangelos Liatsikos, and Jens-Uwe Stolzenburg. To biopsy or not biopsy, that is the question-pi-rads 3 prostate lesions– validation of clinical and radiological parameters for biopsy decision-making.BMC Urology, 25(1):274, 2025

    2025

Show all 58 references
  1. [9]

    Matthew D Greer, Joanna H Shih, Nathan Lay, Tris- tan Barrett, Leonardo Bittencourt, Samuel Borof- sky, Ismail Kabakus, Yan Mee Law, Jamie Marko, Haytham Shebel, et al. Interreader variability of prostate imaging reporting and data system version 2 in detecting and assessing p...

    2019

  2. [10]

    Mixed supervision of histopathology improves prostate cancer classification from mri.IEEE Transactions on Medical Imaging, 43(7):2610–2622, 2024

    Abhejit Rajagopal, Antonio C Westphalen, Nathan Velarde, Jeffry P Simko, Hao Nguyen, Thomas A Hope, Peder EZ Larson, and Kirti Magudia. Mixed supervision of histopathology improves prostate cancer classification from mri.IEEE Transactions on Medical Imaging, 43(7):2610–2622, 2024. 13

    2024

  3. [11]

    Chunguang Yang, Basen Li, Yang Luan, Shiwei Wang, Yang Bian, Junbiao Zhang, Zefeng Wang, Bo Liu, Xin Chen, Marcus Hacker, Zhen Li, Xi- ang Li, and Zhihua Wang. Deep learning model for the detection of prostate cancer and classification of clinically significant disease using m...

    2024 doi

  4. [12]

    Classification of clinically significant prostate cancer on multi- parametric mri: a validation study comparing deep learning and radiomics.Cancers, 14(1):12, 2021

    Jose M Castillo T, Muhammad Arif, Martijn PA Starmans, Wiro J Niessen, Chris H Bangma, Ivo G Schoots, and Jifke F Veenland. Classification of clinically significant prostate cancer on multi- parametric mri: a validation study comparing deep learning and radiomics.Cancers, 14(1...

    2021

  5. [13]

    Joint prostate cancer detection and gleason score prediction in mp-mri via focal- net.IEEE Transactions on Medical Imaging, 38(11): 2496–2506, 2019

    Ruiming Cao, Amirhossein Mohammadian Bajgi- ran, Sohrab Afshari Mirak, Sepideh Shakeri, Xin- ran Zhong, Dieter Enzmann, Steven Raman, and Kyunghyun Sung. Joint prostate cancer detection and gleason score prediction in mp-mri via focal- net.IEEE Transactions on Medical Imaging,...

    2019

  6. [14]

    Prostate classification network (pc-net) for automated classification of prostate can- cer in magnetic resonance imaging.Multimedia Tools and Applications, 84(10):7717–7743, 2025

    Mamta Juneja, Sumindar Kaur Saini, Kunal Sharma, and Prashant Jindal. Prostate classification network (pc-net) for automated classification of prostate can- cer in magnetic resonance imaging.Multimedia Tools and Applications, 84(10):7717–7743, 2025

    2025

  7. [15]

    Advancing prostate cancer diagnosis: A deep learning approach for enhanced detection in mri images.Diagnostics, 14(17):1871, 2024

    Alparslan Horasan and Ali Güne¸ s. Advancing prostate cancer diagnosis: A deep learning approach for enhanced detection in mri images.Diagnostics, 14(17):1871, 2024

    2024

  8. [16]

    Computer- aided detection of prostate cancer in mri.IEEE Transactions on Medical Imaging, 33(5):1083– 1092, 2014

    Geert Litjens, Oscar Debats, Jelle Barentsz, Nico Karssemeijer, and Henkjan Huisman. Computer- aided detection of prostate cancer in mri.IEEE Transactions on Medical Imaging, 33(5):1083– 1092, 2014. doi: 10.1109/TMI.2014.2303821

    2014

  9. [17]

    Zhiwei Wang, Chaoyue Liu, Danpeng Cheng, Liang Wang, Xin Yang, and Kwang-Ting Cheng. Au- tomated detection of clinically significant prostate cancer in mp-mri images based on an end-to-end deep neural network.IEEE Transactions on Medi- cal Imaging, 37(5):1127–1139, 2018. doi: ...

    2018

  10. [18]

    Deep learning re- gression for prostate cancer detection and grading in bi-parametric mri.IEEE Transactions on Biomedi- cal Engineering, 68(2):374–383, 2020

    Coen De Vente, Pieter V os, Matin Hosseinzadeh, Josien Pluim, and Mitko Veta. Deep learning re- gression for prostate cancer detection and grading in bi-parametric mri.IEEE Transactions on Biomedi- cal Engineering, 68(2):374–383, 2020

    2020

  11. [19]

    Deep learning-based pi-rads score estimation to detect prostate cancer using multipara- metric magnetic resonance imaging.Computers and Electrical Engineering, 102:108275, 2022

    Kadir Yildirim, Muhammed Yildirim, Hasan Erye- sil, Muhammed Talo, Ozal Yildirim, Murat Kara- batak, Mehmet Sezai Ogras, Hakan Artas, and U Ra- jendra Acharya. Deep learning-based pi-rads score estimation to detect prostate cancer using multipara- metric magnetic resonance ima...

    2022

  12. [20]

    Multi-modal siamese network for diagnostically similar lesion retrieval in prostate mri.IEEE Trans- actions on Medical Imaging, 40(3):986–995, 2020

    Alberto Rossi, Matin Hosseinzadeh, Monica Bian- chini, Franco Scarselli, and Henkjan Huisman. Multi-modal siamese network for diagnostically similar lesion retrieval in prostate mri.IEEE Trans- actions on Medical Imaging, 40(3):986–995, 2020

    2020

  13. [21]

    Jonathan I Epstein, Lars Egevad, Mahul B Amin, Brett Delahunt, John R Srigley, Peter A Humphrey, Grading Committee, et al. The 2014 international society of urological pathology (isup) consensus conference on gleason grading of prostatic carci- noma: definition of grading patt...

    2014

  14. [22]

    Eau guidelines on prostate cancer.European Urology, 53(1):68–80, 2008

    Axel Heidenreich, Gunnar Aus, Michel Bolla, Steven Joniau, Vsevolod B Matveev, Hans Peter Schmid, and Filliberto Zattoni. Eau guidelines on prostate cancer.European Urology, 53(1):68–80, 2008

    2008

  15. [23]

    Prostate cancer and its mimics—a pictorial review

    Anna ˙Zurowska, Rafał Peksa, Michał Bie ´nkowski, Katarzyna Skrobisz, Marek Sowa, Marcin Ma- tuszewski, Wojciech Biernat, and Edyta Szurowska. Prostate cancer and its mimics—a pictorial review. Cancers, 15(14):3682, 2023

    2023

  16. [24]

    Pitfalls in 14 prostate mri interpretation: A pictorial review.Sem- inars in Roentgenology, 56(4):391–405, 2021

    Andrei S Purysko, Benjamin J Childs, Ryan D Ward, Leonardo K Bittencourt, and Eric A Klein. Pitfalls in 14 prostate mri interpretation: A pictorial review.Sem- inars in Roentgenology, 56(4):391–405, 2021. doi: 10.1053/j.ro.2021.07.005

    2021 doi

  17. [25]

    Fine-grained domain generalization with feature structuralization.IEEE Transactions on Multimedia, 27:8510–8524, 2025

    Wenlong Yu, Dongyue Chen, Qilong Wang, and Qinghua Hu. Fine-grained domain generalization with feature structuralization.IEEE Transactions on Multimedia, 27:8510–8524, 2025. doi: 10.1109/ TMM.2025.3607716

    2025

  18. [26]

    Cat-net: A cross-slice attention transformer model for prostate zonal segmentation in mri.IEEE Transactions on Medical Imaging, 42 (1):291–303, 2022

    Alex Ling Yu Hung, Haoxin Zheng, Qi Miao, Steven S Raman, Demetri Terzopoulos, and Kyunghyun Sung. Cat-net: A cross-slice attention transformer model for prostate zonal segmentation in mri.IEEE Transactions on Medical Imaging, 42 (1):291–303, 2022

    2022

  19. [27]

    Feature reinforcement network for image classification

    Bingxu Lu, Qinghua Hu, Yijing Hui, Quan Wen, and Min Li. Feature reinforcement network for image classification. In2018 IEEE International Confer- ence on Multimedia and Expo (ICME), pages 1–6. IEEE, 2018

    2018

  20. [28]

    Cross-slice attention and evi- dential critical loss for uncertainty-aware prostate cancer detection

    Alex Ling Yu Hung, Haoxin Zheng, Kai Zhao, Kaifeng Pang, Demetri Terzopoulos, and Kyunghyun Sung. Cross-slice attention and evi- dential critical loss for uncertainty-aware prostate cancer detection. InInternational Conference on Medical Image Computing and Computer-Assisted I...

    2024

  21. [29]

    Sam 3: Seg- ment anything with concepts.arXiv preprint arXiv:2511.16719, 2025

    Nicolas Carion, Laura Gustafson, Yuan-Ting Hu, Shoubhik Debnath, Ronghang Hu, Didac Suris, Chaitanya Ryali, Kalyan Vasudev Alwala, Haitham Khedr, Andrew Huang, et al. Sam 3: Seg- ment anything with concepts.arXiv preprint arXiv:2511.16719, 2025

    2025 arXiv

  22. [30]

    Muhammad Arif, Ivo G Schoots, Jose Castillo To- var, Chris H Bangma, Gabriel P Krestin, Monique J Roobol, Wiro Niessen, and Jifke F Veenland. Clin- ically significant prostate cancer detection and seg- mentation in low-risk patients using a convolutional neural network on mult...

    2020

  23. [31]

    Zhaonan Sun, Pengsheng Wu, Yingpu Cui, Xiang Liu, Kexin Wang, Ge Gao, Huihui Wang, Xiaodong Zhang, and Xiaoying Wang. Deep-learning mod- els for detection and localization of visible clini- cally significant prostate cancer on multi-parametric mri.Journal of Magnetic Resonance...

    2023

  24. [32]

    Fully automated deep learning model to detect clin- ically significant prostate cancer at mri.Radiology, 312(2):e232635, 2024

    Jason C Cai, Hirotsugu Nakai, Shiba Kuanar, Adam T Froemming, Candice W Bolan, Akira Kawashima, Hiroaki Takahashi, Lance A Mynderse, Chandler D Dora, Mitchell R Humphreys, et al. Fully automated deep learning model to detect clin- ically significant prostate cancer at mri.Radi...

    2024

  25. [33]

    Multimodal triplet attention network for brain disease diagno- sis.IEEE Transactions on Medical Imaging, 41(12): 3884–3894, 2022

    Qi Zhu, Heyang Wang, Bingliang Xu, Zhiqiang Zhang, Wei Shao, and Daoqiang Zhang. Multimodal triplet attention network for brain disease diagno- sis.IEEE Transactions on Medical Imaging, 41(12): 3884–3894, 2022

    2022

  26. [34]

    Segment anything

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alexander C Berg, Wan- Yen Lo, et al. Segment anything. InProceedings of the IEEE/CVF International Conference on Com- puter Vision, pages 4015–4026, 2023

    2023

  27. [35]

    Sam 2: Segment any- thing in images and videos.arXiv preprint arXiv:2408.00714, 2024

    Nikhila Ravi, Valentin Gabeur, Yuan-Ting Hu, Ronghang Hu, Chaitanya Ryali, Tengyu Ma, Haitham Khedr, Roman Rädle, Chloe Rolland, Laura Gustafson, et al. Sam 2: Segment any- thing in images and videos.arXiv preprint arXiv:2408.00714, 2024

    2024 arXiv

  28. [36]

    Segment everything every- where all at once.Advances in Neural Information Processing Systems, 36:19769–19782, 2023

    Xueyan Zou, Jianwei Yang, Hao Zhang, Feng Li, Linjie Li, Jianfeng Wang, Lijuan Wang, Jianfeng Gao, and Yong Jae Lee. Segment everything every- where all at once.Advances in Neural Information Processing Systems, 36:19769–19782, 2023

    2023

  29. [37]

    Segment and recognize anything at any granularity

    Feng Li, Hao Zhang, Peize Sun, Xueyan Zou, Shi- long Liu, Chunyuan Li, Jianwei Yang, Lei Zhang, and Jianfeng Gao. Segment and recognize anything at any granularity. InEuropean Conference on Com- puter Vision, pages 467–484, 2024

    2024

  30. [38]

    Segment anything in medical images

    Jun Ma, Yuting He, Feifei Li, Lin Han, Chenyu You, and Bo Wang. Segment anything in medical images. Nature Communications, 15(1):654, 2024. 15

    2024

  31. [39]

    Medsam2: Segment anything in 3d medical images and videos

    Jun Ma, Zongxin Yang, Sumin Kim, Bihui Chen, Mohammed Baharoon, Adibvafa Fallahpour, Reza Asakereh, Hongwei Lyu, and Bo Wang. Medsam2: Segment anything in 3d medical images and videos. arXiv preprint arXiv:2504.03600, 2025

    2025 arXiv

  32. [40]

    Seg- ment anything model for medical image analysis: an experimental study.Medical Image Analysis, 89: 102918, 2023

    Maciej A Mazurowski, Haoyu Dong, Hanxue Gu, Jichen Yang, Nicholas Konz, and Yixin Zhang. Seg- ment anything model for medical image analysis: an experimental study.Medical Image Analysis, 89: 102918, 2023

    2023

  33. [41]

    Segment any- thing model for medical images?Medical Image Analysis, 92:103061, 2024

    Yuhao Huang, Xin Yang, Lian Liu, Han Zhou, Ao Chang, Xinrui Zhou, Rusi Chen, Junxuan Yu, Jiongquan Chen, Chaoyu Chen, et al. Segment any- thing model for medical images?Medical Image Analysis, 92:103061, 2024

    2024

  34. [42]

    Ma- sam: Modality-agnostic sam adaptation for 3d med- ical image segmentation.Medical Image Analysis, 98:103310, 2024

    Cheng Chen, Juzheng Miao, Dufan Wu, Aox- iao Zhong, Zhiling Yan, Sekeun Kim, Jiang Hu, Zhengliang Liu, Lichao Sun, Xiang Li, et al. Ma- sam: Modality-agnostic sam adaptation for 3d med- ical image segmentation.Medical Image Analysis, 98:103310, 2024

    2024

  35. [43]

    Francesco Giganti, Clare Allen, Mark Emberton, Caroline M Moore, Veeru Kasivisvanathan, PRECI- SION Study Group, et al. Prostate imaging quality (pi-qual): a new quality control scoring system for multiparametric magnetic resonance imaging of the prostate from the precision tr...

    2020

  36. [44]

    Artificial intelligence and radiologists at prostate cancer detection in mri—the pi-cai challenge

    Anindo Saha, Joeran Bosma, Jasper Twilt, Bram van Ginneken, Derya Yakar, Mattijs Elschot, Jeroen Veltman, Jurgen Fütterer, Maarten de Rooij, et al. Artificial intelligence and radiologists at prostate cancer detection in mri—the pi-cai challenge. In Medical Imaging with Deep L...

    2023

  37. [45]

    Deep residual learning for image recog- nition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recog- nition. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 770–778, 2016

    2016

  38. [46]

    Swin transformer: Hierarchical vision transformer using shifted windows

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shifted windows. InProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 10012–10022, 2021

    2021

  39. [47]

    An image is worth 16x16 words: Transformers for image recognition at scale

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition at...

    2021

  40. [48]

    Biomedclip: a multimodal biomedical founda- tion model pretrained from fifteen million scientific image-text pairs.arXiv preprint arXiv:2303.00915, 2023

    Sheng Zhang, Yanbo Xu, Naoto Usuyama, Han- wen Xu, Jaspreet Bagga, Robert Tinn, Sam Preston, Rajesh Rao, Mu Wei, Naveen Valluri, Cliff Wong, Matthew P Lungren, Tristan Naumann, and Hubert Pien. Biomedclip: a multimodal biomedical founda- tion model pretrained from fifteen mill...

    2023 arXiv

  41. [49]

    Omnirad: A radiological foundation model for multi-task medical image analysis.arXiv preprint arXiv:2602.04547, 2026

    Luca Zedda, Andrea Loddo, and Cecilia Di Ruberto. Omnirad: A radiological foundation model for multi-task medical image analysis.arXiv preprint arXiv:2602.04547, 2026

    2026

  42. [50]

    U-net: Convolutional networks for biomedi- cal image segmentation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U-net: Convolutional networks for biomedi- cal image segmentation. InMedical Image Comput- ing and Computer-Assisted Intervention (MICCAI), volume 9351 ofLNCS, pages 234–241. Springer, 2015

    2015

  43. [51]

    Kensho Hara, Hirokatsu Kataoka, and Yutaka Satoh. Can spatiotemporal 3d cnns retrace the history of 2d cnns and imagenet? InProceedings of the IEEE Conference on Computer Vision and Pattern Recog- nition (CVPR), pages 6546–6555, 2018

    2018

  44. [52]

    Efficientnet: Re- thinking model scaling for convolutional neural networks

    Mingxing Tan and Quoc V Le. Efficientnet: Re- thinking model scaling for convolutional neural networks. InProceedings of the 36th Interna- tional Conference on Machine Learning (ICML), volume 97 ofProceedings of Machine Learning Re- search, pages 6105–6114, 2019. 16

    2019

  45. [53]

    Inter-slice context residual learning for 3d med- ical image segmentation.IEEE Transactions on Medical Imaging, 40(2):661–672, 2020

    Jianpeng Zhang, Yutong Xie, Yan Wang, and Yong Xia. Inter-slice context residual learning for 3d med- ical image segmentation.IEEE Transactions on Medical Imaging, 40(2):661–672, 2020

    2020

  46. [54]

    Mri features of the normal prostatic pe- ripheral zone: the relationship between age and sig- nal heterogeneity on t2wi, dwi, and dce sequences

    Vlad Bura, Iztok Caglic, Ziga Snoj, Nikita Sushent- sev, Alexandra S Berghe, Andrew N Priest, and Tris- tan Barrett. Mri features of the normal prostatic pe- ripheral zone: the relationship between age and sig- nal heterogeneity on t2wi, dwi, and dce sequences. European Radiol...

    2021

  47. [55]

    Amr A Elfattah Hassan Gadalla, Marianne Abdal- lah Sabet, Wael Abdelgawwad Edesa, Manal Halim Wahba, and Sherif M Abolyazid. Diagnos- tic accuracy of the different pulse sequences of multi-parametric prostate mri in the diagnosis of prostate cancer in the peripheral and transi...

    2023

  48. [56]

    Maxime Deforche, Y Lefebvre, Romain Diamand, MA Bali, Marc Lemort, and Nicolas Coquelet. Im- proved diagnostic accuracy of readout-segmented echo-planar imaging for peripheral zone clinically significant prostate cancer: a retrospective 3t mri study.Scientific Reports, 14(1):3...

    2024

  49. [57]

    Explainable artificial intelligence (xai) in medical imaging: a systematic review of techniques, applications, and challenges.BMC Med- ical Imaging, 2026

    Fahad Ahmed, Naila Sammar Naz, Sunawar Khan, Ateeq Ur Rehman, Waleed M Ismael, and Muham- mad Adnan Khan. Explainable artificial intelligence (xai) in medical imaging: a systematic review of techniques, applications, and challenges.BMC Med- ical Imaging, 2026

    2026

  50. [58]

    Explainability enhanced object detec- tion transformer with feature disentanglement.IEEE Transactions on Image Processing, 33:6439–6454, 2024

    Wenlong Yu, Ruonan Liu, Dongyue Chen, and Qinghua Hu. Explainability enhanced object detec- tion transformer with feature disentanglement.IEEE Transactions on Image Processing, 33:6439–6454, 2024. 17

    2024

Pith tools

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