Pith. sign in

REVIEW 2 major objections 4 minor 88 references

Label-Free Whole Slide Virtual Multi-Staining Using Dual-Excitation Photon Absorption Remote Sensing Microscopy

T0 review · 2 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read One label-free dual-excitation PARS scan can generate virtual H&E, Masson's trichrome, PAS, and JMS stains that masked pathologists could not reliably distinguish from chemical stains.

desk verdict Useful dual-excitation PARS multi-stain virtual staining with solid held-out WSI metrics; the 'indistinguishable from chemical' claim however rests on a tiny, biased pathologist survey and an unaddressed UV-pre-exposure control. read the letter →

arxiv 2509.05085 v1 pith:ALKER3VT submitted 2025-09-05 physics.optics

James E.D. Tweel , Benjamin R. Ecclestone , James A. Tummon Simmons , Parsin Haji Reza This is my paper · ORCID
classification physics.optics
keywords virtualstaininglabel-freehistologyphotonabsorptionremotesensingdual-excitationmicroscopyRegGANultravioletwholeslideimagingdigitalpathology
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 reports a label-free microscopy system that produces virtual copies of four histochemical stains—H&E, Masson's trichrome, periodic acid–Schiff, and Jones methenamine silver—from a single scan of an unstained tissue section. The new element is a second ultraviolet wavelength, 355 nm, interlaced with the established 266 nm source so both wavelengths interrogate the tissue in one pass. The 355 nm channel contributes hemoglobin and melanin absorption plus collagen and elastin fluorescence, while 266 nm continues to provide nuclear contrast; together these four contrast channels give a generative model the cues it needs to reproduce each stain's colors and structures. Using the RegGAN framework, which embeds a learned registration network to tolerate imperfect alignment, the authors generate whole-slide virtual stains on tissue never seen during training. In a masked evaluation, three pathologists scored virtual and chemical images at the same diagnostic quality and could not reliably determine which were virtual.

What carries the argument

The load-bearing object is the dual-excitation PARS contrast set: interlaced 266 nm and 355 nm pulses, each producing a non-radiative absorption image (transient modulation of a 405 nm probe beam) and a radiative autofluorescence image. These four channels are fed to a ResNet encoder–decoder generator, which outputs an RGB virtual stain. Training uses RegGAN, a supervised scheme that inserts a registration network between the generated image and the chemically stained target; the registration network predicts a deformation field that aligns the virtual output to the real stain before the L1 and adversarial losses are computed. That built-in alignment is what makes pixel-level supervision usa

What would settle it

Stain a set of paired sections where one half of each slide is shielded from the UV beams and the other half is scanned as usual, then measure with a molecularly specific assay (e.g., antibody binding, lectin histochemistry, RNA integrity, or Congo red for amyloid) whether scanned regions deviate from shielded controls. A simpler version: quantitate stain intensity in adjacent imaged and never-imaged areas of the same slide; if they differ systematically, the ground truth for training is compromised.

Watch

Extended reading notes

Core claim

On its own terms, the paper claims the first demonstration of PARS virtual staining beyond H&E, generating Masson's trichrome, PAS, and Jones' silver from the same label-free scan used for H&E. The central discovery is that adding 355 nm UVA excitation to the established 266 nm UVC source provides biologically distinct absorption contrasts—melanin and hemoglobin in the non-radiative channel, collagen and elastin fluorescence in the radiative channel—and that these complementary cues, combined with 266 nm nuclear contrast, are enough for a supervised image-translation model to reconstruct the color and distribution of four separate chemical stains on unseen whole slides. The evidence includes

Load-bearing premise

The training and validation design assumes the UV pulses do not change the molecules that the post-imaging chemical stains bind to; the paper checks this only by visual inspection of stained slides, not by quantitative comparison of imaged and never-imaged regions.

Editorial extensions

If this is right

  • One label-free scan can provide virtual H&E, Masson's trichrome, PAS, and JMS on the same section, so multiple stains no longer require cutting adjacent sections.
  • The 355 nm channel carries the specific cues needed for correct collagen, RBC, melanin, and fungal-hyphae staining; ablation metrics show dual excitation beats either wavelength alone for every tissue–stain combination.
  • RegGAN's learned deformation field reduces dependence on perfect registration, lowering the curation burden for paired cross-modality data.
  • Because the section remains unstained until after imaging, the same slide can later receive chemical stains or downstream assays.
  • The paper makes the paired chemical and virtual whole-slide images publicly available, allowing direct comparison on unseen tissue.

Reading between the lines

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

  • Beyond the paper: the 355 nm non-radiative channel's sensitivity to hemoglobin and melanin suggests this input could generalize to virtual stains whose diagnostic signal is pigment-based (e.g., iron, hemosiderin) or to fresh-tissue applications such as surgical margin assessment.
  • Beyond the paper: the paper's non-destructive claim rests on visual inspection; a quantitative comparison of imaged versus shielded regions with molecular probes (antibodies, lectins, DNA/RNA integrity metrics) would be the natural certification experiment.
  • Beyond the paper: the same RegGAN-plus-dual-excitation pipeline could be retrained for additional special stains (Congo red, Prussian blue, elastin stains) since the four-channel input already captures the relevant endogenous absorbers; this is an untested extension.
  • Beyond the paper: the masked evaluation used 20 images and three pathologists; translating 'not reliably distinguishable' into a deployment-grade claim would require a larger, multi-site reader study.
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

2 major / 4 minor

Summary. The paper reports a dual-excitation (266 nm and 355 nm) PARS microscope and uses a RegGAN image-translation framework to generate virtual H&E, Masson's trichrome, PAS, and JMS stains from a single label-free whole-slide scan. The authors claim that 355 nm excitation adds complementary contrast for stroma, RBCs, melanin, and fungal elements; that dual-excitation input improves quantitative similarity metrics over single-wavelength inputs; that RegGAN outperforms Pix2Pix and CycleGAN on held-out whole-slide images; and that a masked evaluation by three pathologists showed comparable diagnostic quality and inability to distinguish chemical from virtual stains. The paper positions the system as a non-destructive multi-stain alternative that preserves tissue for downstream assays, with real and virtual WSI pairs made publicly available.

Significance. If the claims are correct, this would be a practically useful advance in label-free digital pathology: a single scan can produce multiple specialized stains without consuming additional sections, with publicly released whole-slide data for benchmarking. The study has notable strengths: tests are performed on entirely held-out WSIs, the datasets are large (tens of thousands of patches per tissue–stain combination), the ablation and model-comparison results are reported per dataset as well as aggregated, and the use of RegGAN to handle residual misalignment is well motivated. The quantitative improvements from dual excitation and from RegGAN over baselines are credible. However, the two headline claims — diagnostic equivalence to chemical staining, and non-destructive preservation of tissue — rest on evidence that is currently under-powered or indirect: a 10-pair, three-pathologist survey with a strong response bias, and only visual inspection for UV-induced tissue alteration. These gaps are real but addressable with additional experiments and analysis.

major comments (2)
  1. [Sec. 3.1 and Sec. 3.2] The paper's ground-truth validity and its 'non-destructive' claim depend on the assumption that 163 pJ at 266 nm and 808 pJ at 355 nm do not alter the molecular targets of the subsequent chemical stains. The only support offered is 'no tissue damage visible in any of the stained samples' (Sec. 3.1). This is insufficient: UV exposure can photochemically modify nucleic acids, proteins, and glycans without producing visible morphological changes. Because the same section is PARS-imaged first and then chemically stained, any UV-induced change in stain-target availability would corrupt both the RegGAN training targets and the pathologist comparison reference. I recommend a quantitative control, e.g., comparing stain optical density or a stain-specific signal in previously imaged versus non-imaged regions of the same section, or comparing imaged sections against adjacent unimaged sections. Thi
  2. [Sec. 2.5 and Table 2] The pathologist equivalence conclusion is drawn from 10 image pairs and three readers (30 responses per category). The response pattern is dominated by a 'chemical' prior: 18 of 30 virtual images and 17 of 30 chemical images were marked as chemical, while only one image of each type was marked virtual (12 and 11 uncertain). This does not demonstrate that pathologists 'could not reliably distinguish' the two; it is equally consistent with a tendency to default to 'chemical' when uncertain. A forced-choice design with known prevalence, or an analysis accounting for response bias, is needed before the abstract's strong claim is supportable. In addition, diagnostic-quality equivalence (DQ 2.600 vs 2.667) is reported without any statistical test or concordance measure. This concern affects the paper's central diagnostic-equivalence claim.
minor comments (4)
  1. [Sec. 3.5] In the paragraph describing Figure 8, the text refers to 'poorly defined glomerulus basement membranes in the JMS-stained kidney (Figure 8d)', but according to the caption, Figure 8d is Masson's trichrome and Figure 8e is JMS. Please correct the cross-reference.
  2. [Table 1 and Sec. 2.3] The paper states that 'multiple WSIs were held out entirely for testing' but does not report the number of held-out WSIs per tissue–stain combination. For reproducibility, please list the WSI counts in Table 1 or in the data-availability section.
  3. [Throughout] The manuscript contains many typographical artifacts from ligature conversion, e.g., 'o!' for 'of', 'di!erent' for 'different', and 'e!ect' for 'effect'. A thorough proofreading pass is needed.
  4. [Sec. 3.3] The claim that 'the majority of images (both chemical or virtual) were marked as chemical' is presented as evidence of indistinguishability, but without knowing the readers' prior or the task's base rate this observation is not informative. This should be reframed or supplemented with a quantitative analysis.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the stain outputs are learned on registered training pairs and evaluated on held-out WSIs; self-citations are context, not load-bearing.

full rationale

The derivation chain is self-contained. The paper trains a RegGAN generator to map multichannel PARS input to a chemical stain image, using paired registered patches; the chemical stain is the training target, not an input at inference. Evaluation is on held-out WSIs ('all testing results presented in this work were generated from entirely unseen WSIs that were held out during model training and validation'), and the pathologist comparison uses the same-section chemical stain as an external reference. No fitted parameter is renamed as a prediction: the reported MS-SSIM/DISTS scores and the pathologist classifications are computed on outputs of models applied to unseen slides. The self-citations (e.g., [36], [38], [47]) support instrumentation details, baseline estimation, and prior PARS H&E context; they do not define or force the multi-stain result, and no uniqueness theorem or ansatz is imported from same-author work. The one notable weakness is the non-destructive claim based only on 'no tissue damage visible in any of the stained samples' (Sec. 3.1): UV exposure before chemical staining could in principle alter the molecular targets of the stains, which would undermine the ground-truth validity. That is a biological/correctness risk, not a circularity, because the prediction is not equal to its input by construction.

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

No new physical entities are postulated. The dual-excitation PARS system is a new instrument, not a new theoretical object. The free-parameter list is empty: the neural network weights are learned from training data, which is standard supervised fitting rather than an ad hoc free parameter. The main assumptions are domain-level: the interpretation of 355 nm contrast, the assumption of non-damaging UV exposure, and the adequacy of the registration network.

assumptions (3)
  • domain assumption 355 nm excitation produces non-radiative contrast from hemoglobin and melanin, and radiative contrast from collagen/elastin.
    Stated in Section 2.1 and Figure 4, based on cited literature [10], [43] rather than direct spectroscopic validation in this paper.
  • domain assumption Exposure to low-energy UV does not alter tissue so that post-imaging chemical staining is a valid ground truth.
    Assumed in Section 2.4 and Section 3.1; only visual inspection for damage, no quantitative equivalence check.
  • domain assumption Warpy registration and artifact annotation leave residual misalignments that the RegGAN registration network can model.
    The training scheme depends on this; empirically validated via improved metrics vs Pix2Pix.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Label-Free Whole Slide Virtual Multi-Staining Using Dual-Excitation Photon Absorption Remote Sensing Microscopy." pith.science (2026). https://pith.science/paper/ALKER3VT

@misc{pith2026250905085,
  author       = {Pith},
  title        = {Pith review of: Label-Free Whole Slide Virtual Multi-Staining Using Dual-Excitation Photon Absorption Remote Sensing Microscopy},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ALKER3VT}},
  note         = {Machine review of arXiv:2509.05085}
}
read the original abstract

Histochemical staining is essential for visualizing tissue architecture and cellular morphology but is destructive and limited by the availability of tissue for multiple stains. Virtual staining with label-free microscopy offers a non-destructive alternative, enabling multiple stains to be generated from the same section while reducing stain variability and preserving tissue for downstream assays. Here, a new dual-excitation Photon Absorption Remote Sensing (PARS) system is presented, representing the first application of long-wave ultraviolet A (UVA) 355 nm excitation alongside the established UVC 266 nm source. The addition of 355 nm extends PARS contrast beyond 266 nm, enhancing stromal visualization (e.g., collagen, elastin) and capturing red blood cells, melanin, and other features through complementary radiative and non-radiative absorption. The 266 nm and 355 nm pulses interrogate the sample in an interlaced fashion, enabling concurrent acquisition without compromising imaging speed. Using the RegGAN image-translation framework, this work presents the first demonstration of PARS virtual staining across multiple specialized stains, including Masson's trichrome, periodic acid-Schiff (PAS), and Jones' silver, in addition to hematoxylin and eosin (H&E), across diverse human and murine tissues. A masked evaluation by expert pathologists showed that virtual stains achieved the same diagnostic quality as their chemical counterparts, and pathologists could not reliably distinguish real from virtual stains. By providing label-free multi-stain outputs from a single scan, dual-excitation PARS virtual staining could integrate into digital pathology workflows, expanding diagnostic utility. Real and virtual whole-slide image (WSI) pairs are publicly available at the BioImage Archive (https://doi.org/10.6019/S-BIAD2232).

Figures

Figures reproduced from arXiv: 2509.05085 by the authors.

Figure 1
Figure 1. Dual-excitation PARS architecture, interlaced scanning pattern, and signal extraction. (a) Overview of the PARS microscope optical system. Component labels: AC (achromatic collimator), BT (beam trap), BS (beam sampler), Col. (collimator), HS (harmonic separator), MMF (multi-mode fiber), M (mirror), NF (notch filter), OL (objective lens), PD (photodiode detector), VBE (variable beam expander), Cond. (condenser). (b) … view at source ↗
Figure 2
Figure 2. Workflow for preparing training data and performing whole slide virtual staining. (1) PARS and stained WSIs are aligned using the Warpy registration framework. (2) Tissue borders and artifacts are annotated in QuPath, and the slides are tiled into artifact-free patches. (3) These patches are used to train the virtual staining model. (4) Virtual staining is applied to full PARS WSIs using overlapping patch-wise infer… view at source ↗
Figure 3
Figure 3. Overview of the RegGAN-based virtual staining framework and network architectures. (a) Schematic of the RegGAN training workflow. The generator network 𝐿 receives a multichannel PARS input 𝑀 and produces a virtual stain 𝐿(𝑀). This is concatenated with the misaligned real stain ˜𝑁 and passed through the registration network 𝑂, which predicts a deformation field 𝑃 that is applied to align the virtual stain. A correcti… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Summary of label-free contrasts captured by 266 nm and 355 nm excitation in the PARS system, showing strong correspondence with H&E and specialized stains. (a) PARS images of human skin malignancy (nodular melanoma) showing separate non-radiative (NR) and radiative (R)…
Figure 5
Figure 5. Figure 5: Virtual staining results using the dual excitation PARS input and the RegGAN framework across healthy and diseased tissue cases. Examples span multiple tissue types and include both routine H&E and specialized histochemical stains. For each row: (left) label-free PARS …
Figure 6
Figure 6. Figure 6: Simultaneous virtual H&E, Masson’s trichrome, PAS, and JMS staining of mouse kidney tissue from a single PARS input. (a) Region containing two abnormal glomeruli with nodular expansion of mesangial areas, along with a blood vessel and surrounding renal tubules. (b) Nor…
Figure 7
Figure 7. Figure 7: Dual-excitation models improve staining of histological features across multiple stains. (a) Masson’s trichrome, ccRCC kidney: Each region shows a collagen segmentation map (bottom left) and Collagen Proportionate Area (CPA). Violin plots of CPA distribution (N=4,205, …
Figure 8
Figure 8. Figure 8: Visual and quantitative comparison of virtual staining performance for RegGAN, Pix2Pix, and CycleGAN models. Each example shows the PARS input image, ground truth chemical stain, and outputs from all three models, along with distributions of quantitative metrics (DISTS…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

88 extracted references · 80 canonical work pages

  1. [1]

    Day, editor

    Christina E. Day, editor. Histopathology: Methods and Protocols , volume 1180 of Methods in Molecular Biology . Springer, New Y ork, NY , 2014

    2014

  2. [2]

    Kerr, Tad T

    Fatemeh Ghezloo, Pin-Chieh Wang, Kathleen F. Kerr, Tad T. Bruny´e, Trafton Drew, Oliver H. Chang, Lisa M. Reisch, Linda G. Shapiro, and Joann G. Elmore. An analysis of pathologists’ viewing processes as they diagnose whole slide digital images. Journal of Pathology Informatics, 13:100104, January 2022

    2022

  3. [3]

    Whole Slide Imaging (WSI) in Pathology: Current Perspectives and Future Directions

    Neeta Kumar, Ruchika Gupta, and Sanjay Gupta. Whole Slide Imaging (WSI) in Pathology: Current Perspectives and Future Directions. Journal of Digital Imaging , 33(4):1034–1040, August 2020

    2020

  4. [4]

    Gurina and Lary Simms

    Tatyana S. Gurina and Lary Simms. Histology, Staining. In StatPearls. StatPearls Publishing, Treasure Island (FL), 2023

    2023

  5. [5]

    Morrison, Mark R

    Larry E. Morrison, Mark R. Lefever, Heather N. Lewis, Monesh J. Kapadia, and Daniel R. Bauer. Conventional histological and cytological staining with simultaneous immunohistochemistry enabled by invisible chromogens.Laboratory Investigation, 102(5):545–553, May 2022

    2022

  6. [6]

    Stain normalization in digital pathology: Clinical multi-center evaluation of image quality

    Nicola Michielli, Alessandro Caputo, Manuela Scotto, Alessandro Mogetta, Orazio Antonino Maria Pennisi, Filippo Molinari, Davide Balmativola, Martino Bosco, Alessandro Gambella, Jasna Metovic, Daniele Tota, Laura Carpenito, Paolo Gasparri, and Massimo Salvi. Stain normalization in digital pathology: Clinical multi-center evaluation of image quality. Journ...

    2022

  7. [7]

    Deep learning-enabled virtual histological staining of biological samples

    Bijie Bai, Xilin Y ang, Yuzhu Li, Yijie Zhang, Nir Pillar, and Aydogan Ozcan. Deep learning-enabled virtual histological staining of biological samples. Light: Science & Applications , 12(1):57, March 2023

    2023

  8. [8]

    Quantitative phase imaging in biomedicine

    Y ongKeun Park, Christian Depeursinge, and Gabriel Popescu. Quantitative phase imaging in biomedicine. Nature Photonics, 12(10):578–589, October 2018

    2018

Show all 88 references
  1. [9]

    Croce and G

    A.C. Croce and G. Bottiroli. Autofluorescence Spectroscopy and Imaging: A Tool for Biomedical Research and Diagnosis. European Journal of Histochemistry : EJH , 58(4):2461, December 2014

    2014

  2. [10]

    Junjie Y ao and Lihong V . Wang. Photoacoustic Microscopy. Laser & photonics reviews , 7(5):10.1002/lpor.201200060, September 2013

    2013 doi

  3. [11]

    Photothermal Microscopy for High Sensitivity and High Resolution Absorption Contrast Imaging of Biological Tissues

    Jun Miyazaki and Takayoshi Kobayahsi. Photothermal Microscopy for High Sensitivity and High Resolution Absorption Contrast Imaging of Biological Tissues. Photonics, 4(2):32, June 2017

    2017

  4. [12]

    Miller, Jeremy W

    David R. Miller, Jeremy W . Jarrett, Ahmed M. Hassan, and Andrew K. Dunn. Deep Tissue Imaging with Multiphoton Fluorescence Microscopy. Current opinion in biomedical engineering , 4:32–39, December 2017

    2017

  5. [13]

    Second harmonic generation microscopy: A powerful tool for bio-imaging

    Arash Aghigh, St ´ephane Bancelin, Maxime Rivard, Maxime Pinsard, Heide Ibrahim, and Franc ¸ois L ´egar´e. Second harmonic generation microscopy: A powerful tool for bio-imaging. Biophysical Reviews, 15(1):43–70, January 2023

    2023

  6. [14]

    Theory, innovations and applications of stimulated Raman scattering microscopy

    Wei Min, Ji-Xin Cheng, and Y asuyuki Ozeki. Theory, innovations and applications of stimulated Raman scattering microscopy. Nature Photonics, pages 1–14, August 2025

    2025

  7. [15]

    Zuckerman, Thomas Chong, Anthony E

    Y air Rivenson, Hongda Wang, Zhensong Wei, Kevin de Haan, Yibo Zhang, Yichen Wu, Harun G¨ unaydın, Jonathan E. Zuckerman, Thomas Chong, Anthony E. Sisk, Lindsey M. Westbrook, W . Dean Wallace, and Aydogan Ozcan. Virtual histological staining of unlabelled tissue-autofluorescenc...

    2019

  8. [16]

    Digital synthesis of histological stains using micro-structured and multiplexed virtual staining of label-free tissue

    Yijie Zhang, Kevin de Haan, Y air Rivenson, Jingxi Li, Apostolos Delis, and Aydogan Ozcan. Digital synthesis of histological stains using micro-structured and multiplexed virtual staining of label-free tissue. Light: Science & Applications , 9(1):78, May 2020

    2020

  9. [17]

    Unsupervised content-preserving transformation for optical microscopy

    Xinyang Li, Guoxun Zhang, Hui Qiao, Feng Bao, Yue Deng, Jiamin Wu, Y angfan He, Jingping Yun, Xing Lin, Hao Xie, Haoqian Wang, and Qionghai Dai. Unsupervised content-preserving transformation for optical microscopy. Light: Science & Applications, 10(1):44, March 2021

    2021

  10. [18]

    S ´anchez-Peralta, Riccardo Cicchi, Roberto Bilbao, Domenico Alfieri, Andoni Elola, Ben Glover, and Cristina L

    Artzai Picon, Alfonso Medela, Luisa F. S ´anchez-Peralta, Riccardo Cicchi, Roberto Bilbao, Domenico Alfieri, Andoni Elola, Ben Glover, and Cristina L. Saratxaga. Autofluorescence Image Reconstruction and Virtual Staining for In-Vivo Optical Biopsying. IEEE Access, 9:32081–32093, 2021

    2021

  11. [19]

    A Computationally Virtual Histological Staining Method to Ovarian Cancer Tissue by Deep Generative Adversarial Networks

    Xiangyu Meng, Xin Li, and Xun Wang. A Computationally Virtual Histological Staining Method to Ovarian Cancer Tissue by Deep Generative Adversarial Networks. Computational and Mathematical Methods in Medicine , 2021(1):4244157, 2021

    2021

  12. [20]

    Pixel super-resolved virtual staining of label-free tissue using di!usion models

    Yijie Zhang, Luzhe Huang, Nir Pillar, Yuzhu Li, Hanlong Chen, and Aydogan Ozcan. Pixel super-resolved virtual staining of label-free tissue using di!usion models. Nature Communications, 16(1):5016, May 2025

    2025

  13. [21]

    Doan, Xiaoran Zhang, Yijie Zhang, Jingxi Li, Xilin Y ang, Wenjie Dong, Morgan Angus Darrow, Elham Kamangar, Han Sung Lee, Y air Rivenson, and Aydogan Ozcan

    Bijie Bai, Hongda Wang, Yuzhu Li, Kevin de Haan, Francesco Colonnese, Yujie Wan, Jingyi Zuo, Ngan B. Doan, Xiaoran Zhang, Yijie Zhang, Jingxi Li, Xilin Y ang, Wenjie Dong, Morgan Angus Darrow, Elham Kamangar, Han Sung Lee, Y air Rivenson, and Aydogan Ozcan. Label-free virtual ...

    2022

  14. [22]

    Akram, David A

    Qiang Wang, Ahsan R. Akram, David A. Dorward, Sophie Talas, Basil Monks, Chee Thum, James R. Hopgood, Malihe Javidi, and Marta Vallejo. Deep learning-based virtual H& E staining from label-free autofluorescence lifetime images. npj Imaging, 2(1):1–11, June 2024

    2024

  15. [23]

    Bower, Stephen A

    Navid Borhani, Andrew J. Bower, Stephen A. Boppart, and Demetri Psaltis. Digital staining through the application of deep neural networks to multi-modal multi-photon microscopy. Biomedical Optics Express, 10(3):1339–1350, March 2019

    2019

  16. [24]

    PhaseStain: The digital staining of label-free quantitative phase microscopy images using deep learning

    Y air Rivenson, Tairan Liu, Zhensong Wei, Yibo Zhang, Kevin de Haan, and Aydogan Ozcan. PhaseStain: The digital staining of label-free quantitative phase microscopy images using deep learning. Light: Science & Applications , 8(1):23, February 2019

    2019

  17. [25]

    Nygate, Mattan Levi, Simcha K

    Y oav N. Nygate, Mattan Levi, Simcha K. Mirsky, Nir A. T urko, Moran Rubin, Itay Barnea, Gili Dardikman- Y o!e, Miki Haifler, Alon Shalev, and Natan T. Shaked. Holographic virtual staining of individual biological cells. Proceedings of the National Academy of Sciences, 117(17):...

    2020

  18. [26]

    Deep Learning for Virtual Histological Staining of Bright-Field Microscopic Images of Unlabeled Carotid Artery Tissue

    Dan Li, Hui Hui, Yingqian Zhang, Wei Tong, Feng Tian, Xin Y ang, Jie Liu, Yundai Chen, and Jie Tian. Deep Learning for Virtual Histological Staining of Bright-Field Microscopic Images of Unlabeled Carotid Artery Tissue. Molecular Imaging and Biology, 22(5):1301–1309, October 2020

    2020

  19. [27]

    Image-to-Images Translation for Multiple Virtual Histological Staining of Unlabeled Human Carotid Atherosclerotic Tissue

    Guanghao Zhang, Bin Ning, Hui Hui, Tengfei Yu, Xin Y ang, Hongxia Zhang, Jie Tian, and Wen He. Image-to-Images Translation for Multiple Virtual Histological Staining of Unlabeled Human Carotid Atherosclerotic Tissue. Molecular Imaging and Biology, 24(1):31–41, February 2022

    2022

  20. [28]

    Lei Kang, Xiufeng Li, Y an Zhang, and Terence T. W . Wong. Deep learning enables ultraviolet photoacoustic microscopy based histological imaging with near real-time virtual staining. Photoacoustics, 25:100308, March 2022

    2022

  21. [29]

    Nelson, Samuel Davis, Yu Liang, Yilin Luo, Yide Zhang, Brooke Crawford, and Lihong V

    Rui Cao, Scott D. Nelson, Samuel Davis, Yu Liang, Yilin Luo, Yide Zhang, Brooke Crawford, and Lihong V . Wang. Label-free intraoperative histology of bone tissue via deep-learning-assisted ultraviolet photoacoustic microscopy. Nature Biomedical Engineering, 7(2):124–134, February 2023

    2023

  22. [30]

    Deep learning-based virtual staining, segmentation, and classification in label-free photoacoustic histology of human specimens

    Chiho Y oon, Eunwoo Park, Sampa Misra, Jin Y oung Kim, Jin Woo Baik, Kwang Gi Kim, Chan Kwon Jung, and Chulhong Kim. Deep learning-based virtual staining, segmentation, and classification in label-free photoacoustic histology of human specimens. Light: Science & Applications , ...

    2024

  23. [31]

    Martell, Nathaniel J

    Matthew T. Martell, Nathaniel J. M. Haven, Brendyn D. Cikaluk, Brendon S. Restall, Ewan A. McAlister, Rohan Mittal, Benjamin A. Adam, Nadia Giannakopoulos, Lashan Peiris, Sveta Silverman, Jean Deschenes, Xingyu Li, and Roger J. Zemp. Deep learning-enabled realistic virtual his...

    2023

  24. [32]

    Virtual formalin-fixed and para”n-embedded staining of fresh brain tissue via stimulated Raman CycleGAN model.Science Advances, 10(13):eadn3426, March 2024

    Zhijie Liu, Lingchao Chen, Haixia Cheng, Jianpeng Ao, Ji Xiong, Xing Liu, Y axin Chen, Ying Mao, and Minbiao Ji. Virtual formalin-fixed and para”n-embedded staining of fresh brain tissue via stimulated Raman CycleGAN model.Science Advances, 10(13):eadn3426, March 2024

    2024

  25. [33]

    Mukherjee, Sounak Gupta, Loren Herrera-Hernandez, Michael R

    Kianoush Falahkheirkhah, Sudipta S. Mukherjee, Sounak Gupta, Loren Herrera-Hernandez, Michael R. McCarthy, Rafael E. Jimenez, John C. Cheville, and Rohit Bhargava. Accelerating Cancer Histopathology Workflows with Chemical Imaging and Machine Learning. Cancer Research Communica...

    2023

  26. [34]

    Computational tissue staining of non-linear multimodal imaging using supervised and unsupervised deep learning

    Pranita Pradhan, Tobias Meyer, Michael Vieth, Andreas Stallmach, Maximilian Waldner, Michael Schmitt, Juergen Popp, and Thomas Bocklitz. Computational tissue staining of non-linear multimodal imaging using supervised and unsupervised deep learning. Biomedical Optics Express, 1...

    2021

  27. [35]

    Marian Boktor, James E. D. T weel, Benjamin R. Ecclestone, Jennifer Ai Y e, Paul Fieguth, and Parsin Haji Reza. Multi-channel feature extraction for virtual histological staining of photon absorption remote sensing images. Scientific Reports, 14(1):2009, January 2024

    2009

  28. [36]

    Ecclestone, Kevan Bell, Sarah Sparkes, Deepak Dinakaran, John R

    Benjamin R. Ecclestone, Kevan Bell, Sarah Sparkes, Deepak Dinakaran, John R. Mackey, and Parsin Haji Reza. Label-free complete absorption microscopy using second generation photoacoustic remote sensing. Scientific Reports, 12(1):8464, May 2022

    2022

  29. [37]

    Ecclestone, James A

    Benjamin R. Ecclestone, James A. T ummon Simmons, James E. D. T weel, Deepak Dinakaran, and Parsin Haji Reza. Photon Absorption Remote Sensing (PARS): Comprehensive Absorption Imaging Enabling Label-Free Biomolecule Characterization and Mapping, June 2025

    2025

  30. [38]

    T weel, Benjamin R

    James E.D. T weel, Benjamin R. Ecclestone, Marian Boktor, Deepak Dinakaran, John R. Mackey, and Parsin Haji Reza. Automated Whole Slide Imaging for Label-Free Histology Using Photon Absorption Remote Sensing Microscopy. IEEE Transactions on Biomedical Engineering, 71(6):1901–1...

    1901

  31. [39]

    Ecclestone, Vlad Pekar, Deepak Dinakaran, John R

    Marian Boktor, Benjamin R. Ecclestone, Vlad Pekar, Deepak Dinakaran, John R. Mackey, Paul Fieguth, and Parsin Haji Reza. Virtual histological staining of label-free total absorption photoacoustic remote sensing (TA-PARS). Scientific Reports , 12(1):10296, June 2022. 24

    2022

  32. [40]

    Ecclestone, James E

    Benjamin R. Ecclestone, James E. D. T weel, Marie Abi Daoud, Hager Gaouda, Deepak Dinakaran, Michael P . Wallace, Ally Khan Somani, Gilbert Bigras, John R. Mackey, and Parsin Haji Reza. Photon Absorption Remote Sensing Virtual Histopathology: Diagnostic Equivalence to Gold-Sta...

    2025

  33. [41]

    James E. D. T weel, Benjamin R. Ecclestone, Marian Boktor, James Alexander T ummon Simmons, Paul Fieguth, and Parsin Haji Reza. Virtual Histology with Photon Absorption Remote Sensing using a Cycle-Consistent Generative Adversarial Network with Weakly Registered Pairs, June 2023

    2023

  34. [42]

    James E. D. T weel, Benjamin R. Ecclestone, Hager Gaouda, Deepak Dinakaran, Michael P . Wallace, Gilbert Bigras, John R. Mackey, and Parsin Haji Reza. Photon Absorption Remote Sensing Imaging of Breast Needle Core Biopsies Is Diagnostically Equivalent to Gold Standard H&E Hist...

    2023

  35. [43]

    Deep UV autofluorescence microscopy for cell biology and tissue histology

    Fr´ed´eric Jamme, Slavka Kascakova, Sandrine Villette, Fatma Allouche, St ´ephane Pallu, Val ´erie Rouam, and Matthieu R´efr´egiers. Deep UV autofluorescence microscopy for cell biology and tissue histology. Biology of the Cell, 105(7):277–288, 2013

    2013

  36. [44]

    Cell and tissue autofluorescence research and diagnostic applications

    Monica Monici. Cell and tissue autofluorescence research and diagnostic applications. In Biotechnology Annual Review , volume 11, pages 227–256. Elsevier, January 2005

    2005

  37. [45]

    Breaking the Dilemma of Medical Image-to-image Translation, November 2021

    Lingke Kong, Chenyu Lian, Detian Huang, Zhenjiang Li, Y anle Hu, and Qichao Zhou. Breaking the Dilemma of Medical Image-to-image Translation, November 2021

    2021

  38. [46]

    The bioimage archive–building a home for life-sciences microscopy data

    Matthew Hartley, Gerard J Kleywegt, Ardan Patwardhan, Ugis Sarkans, Jason R Swedlow, and Alvis Brazma. The bioimage archive–building a home for life-sciences microscopy data. Journal of Molecular Biology , 434(11):167505, 2022

    2022

  39. [47]

    T ummon Simmons, Sarah J

    James A. T ummon Simmons, Sarah J. Werezak, Benjamin R. Ecclestone, James E. D. T weel, Hager Gaouda, and Parsin Haji Reza. Label-Free Non-Contact Vascular Imaging Using Photon Absorption Remote Sensing.IEEE Transactions on Biomedical Engineering, 72(3):1160–1169, March 2025

    2025

  40. [48]

    An Open-Source Whole Slide Image Registration Workflow at Cellular Precision Using Fiji, QuPath and Elastix.Frontiers in Computer Science, 3, January 2022

    Nicolas Chiaruttini, Olivier Burri, Peter Haub, Romain Guiet, Jessica Sordet-Dessimoz, and Arne Seitz. An Open-Source Whole Slide Image Registration Workflow at Cellular Precision Using Fiji, QuPath and Elastix.Frontiers in Computer Science, 3, January 2022

    2022

  41. [49]

    Phillip Isola, Jun- Y an Zhu, Tinghui Zhou, and Alexei A. Efros. Image-to-Image Translation with Conditional Adversarial Networks, November 2018

    2018

  42. [50]

    High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs, August 2018

    Ting-Chun Wang, Ming- Yu Liu, Jun- Y an Zhu, Andrew Tao, Jan Kautz, and Bryan Catanzaro. High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs, August 2018

    2018

  43. [51]

    Jun- Y an Zhu, Taesung Park, Phillip Isola, and Alexei A. Efros. Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks, August 2020

    2020

  44. [52]

    Virtual staining for pathology: Challenges, limitations and perspectives

    Weiping Lin, Yihuang Hu, Runchen Zhu, Baoshun Wang, and Liansheng Wang. Virtual staining for pathology: Challenges, limitations and perspectives. Intelligent Oncology, 1(2):105–119, April 2025

    2025

  45. [53]

    Xudong Mao, Qing Li, Haoran Xie, Raymond Y . K. Lau, Zhen Wang, and Stephen Paul Smolley. Least Squares Generative Adversarial Networks, April 2017

    2017

  46. [54]

    Unsupervised Multi-Modal Image Registration via Geometry Preserving Image-to-Image Translation, March 2020

    Moab Arar, Yiftach Ginger, Dov Danon, Ilya Leizerson, Amit Bermano, and Daniel Cohen-Or. Unsupervised Multi-Modal Image Registration via Geometry Preserving Image-to-Image Translation, March 2020

    2020

  47. [55]

    Learning from Simulated and Unsupervised Images through Adversarial Training, July 2017

    Ashish Shrivastava, Tomas Pfister, Oncel T uzel, Josh Susskind, Wenda Wang, and Russ Webb. Learning from Simulated and Unsupervised Images through Adversarial Training, July 2017

    2017

  48. [56]

    Sellaro, Robert Filkins, Chelsea Ho!man, Je!rey L

    Ti!any L. Sellaro, Robert Filkins, Chelsea Ho!man, Je!rey L. Fine, Jon Ho, Anil V . Parwani, Liron Pantanowitz, and Michael Montalto. Relationship between magnification and resolution in digital pathology systems. Journal of Pathology Informatics, 4(1):21, January 2013

    2013

  49. [57]

    November 2024

    Weedon’s Skin Pathology. November 2024

    2024

  50. [58]

    Charles Jennette MD and Vivette D

    J. Charles Jennette MD and Vivette D. D’ Agati MD. Heptinstall’s Pathology of the Kidney. Wolters Kluwer, Philadelphia, 2024

    2024

  51. [59]

    Dermatomycosis from the perspective of dermatopathology (version 1.1)

    Taekwoon Kim, Jeongsoo Lee, and Joonsoo Park. Dermatomycosis from the perspective of dermatopathology (version 1.1). Journal of, 26(3), 2021

    2021

  52. [60]

    Silva’s diagnostic renal pathology

    Xin J Zhou, Zoltan G Laszik, Tibor Nadasdy, Vivette D D’ Agati, et al. Silva’s diagnostic renal pathology . Cambridge University Press, 2017

    2017

  53. [61]

    Historical control data of spontaneous pathological findings in c57bl/6j mice used in 18-month dietary carcinogenicity assays.Toxicologic Pathology, 52(2-3):99–113, 2024

    La¨etitia Elies, Elise Guillaume, Mathilde Gorieu, Patricia Neves, and Fr´ed´eric Schorsch. Historical control data of spontaneous pathological findings in c57bl/6j mice used in 18-month dietary carcinogenicity assays.Toxicologic Pathology, 52(2-3):99–113, 2024. 25

    2024

  54. [62]

    Renal function in aged c57bl/6j mice is impaired by deposition of age-related apolipoprotein a-ii amyloid independent of kidney aging

    Ying Li, Jian Dai, Fuyuki Kametani, Masahide Y azaki, Akihito Ishigami, Masayuki Mori, Hiroki Miyahara, and Keiichi Higuchi. Renal function in aged c57bl/6j mice is impaired by deposition of age-related apolipoprotein a-ii amyloid independent of kidney aging. The American Jour...

    2023

  55. [63]

    Hoane, Crystal L

    Jessica S. Hoane, Crystal L. Johnson, James P . Morrison, and Susan A. Elmore. Comparison of Renal Amyloid and Hyaline Glomerulopathy in B6C3F1 Mice: An NTP Retrospective Study. Toxicologic pathology, 44(5):687–704, July 2016

    2016

  56. [64]

    Zuckerman, Tairan Liu, Anthony E

    Kevin de Haan, Yijie Zhang, Jonathan E. Zuckerman, Tairan Liu, Anthony E. Sisk, Miguel F. P . Diaz, Kuang- Yu Jen, Alexander Nobori, Sofia Liou, Sarah Zhang, Rana Riahi, Y air Rivenson, W . Dean Wallace, and Aydogan Ozcan. Deep learning-based transformation of H&E stained tissu...

    2021

  57. [65]

    Wang, E.P

    Z. Wang, E.P . Simoncelli, and A.C. Bovik. Multiscale structural similarity for image quality assessment. InThe Thrity-Seventh Asilomar Conference on Signals, Systems & Computers, 2003 , volume 2, pages 1398–1402 Vol.2, November 2003

    2003

  58. [66]

    Simoncelli

    Keyan Ding, Kede Ma, Shiqi Wang, and Eero P . Simoncelli. Image Quality Assessment: Unifying Structure and Texture Similarity. IEEE Transactions on Pattern Analysis and Machine Intelligence, pages 1–1, 2020

    2020

  59. [67]

    Manesis, Massimo Pinzani, Amar P

    Elena Buzzetti, Andrew Hall, Mattias Ekstedt, Roberta Manuguerra, Marta Guerrero Misas, Claudia Covelli, Gioacchino Leandro, T uVinh Luong, Stergios Kechagias, Emanuel K. Manesis, Massimo Pinzani, Amar P . Dhillon, and Emmanuel A. Tsochatzis. Collagen proportionate area is an ...

    2019

  60. [68]

    Cell Detection with Star-Convex Polygons

    Uwe Schmidt, Martin Weigert, Coleman Broaddus, and Gene Myers. Cell Detection with Star-Convex Polygons. In Alejan- dro F. Frangi, Julia A. Schnabel, Christos Davatzikos, Carlos Alberola-L ´opez, and Gabor Fichtinger, editors, Medical Image Computing and Computer Assisted Inte...

    2018

  61. [70]

    T ype – Stain – Sample Path

    Supplemental Information Supplementary Table 1: Results of masked pathologist evaluation showing responses for diagnostic quality (DQ) and image origin (IO). T ype – Stain – Sample Path. 1 Path. 2 Path. 3 Avg. DQ DQ IO DQ IO DQ IO

  62. [71]

    C – H&E – Kidney ccRCC 3Y 2U 3U 2.67

  63. [73]

    C – MT – Kidney ccRCC 2Y 2Y 3U 2.33

  64. [74]

    V – MT – Kidney ccRCC 2Y 2U 3U 2.33

  65. [75]

    C – PAS – Mouse Kidney 3Y 1N 3U 2.33

  66. [76]

    V – PAS – Mouse Kidney 3Y 2Y 3U 2.67

  67. [77]

    C – JMS – Mouse Kidney 3Y 1Y 2U 2.00

  68. [78]

    V – JMS – Mouse Kidney 2Y 2Y 2U 2.00

  69. [79]

    C – H&E – Skin Melanoma 3Y 3Y 3U 3.00

  70. [80]

    V – H&E – Skin Melanoma 3Y 3Y 3U 3.00

  71. [81]

    C – PAS – Mouse GI 3Y 2U 3U 2.67

  72. [82]

    V – PAS – Mouse GI 3Y 2N 3U 2.67

  73. [84]

    V – MT – Kidney ccRCC 3Y 3Y 3U 3.00

  74. [85]

    C – PAS – Skin Fungal 3Y 3Y 3U 3.00

  75. [86]

    V – PAS – Skin Fungal 3Y 3Y 3U 3.00

  76. [87]

    C – H&E – Kidney ccRCC 3Y 2Y 3U 2.67

  77. [88]

    V – H&E – Kidney ccRCC 3Y 2Y 3U 2.67

  78. [89]

    C – MT – Kidney ccRCC 3Y 2Y 3U 2.67

  79. [90]

    IO = Image Origin (Y : Y es, N: No, U: Uncertain)

    V – MT – Kidney ccRCC 3Y 2Y 3U 2.67 DQ = Diagnostic Quality (1: poor, 2: good, 3: excellent). IO = Image Origin (Y : Y es, N: No, U: Uncertain). C = Chemical stain. V = Virtual stain. MT = Masson’s Trichrome; 27 Supplementary Figure 1: Virtual staining results using the dual e...

  80. [2018]

    Springer International Publishing. 26

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.