Pith. sign in

REVIEW 4 major objections 5 minor 66 references

CorePath: A Breast-Specialized Pathology Foundation Model for Core Needle Biopsy Diagnosis and Risk-Controlled Report Generation

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read A breast-specialized fine-tune of a pathology foundation model outperforms the base model on zero-shot breast core needle biopsy diagnosis and, with conformal risk control, releases zero non-breast hallucinations.

desk verdict Solid classification results with a credible breast-specialized fine-tune of PRISM, but the headline 'zero hallucinations' claim for CorePath-CRG is asserted rather than measured and needs an honest fix before the paper is accepted. read the letter →

arxiv 2608.03079 v1 pith:2GNWFJUQ submitted 2026-08-04 cs.CV cs.AIcs.LGstat.AP

Ting Yin , Danning Li , Chen Shu , Xiaoxia Yao , Boyu Fu , Yujing Chang , Tianyu Shi , Mengna Feng
show 12 more authors
Jie Chen Jing Fu Xiuli Xiao Tianlin Li Mumin Shao Jiaxin Bi Wenchuan Zhang Xiaoyan Wu Xiao Han Zhang Zhang Yuhao Yi Hong Bu
This is my paper · ORCID
classification cs.CVcs.AIcs.LGstat.AP
keywords breastcoreneedlebiopsypathologyfoundationmodelzero-shotclassificationwhole-slideimagingconformalpredictionrisk-controlledreportgenerationhallucinationreductionmultimodalfine-tuning
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 argues that a pathology foundation model specialized to breast core needle biopsy (CNB) can beat a general-purpose foundation model at zero-shot diagnosis and do so safely enough for clinical report generation. The authors fine-tune PRISM on 7,901 paired whole-slide images and diagnostic reports from two centers, then test it without task-specific retraining on six private cohorts and two public benchmarks. Across cancer detection, invasion assessment, and five-class histological subtyping, the specialized model consistently achieves higher weighted AUCs than the base model, with subtyping AUCs between 0.9526 and 0.9735. For report generation, the specialization cuts severe non-breast hallucinations from 30.1% to 2.8%, and a conformal risk-control extension eliminates them entirely among released reports.

What carries the argument

The central object is CorePath-CRG, a three-tier selective report generation pipeline built on the fine-tuned CorePath model. CorePath itself is produced by Perceiver-only fine-tuning: small adapter modules are inserted into the feed-forward sublayers of the Perceiver slide encoder, which aggregates patch-level features into a whole-slide representation, while the language encoder and all cross-modal components stay frozen. CorePath-CRG adds conformal calibration of diagnostic probabilities, a Self-Confidence Score that rates candidate reports without ground truth, a Judge Score used only during calibration, and a Learn-Then-Test procedure that selects a release threshold with a finite-sample guarantee on the risk of auto-releasing an unacceptable report. At inference, cases above the threshold go to trusted synthesis, cases below it with a confident conformal subtype fall back to subtype-only output, and cases with neither are deferred as Unknown for pathologist review.

What would settle it

Have a pathologist who was not involved in the study re-derive the five-class subtype labels directly from the slides of one entire private cohort, then rerun CorePath's zero-shot prompts and compare weighted AUCs; a meaningful drop below the reported 0.9526-0.9735 band would show that label extraction, not morphology, drove part of the result. Alternatively, on a deliberately shifted cohort, count how many auto-released CorePath-CRG reports a panel judges clinically unacceptable; if substantially more than the declared 10% risk bound fail, the exchangeability assumption behind the guarantee did not hold.

Watch

Extended reading notes

Core claim

The authors report that a breast-specialized pathology foundation model, obtained by fine-tuning only the Perceiver-based slide encoder of PRISM on paired CNB whole-slide images and diagnostic reports, consistently outperforms the unmodified PRISM model in zero-shot breast CNB classification across six private centers and two public benchmarks. The gains appear at every diagnostic level, from cancer detection to invasion assessment to five-class histological subtyping, with weighted subtyping AUCs of 0.9526-0.9735 on private cohorts and 0.7780, 0.8178, and 0.8252 on the BCNB and BRACS public tasks. The same model reduces severe non-breast hallucinations in generated reports from 30.1% to 2.8%, and the CorePath-CRG extension, which applies conformal subtype gating and Learn-Then-Test threshold calibration to a three-tier selective release policy, achieves zero non-breast hallucinations among released reports while preserving high-quality outputs when confidence is sufficient.

Load-bearing premise

The whole pipeline, including training labels, evaluation ground truth, and translated report text, depends on what language models extracted from the Chinese pathology reports, and only a fraction of that extracted text was manually checked; if those extractions are systematically wrong, the reported accuracy is measuring mistakes as ground truth.

Editorial extensions

If this is right

  • A breast-specialized pathology foundation model can be used for zero-shot hierarchical diagnosis, meaning new diagnostic label spaces can be defined by text prompts without training task-specific classifiers.
  • Fine-tuning only the visual encoder on paired WSI-report data preserves the language capabilities of the base model while anchoring its output to the breast pathology domain, cutting non-breast hallucinations by an order of magnitude.
  • Adding conformal gating and Learn-Then-Test risk control to report generation makes it possible to guarantee, with user-chosen confidence, that auto-released reports stay below a target risk level, while uncertain cases are explicitly deferred to pathologists.
  • The three-tier release policy separates diagnostic inference from output authorization, so a model can refuse to narrate a case it is unsure about while still returning a confident subtype label.
  • LLM-based Evaluation Scores validated against pathologist scores (Spearman rho 0.911, ICC(A,1) 0.843) can serve as a scalable proxy for clinical report quality in this setting.

Reading between the lines

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

  • Editorial inference: The same recipe, fine-tuning only the visual encoder of a multimodal foundation model on paired WSI-report data and then gating releases by conformal risk control, could transfer to other organ systems where core needle biopsy diagnosis is difficult, such as prostate or thyroid, although the paired data requirement would need to be met anew.
  • Editorial inference: The risk-control layer is largely independent of the underlying generative model, so Learn-Then-Test gating could be attached to any pathology report generator as a plug-in safety mechanism, not just to CorePath.
  • Editorial inference: Because the reported labels and English translations were produced by language models and only partly reviewed by humans, the cleanest external check is to re-label a full cohort by pathologists directly from slides and rerun the zero-shot evaluation; if the extracted labels were systematically biased, the reported accuracy numbers would shift.
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 / 5 minor

Summary. This manuscript introduces CorePath, a breast-specific multimodal pathology foundation model obtained by Perceiver-only fine-tuning of PRISM on 7901 paired breast core-needle-biopsy whole-slide images and diagnostic reports from two centers. The authors evaluate zero-shot classification across six private cohorts and two public benchmarks (BCNB, BRACS), reporting that CorePath consistently outperforms PRISM and TITAN on cancer detection, invasion assessment, and histological subtyping (weighted AUCs 0.9526-0.9735 on five-class subtyping). They further develop CorePath-CRG, a conformalized and Learn-Then-Test-gated report-generation framework with three-tier selective release (trusted synthesis, subtype-only fallback, and deferral), and claim zero non-breast hallucinations among released outputs. The paper includes pathologist validation of an LLM-based Evaluation Score (ICC(A,1)=0.843 on 144 reports) and reports improved Evaluation Scores and text-similarity metrics for CorePath-CRG.

Significance. If confirmed, the classification results are a strong demonstration that domain-specialized fine-tuning of a histopathology foundation model improves zero-shot generalization across institutions and public datasets; the use of bootstrap confidence intervals and multiple independent test cohorts is a notable strength. The report-generation risk-control architecture is a timely contribution, and the paper appropriately separates diagnosis from output authorization. However, the headline safety claim of a 'strict 0% hallucination rate' is not empirically supported as written, and the ground-truth labels depend on unvalidated LLM extraction. The statistical proof of the LTT guarantee also needs strengthening.

major comments (4)
  1. [2.6, 3.3, Abstract] The claimed 'strict 0% hallucination rate' for CorePath-CRG is not measured: Section 2.6 states that CorePath-CRG is excluded from the hallucination evaluation, and Figure 3a reports only PRISM and CorePath. The LTT guarantee in Supplementary A.5 controls the rate of Judge-Score-defined clinical risk on the SPH-2 calibration set, not the hallucination rate on the four external centers. Please measure the hallucination rate of released CorePath-CRG reports using the same Qwen3-based detector or a pathologist audit, report the observed count with a confidence interval, and revise the abstract and Section 3.3 accordingly.
  2. [2.2] The ground-truth labels for both fine-tuning and evaluation are derived from LLM extraction (DeepSeek-R1-Distill-Qwen-32B) and translation (OpenBioLLM-70B), with manual review only 'when needed'. The manuscript does not report the proportion of reports manually reviewed, inter-method agreement, or an error analysis. Since every reported AUC depends on these labels, a systematic extraction error would directly bias the classification results. Please provide a statistical validation of the label-extraction pipeline (e.g., a random sample audited by a pathologist, with agreement rates and a breakdown of discrepancies) or otherwise justify the reliability of the labels.
  3. [A.5.4, Theorem A.3] The proof sketch for the distribution-free risk-control guarantee is incomplete: the p-values p_k defined in Equation (11) are not shown to be super-uniform or monotone in k, and the claim that the nested structure of the acceptance sets A_k suffices to control FWER for the maximum-rejecting rule is not justified. The guarantee in Corollary A.4 is load-bearing for the safety claims of CorePath-CRG, so please provide a rigorous derivation, or state precisely which theorem in [17] is being invoked and verify that its monotonicity conditions hold for this choice of acceptance sets and risk indicators.
  4. [2.6, A.3, A.4] The LLM-based Evaluation Score used for the main report-quality comparison is computed by the DeepSeek API, the same model family used for the Self-Confidence Score and Judge Score inside CorePath-CRG. Although the Evaluation Score was validated against a pathologist on 144 reports (Table S18, ICC(A,1)=0.843), the validation subset is small and the possibility remains that the evaluator is biased toward outputs from the DeepSeek-based synthesis pipeline. Please report per-center agreement statistics for the pathologist validation, or compare models with an independent human-annotated subset across all centers, to rule out scoring bias as an explanation for the reported improvements.
minor comments (5)
  1. [References] Reference [16] contains a typo: 'Vladimir V ovk' should read 'Vladimir Vovk'.
  2. [Figure 3] Figure 3(a) shows hallucination rates only for PRISM and CorePath; the caption should state that CorePath-CRG is omitted and explain why.
  3. [2.2] Section 2.2 states that translations were manually reviewed 'when needed'; please specify the criteria and the number of reports audited.
  4. [Abstract] The abstract's 'zero non-breast hallucinations among released outputs' should be revised to reflect the actual measurement once provided; as written it overstates the evidence.
  5. [Table S20] Table S20 reports retained-sample metrics; consider adding a column with the N per center to the main text to make the selective-release behavior transparent.

Circularity Check

1 steps flagged · score 4.0 of 10

Diagnostic evaluation is externally grounded, but the CorePath-CRG 'strict 0% hallucination' headline is an unmeasured, design-level claim that is partially circular because released outputs are defined by the same gating score that is supposed to guarantee the result.

  1. other [Section 2.6 (Report Quality Assessment), Section 3.3 (Safety and Clinical Quality), Figure 3 caption, Supplementary Methods A.5]
    "However, evaluation for CorePath-CRG is excluded from the primary comparison to maintain a fair evaluation under identical unconstrained generation conditions. Furthermore, CorePath-CRG eliminates the residual errors entirely via constrained decoding, achieving a strict 0% hallucination rate among released outputs."

    Section 2.6 restricts the Qwen3-based hallucination measurement to PRISM and CorePath and explicitly excludes CorePath-CRG; Section 3.3 nevertheless asserts a strict 0% hallucination rate for CorePath-CRG. The risk-control guarantee in Supplementary Methods A.5 bounds the rate of Judge-Score-defined clinical risk on the SPH-2 calibration distribution, not the rate of non-breast hallucinations on SWH/WCH-2/WTH/SJH, and the Self-Confidence gate used to define released outputs explicitly scores 'breast relevance' and assigns near-zero scores to off-topic content.

full rationale

CorePath's diagnostic claims are non-circular: the model is fine-tuned only on WCH-1/SPH-1 WSI-report pairs and is evaluated zero-shot on temporally separated SPH-2, on other-institution cohorts (SWH, WCH-2, WTH, SJH, SZH), and on public BCNB/BRACS benchmarks; these test examples do not overlap the training set, and the comparison to PRISM/TITAN is external. The LTT/conformal machinery is standard external methodology and is not used to conjure the classification AUCs. The only material circularity is in the report-generation safety headline. The manuscript itself discloses that CorePath-CRG was excluded from the Qwen3 hallucination measurement 'to maintain a fair evaluation under identical unconstrained generation conditions,' then asserts 'strict 0% hallucination rate among released outputs' in Section 3.3 and the abstract. Because the release gate's Self-Confidence Score explicitly rewards 'breast relevance' and penalizes off-topic reports, 'released outputs' are precisely the outputs that the system's own judge deems breast-relevant; without an independent detector or pathologist audit on released CorePath-CRG outputs, the 0% figure is an assertion about the system's design rather than a measured result. The LTT guarantee (Supplementary Methods A.5) is a bound on Judge-Score risk on the calibration distribution and does not imply zero hallucination on shifted test cohorts, a limitation the Discussion itself concedes. The score is moderate (4) rather than higher because the central diagnostic and comparative report-generation results are externally grounded.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

No new physical entities are introduced. The model itself is a configuration of existing components. The load-bearing assumptions are exchangeability, the accuracy of LLM-extracted labels/translations, and the choice of free hyperparameters in the release rule.

free parameters (3)
  • lambda (margin weight in fused score) = not reported
    Appears in Eq. (2) as S_i = P_i + lambda*M_i + epsilon_i and is required for threshold selection, but no value is provided in the visible text.
  • tau_judge (clinical quality threshold for risk indicator) = not reported
    Defines R(z)=I(J(z)<tau_judge) in Eq. (1); the risk-control guarantee is with respect to this threshold, but its value is not given.
  • alpha and delta (risk and confidence levels) = 0.10, 0.10
    Chosen by hand as default policy values in A.5.6; the guarantee depends on these choices.
assumptions (4)
  • domain assumption Calibration and test samples are exchangeable (Assumption A.1)
    Required for the Learn-Then-Test finite-sample guarantee; acknowledged in the limitations as imperfect for cross-center deployment.
  • domain assumption LLM-extracted diagnostic labels are accurate enough for supervision and evaluation
    Section 2.2 uses DeepSeek-R1-Distill-Qwen-32B to extract structured diagnostic information without reporting a validation study of the extraction.
  • domain assumption LLM-based English translations preserve the diagnostic content of the original Chinese reports
    Translations by OpenBioLLM-70B were only manually reviewed 'when needed', leaving the risk of undetected semantic drift.
  • standard math Clopper-Pearson bound and binomial model are valid under exchangeability
    Used in Theorem A.2 and Theorem A.3; standard statistical results cited from the literature.

how reviews work

0 comments
Cite this review

Pith. "Pith review of CorePath: A Breast-Specialized Pathology Foundation Model for Core Needle Biopsy Diagnosis and Risk-Controlled Report Generation." pith.science (2026). https://pith.science/paper/2GNWFJUQ

@misc{pith2026260803079,
  author       = {Pith},
  title        = {Pith review of: CorePath: A Breast-Specialized Pathology Foundation Model for Core Needle Biopsy Diagnosis and Risk-Controlled Report Generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2GNWFJUQ}},
  note         = {Machine review of arXiv:2608.03079}
}
read the original abstract

Breast core needle biopsy (CNB) is central to breast cancer diagnosis yet remains challenging because limited tissue sampling, lesion heterogeneity, and subtle morphologic overlap can obscure subtype distinctions. We developed CorePath, a breast-specialized multimodal pathology foundation model fine-tuned from PRISM using 7901 paired CNB whole-slide images and diagnostic reports from two centers. Evaluated across six CNB cohorts and two public breast pathology benchmarks without task-specific retraining, CorePath consistently outperformed PRISM across cancer detection, invasion assessment, and histological subtyping. It achieved weighted area under the receiver operating characteristic curves (AUCs) of 0.9526-0.9735 for five-class CNB histological subtyping across private centers. On public benchmarks, CorePath outperformed leading pathology foundation models, achieving the highest weighted AUCs of 0.7780 for BCNB invasive carcinoma subtyping, 0.8178 for BRACS lesion stratification, and 0.8252 for BRACS fine-grained classification. In report generation, CorePath reduced the overall non-breast hallucinations from 30.1% to 2.8%, demonstrating improved domain fidelity after breast-specific adaptation. CorePath-CRG further combined conformal subtype-confidence gating with Learn-Then-Test risk control to enable selective report release, subtype-level fallback, and deferral. CorePath-CRG achieved zero non-breast hallucinations among released outputs and showed the strongest overall performance in pathologist-validated LLM-based Evaluation Scores and quantitative report-generation metrics across most centers. These results demonstrate that domain-specialized foundation models with statistical risk control offer a promising approach for accurate breast CNB diagnosis and reliable report generation.

Figures

Figures reproduced from arXiv: 2608.03079 by the authors.

Figure 1
Figure 1. Overview of CorePath and CorePath-CRG for breast CNB diagnosis and conformal￾ized report generation. CorePath was developed by fine-tuning the PRISM pathology foundation model using paired CNB WSIs and pathology reports from two institutions. The resulting CorePath model was used for zero-shot diagnostic classification across predefined breast CNB tasks. For report generation, CorePath produced diagnostic probabilit… view at source ↗
Figure 2
Figure 2. Zero-shot diagnostic classification performance across independent private and public breast pathology cohorts. (a) Cancer detection on independent private CNB cohorts. (b) Invasion assessment on independent private CNB cohorts. (c) Histological subtyping on independent private CNB cohorts. (d) Public benchmark evaluation on BCNB and BRACS, where the BCNB 3-class task corresponds to invasive carcinoma subtype catego… view at source ↗
Figure 3
Figure 3. Hallucination reduction and LLM-based Evaluation Score for generated pathology reports. (a) Hallucination rates of unconstrained PRISM and CorePath outputs. CorePath-CRG is omitted because its 0% rate reflects risk-controlled selective outputs and is not directly comparable. (b-e) Violin plots of LLM-based Evaluation Score distributions for PRISM, CorePath, and CorePath￾CRG in SWH, WCH-2, WTH, and SJH. Dashed horizo… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Representative report-generation examples across different models. For each case, the left panels show the whole-slide image and a magnified region of interest, and the green panel summarizes the reference diagnosis. Model-generated outputs are shown below each case. R…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

66 extracted references · 44 canonical work pages

  1. [17]

    Learn then test: Calibrating predictive algorithms to achieve risk control.The Annals of Applied Statistics, 19(2):1641–1662, 2025

    Anastasios N Angelopoulos, Stephen Bates, Emmanuel J Candès, Michael I Jordan, and Lihua Lei. Learn then test: Calibrating predictive algorithms to achieve risk control.The Annals of Applied Statistics, 19(2):1641–1662, 2025

    work page 2025

  2. [1]

    Breast cancer, version 3.2024, NCCN clinical practice guidelines in oncology.Journal of the National Comprehensive Cancer Network, 22(5):331–357, 2024

    William J Gradishar, Meena S Moran, Jame Abraham, Vandana Abramson, Rebecca Aft, Doreen Agnese, Kimberly H Allison, Bethany Anderson, Janet Bailey, Harold J Burstein, et al. Breast cancer, version 3.2024, NCCN clinical practice guidelines in oncology.Journal of the National Comprehensive Cancer Network, 22(5):331–357, 2024

    work page 2024

  3. [2]

    Siegel, Isabelle Soerjomataram, and Ahmedin Jemal

    Freddie Bray, Mathieu Laversanne, Hyuna Sung, Jacques Ferlay, Rebecca L. Siegel, Isabelle Soerjomataram, and Ahmedin Jemal. Global cancer statistics 2022: Globocan estimates of incidence and mortality worldwide for 36 cancers in 185 countries.CA: A Cancer Journal for Clinicians, 74:229 – 263, 2024

    work page 2022

  4. [3]

    Histological diagnosis of unprocessed breast core-needle biopsy via stimulated raman scattering microscopy and multi-instance learning

    Yifan Yang, Zhijie Liu, Jing Huang, Xiangjie Sun, Jianpeng Ao, Bin Zheng, Wanyuan Chen, Zhiming Shao, Hao Hu, Yinlong Yang, et al. Histological diagnosis of unprocessed breast core-needle biopsy via stimulated raman scattering microscopy and multi-instance learning. Theranostics, 13(4):1342, 2023

    work page 2023

  5. [4]

    Precision pathology as applied to breast core needle biopsy evaluation: implications for management.Modern Pathology, 34:48–61, 2021

    Laura C Collins. Precision pathology as applied to breast core needle biopsy evaluation: implications for management.Modern Pathology, 34:48–61, 2021

    work page 2021

  6. [5]

    Diagnostic pitfalls in needle core biopsy of the breast: an update.Diagnostic Histopathology, 2025

    Abeer M Shaaban. Diagnostic pitfalls in needle core biopsy of the breast: an update.Diagnostic Histopathology, 2025

    work page 2025

  7. [6]

    Breast core needle biopsy: issues and controversies.Modern Pathology, 23:S36–S45, 2010

    Michael Bilous. Breast core needle biopsy: issues and controversies.Modern Pathology, 23:S36–S45, 2010

    work page 2010

  8. [7]

    Assessing intraductal proliferations in breast core needle biopsies.Diagnostic Histopathology, 24(2):49–57, 2018

    Liza M Quintana and Laura C Collins. Assessing intraductal proliferations in breast core needle biopsies.Diagnostic Histopathology, 24(2):49–57, 2018

    work page 2018

Show all 66 references
  1. [8]

    Self-supervised stain normalization empowers privacy-preserving and model generalization in digital pathology.npj Digital Medicine, 2025

    Jianhang Wang, Jiahui Yu, Haixu Yang, Yunqi Zhu, Lei Jiang, Xiaoxiao Li, Jing Zhang, and Yingke Xu. Self-supervised stain normalization empowers privacy-preserving and model generalization in digital pathology.npj Digital Medicine, 2025

    2025

  2. [9]

    A visual–language foundation model for pathology image analysis using medical twitter.Nature medicine, 29(9):2307–2316, 2023

    Zhi Huang, Federico Bianchi, Mert Yuksekgonul, Thomas J Montine, and James Zou. A visual–language foundation model for pathology image analysis using medical twitter.Nature medicine, 29(9):2307–2316, 2023

    2023

  3. [10]

    A visual-language foundation model for computational pathology.Nature Medicine, 30:863–874, 2024

    Ming Y Lu, Bowen Chen, Drew FK Williamson, Richard J Chen, Ivy Liang, Tong Ding, Guil- laume Jaume, Igor Odintsov, Long Phi Le, Georg Gerber, et al. A visual-language foundation model for computational pathology.Nature Medicine, 30:863–874, 2024

    2024

  4. [11]

    Prism: A multi-modal generative foundation model for slide-level histopathology.arXiv preprint arXiv:2405.10254, 2024

    George Shaikovski, Adam Casson, Kristen Severson, Eric Zimmermann, Yi Kan Wang, Jeremy D Kunz, Juan A Retamero, Gerard Oakley, David Klimstra, Christopher Kanan, et al. Prism: A multi-modal generative foundation model for slide-level histopathology.arXiv preprint arXiv:2405.10...

    2024 arXiv

  5. [12]

    Estimating diagnostic uncertainty in artificial intelligence assisted pathology using conformal prediction.Nature communications, 13(1):7761, 2022

    Henrik Olsson, Kimmo Kartasalo, Nita Mulliqi, Marco Capuccini, Pekka Ruusuvuori, Hema- mali Samaratunga, Brett Delahunt, Cecilia Lindskog, Emiel AM Janssen, Anders Blilie, et al. Estimating diagnostic uncertainty in artificial intelligence assisted pathology using conformal pr...

    2022

  6. [13]

    Vl-adapter: Parameter-efficient transfer learning for vision-and-language tasks

    Yi-Lin Sung, Jaemin Cho, and Mohit Bansal. Vl-adapter: Parameter-efficient transfer learning for vision-and-language tasks. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 5227–5237, June 2022

    2022

  7. [14]

    A multimodal whole-slide foundation model for pathology.Nature medicine, pages 1–13, 2025

    Tong Ding, Sophia J Wagner, Andrew H Song, Richard J Chen, Ming Y Lu, Andrew Zhang, Anurag J Vaidya, Guillaume Jaume, Muhammad Shaban, Ahrong Kim, et al. A multimodal whole-slide foundation model for pathology.Nature medicine, pages 1–13, 2025

    2025

  8. [15]

    Se- mantic consistency-based uncertainty quantification for factuality in radiology report generation

    Chenyu Wang, Weichao Zhou, Shantanu Ghosh, Kayhan Batmanghelich, and Wenchao Li. Se- mantic consistency-based uncertainty quantification for factuality in radiology report generation. InFindings of the Association for Computational Linguistics: NAACL 2025, pages 1739–1754, 2025

    2025

  9. [16]

    Springer, 2005

    Vladimir V ovk, Alexander Gammerman, and Glenn Shafer.Algorithmic learning in a random world. Springer, 2005

    2005

  10. [18]

    Xiaoge Zhang, Tao Wang, Chao Yan, Fedaa Najdawi, Kai Zhou, Yuan Ma, Yiu-ming Cheung, Maximus C. F. Yeung, and Bradley A. Malin. Implementing trust in non-small cell lung cancer diagnosis with a conformalized uncertainty-aware ai framework.Nature Biomedical Engineering, 2026

    2026

  11. [19]

    Harrison, Gabriel Skogberg, Sonya Jackson, Markus Fridén, Johan Karlsson, Ola Spjuth, and Carolina Wählby

    Håkan Wieslander, Philip J. Harrison, Gabriel Skogberg, Sonya Jackson, Markus Fridén, Johan Karlsson, Ola Spjuth, and Carolina Wählby. Deep learning with conformal prediction for hierarchical analysis of large-scale whole-slide tissue images.IEEE Journal of Biomedical and Heal...

    2021

  12. [20]

    Mitosis domain generalization in histopathology images-the MIDOG challenge.Medical Image Analysis, 84:102699, 2023

    Marc Aubreville, Nikolas Stathonikos, Christof A Bertram, Robert Klopfleisch, Natalie Ter Ho- eve, Francesco Ciompi, Frauke Wilm, Christian Marzahl, Taryn A Donovan, Andreas Maier, et al. Mitosis domain generalization in histopathology images-the MIDOG challenge.Medical Image ...

    2023

  13. [21]

    Diagnostic assessment of deep learning algorithms for detection of lymph node metastases in women with breast cancer.Jama, 318(22):2199–2210, 2017

    Babak Ehteshami Bejnordi, Mitko Veta, Paul Johannes van Diest, Bram Van Ginneken, Nico Karssemeijer, Geert Litjens, Jeroen AWM Van Der Laak, CAMELYON16 consortium, Meyke Hermsen, Quirine F Manson, et al. Diagnostic assessment of deep learning algorithms for detection of lymph ...

    2017

  14. [22]

    Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning, 2025

    DeepSeek-AI. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning, 2025

    2025

  15. [23]

    Openbiollms: Advancing open-source large lan- guage models for healthcare and life sciences

    Malaikannan Sankarasubbu Ankit Pal. Openbiollms: Advancing open-source large lan- guage models for healthcare and life sciences. https://huggingface.co/aaditya/ OpenBioLLM-Llama3-70B, 2024

    2024

  16. [24]

    Perceiver: General perception with iterative attention

    Andrew Jaegle, Felix Gimeno, Andy Brock, Oriol Vinyals, Andrew Zisserman, and Joao Carreira. Perceiver: General perception with iterative attention. InInternational conference on machine learning, pages 4651–4664. PMLR, 2021

    2021

  17. [25]

    Overcoming catastrophic forgetting in neural networks.Proceedings of the national academy of sciences, 114(13):3521–3526, 2017

    James Kirkpatrick, Razvan Pascanu, Neil Rabinowitz, Joel Veness, Guillaume Desjardins, Andrei A Rusu, Kieran Milan, John Quan, Tiago Ramalho, Agnieszka Grabska-Barwinska, et al. Overcoming catastrophic forgetting in neural networks.Proceedings of the national academy of scienc...

    2017

  18. [26]

    Vl-adapter: Parameter-efficient transfer learning for vision-and-language tasks

    Yi-Lin Sung, Jaemin Cho, and Mohit Bansal. Vl-adapter: Parameter-efficient transfer learning for vision-and-language tasks. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 5227–5237, 2022

    2022

  19. [27]

    Parameter-efficient transfer learning for NLP

    Neil Houlsby, Andrei Giurgiu, Stanislaw Jastrzebski, Bruna Morrone, Quentin De Laroussilhe, Andrea Gesmundo, Mona Attariyan, and Sylvain Gelly. Parameter-efficient transfer learning for NLP. InInternational conference on machine learning, pages 2790–2799. PMLR, 2019

    2019

  20. [28]

    Learning to prompt for vision-language models.International Journal of Computer Vision, 130(9):2337–2348, 2022

    Kaiyang Zhou, Jingkang Yang, Chen Change Loy, and Ziwei Liu. Learning to prompt for vision-language models.International Journal of Computer Vision, 130(9):2337–2348, 2022

    2022

  21. [29]

    Learning transferable visual models from natural language supervision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervision. InInternational conference on machine learning, pag...

    2021

  22. [30]

    Predicting axillary lymph node metastasis in early breast cancer using deep learning on primary tumor biopsy slides.Frontiers in oncology, 11:759007, 2021

    Feng Xu, Chuang Zhu, Wenqi Tang, Ying Wang, Yu Zhang, Jie Li, Hongchuan Jiang, Zhongyue Shi, Jun Liu, and Mulan Jin. Predicting axillary lymph node metastasis in early breast cancer using deep learning on primary tumor biopsy slides.Frontiers in oncology, 11:759007, 2021

    2021

  23. [31]

    Bracs: A dataset for breast carcinoma subtyping in h&e histology images.Database, 2022:baac093, 2022

    Nadia Brancati, Anna Maria Anniciello, Pushpak Pati, Daniel Riccio, Giosuè Scognamiglio, Guillaume Jaume, Giuseppe De Pietro, Maurizio Di Bonito, Antonio Foncubierta, Gerardo Botti, et al. Bracs: A dataset for breast carcinoma subtyping in h&e histology images.Database, 2022:b...

    2022

  24. [32]

    Qwen3 technical report, 2025

    Qwen Team. Qwen3 technical report, 2025

    2025

  25. [33]

    Modaltune: Fine-tuning slide-level foundation models with multi-modal information for multi-task learning in digital pathology

    Vishwesh Ramanathan, Tony Xu, Pushpak Pati, Faruk Ahmed, Maged Goubran, and Anne L Martel. Modaltune: Fine-tuning slide-level foundation models with multi-modal information for multi-task learning in digital pathology. InProceedings of the IEEE/CVF International Conference on ...

    2025

  26. [34]

    Problematic issues in breast core needle biopsies.Modern Pathology, 32:71–76, 2019

    Stuart J Schnitt. Problematic issues in breast core needle biopsies.Modern Pathology, 32:71–76, 2019

    2019

  27. [35]

    Improving radiology report generation systems by removing hallucinated references to non-existent priors

    Vignav Ramesh, Nathan A Chi, and Pranav Rajpurkar. Improving radiology report generation systems by removing hallucinated references to non-existent priors. InMachine Learning for Health, pages 456–473. PMLR, 2022

    2022

  28. [36]

    A framework to assess clinical safety and hallucination rates of llms for medical text summarisation.NPJ digital medicine, 8(1):274, 2025

    Elham Asgari, Nina Montaña-Brown, Magda Dubois, Saleh Khalil, Jasmine Balloch, Joshua Au Yeung, and Dominic Pimenta. A framework to assess clinical safety and hallucination rates of llms for medical text summarisation.NPJ digital medicine, 8(1):274, 2025

    2025

  29. [37]

    Second opinion needed: communicating uncertainty in medical machine learning.NPJ Digital Medicine, 4(1):4, 2021

    Benjamin Kompa, Jasper Snoek, and Andrew L Beam. Second opinion needed: communicating uncertainty in medical machine learning.NPJ Digital Medicine, 4(1):4, 2021

    2021

  30. [38]

    Conformal prediction: A gentle introduction

    Anastasios N Angelopoulos, Stephen Bates, et al. Conformal prediction: A gentle introduction. Foundations and trends® in machine learning, 16(4):494–591, 2023

    2023

  31. [39]

    Confor- mal risk control.arXiv preprint arXiv:2208.02814, 2022

    Anastasios N Angelopoulos, Stephen Bates, Adam Fisch, Lihua Lei, and Tal Schuster. Confor- mal risk control.arXiv preprint arXiv:2208.02814, 2022

    2022 arXiv

  32. [40]

    A gentle introduction to conformal prediction and distribution-free uncertainty quantification.arXiv preprint arXiv:2107.07511, 2021

    Anastasios N Angelopoulos and Stephen Bates. A gentle introduction to conformal prediction and distribution-free uncertainty quantification.arXiv preprint arXiv:2107.07511, 2021

    2021 arXiv

  33. [41]

    Uncertainty sets for image classifiers using conformal prediction.arXiv preprint arXiv:2009.14193, 2020

    Anastasios Angelopoulos, Stephen Bates, Jitendra Malik, and Michael I Jordan. Uncertainty sets for image classifiers using conformal prediction.arXiv preprint arXiv:2009.14193, 2020

    2009 arXiv

  34. [42]

    Springer, 2005

    Erich Leo Lehmann and Joseph P Romano.Testing statistical hypotheses. Springer, 2005

    2005

  35. [43]

    Conformal prediction under covariate shift.Advances in neural information processing systems, 32, 2019

    Ryan J Tibshirani, Rina Foygel Barber, Emmanuel Candes, and Aaditya Ramdas. Conformal prediction under covariate shift.Advances in neural information processing systems, 32, 2019

    2019

  36. [44]

    The use of confidence or fiducial limits illustrated in the case of the binomial.Biometrika, 26(4):404–413, 1934

    Charles J Clopper and Egon S Pearson. The use of confidence or fiducial limits illustrated in the case of the binomial.Biometrika, 26(4):404–413, 1934. 19 Supplementary A. Supplementary Methods A.1. Implementation Details Hardware and Training EnvironmentAll fine-tuning experi...

    1934

  37. [45]

    Breast relevance -- do the reports clearly discuss breast tissue or breast disease?

  38. [46]

    Diagnostic completeness -- do they contain a meaningful clinical conclusion?

  39. [47]

    Consistency -- do the reports agree with each other on malignancy and subtype ?

  40. [48]

    Information alignment -- do the generated reports align with the reference classifier?

  41. [49]

    reference classifier unknown -- skipped

    Linguistic coherence -- are the reports written as coherent medical prose? Scoring guidance (use the full continuous range, avoid rounding to 0.1 steps): - Reports that are relevant, complete, consistent, and reference classifier output-aligned -> near 1.00 - Partial relevance...

  42. [50]

    GT Alignment -- do the AI reports correctly identify the core malignancy and histological subtype stated in the GT translations?

  43. [51]

    Inter-Report Consistency -- do the AI reports agree with each other on the core diagnosis?

  44. [52]

    Figure S2: System and User prompts used for calculating the Judge Score via the LLM auditor

    Linguistic Coherence -- are the AI reports written as coherent, professional medical prose? Scoring guidance (use the full continuous range, avoid rounding to 0.1 steps): - Perfect GT alignment, high consistency, and coherent -> near 1.00 - Minor GT mismatches (e.g., subtype a...

  45. [53]

    ASynthesizer Agentintegrates the candidate reports and the classifier’s subtype into a single, professional diagnostic statement for auto-release

    Trusted Synthesis (Si≥ ˆλthresh):The reports are deemed clinically safe. ASynthesizer Agentintegrates the candidate reports and the classifier’s subtype into a single, professional diagnostic statement for auto-release

  46. [54]

    The system discards the LLM-generated text and falls back to outputting only the classifier’s subtype label, prioritizing safety over narrative detail

    Confident Fallback (Si < ˆλthresh, butMi is sufficient):The narrative reports are deemed potentially hallucinated or inconsistent, but the upstream classifier remains confident. The system discards the LLM-generated text and falls back to outputting only the classifier’s subty...

  47. [55]

    Human-in-the-Loop Rejection (Both uncertain):The system outputs an “Unknown” status, strictly intercepting the case and flagging it for mandatory expert pathological review. LLM Agents and Prompt Design.Both the Self-Confidence Scoring and Synthesizer agents are powered by the...

  48. [56]

    Invasive carcinoma, favor invasive lobular carcinoma

  49. [57]

    Fibroadenoma with focal calcification

  50. [58]

    Figure S3: Prompt for the Synthesizer Agent used to generate the final auto-released clinical diagnosis under the trusted pathway

    The breast tissue shows extensive acute and chronic inflammatory cell infiltration. Figure S3: Prompt for the Synthesizer Agent used to generate the final auto-released clinical diagnosis under the trusted pathway. Response Parsing and Execution Guarantee.The system employs st...

  51. [59]

    Clinical Factuality & Honesty (40%):Accuracy of medical entities and strict adherence to the honesty priority

  52. [60]

    Clinical Completeness (30%):Coverage of essential diagnostic information present in the ground truth, including malignancy status, histological subtype, grade when available, and relevant morphologic findings

  53. [61]

    Logical Consistency (20%):Internal coherence, ensuring the microscopic description aligns with the final diagnosis

  54. [62]

    Professionalism & Fluency (10%):Use of standard breast pathology terminology and coherent medical prose. LLM Auditor and Prompt Design.We utilize the DeepSeek API with a temperature of 0.0 and an increased maximum token limit (1024 tokens) to allow for more detailed analytical...

  55. [63]

    Strictly apply the honesty priority

    Clinical Factuality & Honesty (40%): Accuracy of stated diagnostic information. Strictly apply the honesty priority

  56. [64]

    Do not penalize omission of information that is not present in the ground truth or not assessable on core needle biopsy

    Diagnostic Coverage (30%): Coverage of essential diagnostic information present in the ground truth, including malignancy status, histological subtype, grade when available, and relevant morphologic findings. Do not penalize omission of information that is not present in the g...

  57. [65]

    Logical Consistency (20%): Internal coherence (e.g., microscopic description aligns with final diagnosis)

  58. [66]

    [Scoring Guidance] - Near 1.00: Accurate, honest, complete, and professional

    Professionalism & Fluency (10%): Standard breast pathology terminology and coherent medical prose. [Scoring Guidance] - Near 1.00: Accurate, honest, complete, and professional. - 0.60-0.89: Minor omissions or slight terminology flaws, but NO critical factual errors. - 0.20-0.5...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.