Automated grading and staging of ovarian cancer using deep learning on the transmission optical microscopy bright-field images of thin biopsy tissue samples

Ashmit K Mishra; Mousa Alrubayan; Prabhakar Pradhan

REVIEW 3 major objections 5 minor 19 references

Automated grading and staging of ovarian cancer using deep learning on the transmission optical microscopy bright-field images of thin biopsy tissue samples

T0 review · 3 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 ResNet-101 fine-tuned on brightfield microscopy images of thin ovarian tissue cores assigns FIGO stage (0–IV) with 97.62% accuracy on an independent test set of 126 images.

desk verdict A standard transfer-learning result whose headline accuracy is not yet evidence of staging; the missing patient-level split and the questionable link between a single tissue core and FIGO stage are the load-bearing weaknesses. read the letter →

arxiv 2505.09993 v1 pith:FALPZ7CZ submitted 2025-05-15 physics.med-ph physics.bio-phphysics.optics

Ashmit K Mishra , Mousa Alrubayan , Prabhakar Pradhan This is my paper · ORCID

classification physics.med-phphysics.bio-phphysics.optics

keywords deeplearningbrightfieldmicroscopyhistopathologyovariancancerstagingFIGOResNet-101transferdigitalpathology

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

This paper claims that a convolutional neural network can read the stage of ovarian cancer directly from ordinary brightfield microscopy images of thin tissue cores, without molecular or clinical inputs. On an independent test set of 126 images spanning control tissue and FIGO Stages I–IV, the fine-tuned ResNet-101 classified 123 images correctly, an accuracy of 97.62%, with all three errors falling between adjacent stages. The claim matters because ovarian cancer treatment and prognosis depend heavily on stage, and manual staging is time-consuming and subject to inter-observer variability. If the result holds, automated staging could offer a fast, consistent first-pass read in digital pathology workflows, flagging cases for pathologist review rather than replacing the pathologist.

What carries the argument

The load-bearing object is a ResNet-101 convolutional neural network pre-trained on ImageNet and fine-tuned for five-way stage classification. Its classification head replaces the original last layer with a dropout layer (rate $0.303$) followed by a linear layer to five outputs and a softmax. The other mechanisms are data augmentation during training (flips, small rotations, color jitter, affine transforms), class-imbalance handling through a weighted random sampler and weighted cross-entropy loss, early stopping monitored on validation accuracy, and a genetic algorithm that selected the learning rate ($2.00\times10^{-4}$), dropout rate, and weight decay ($3.72\times10^{-5}$). Together these let a small histopathology dataset adapt a generic visual-feature extractor to stage-associated tissue morphology.

What would settle it

Take the trained model to a cohort scanned with a different microscope, stain batch, or institution, with pathologist-confirmed FIGO labels, and compare accuracy: a collapse toward chance would show the 97.62% result came from acquisition-specific cues rather than stage morphology. A second, cheaper check is to inspect saliency maps: if the model attends to glass edges, background, or stain color patches rather than tissue structures, the stage-learning claim is falsified.

Watch

Extended reading notes

Core claim

The central discovery is that the FIGO surgical stage label, mapped to five classes (0 for control, 1–4 for Stages I–IV), is recoverable from the morphology of a single 1.5 mm-diameter, 5 µm-thick tissue core imaged in transmission brightfield. The trained model, optimized with data augmentation, weighted sampling, weighted cross-entropy loss, and hyperparameters found by a genetic algorithm, achieved perfect precision and recall on the extreme classes (control and Stage IV) and F1-scores of 0.98, 0.91, and 0.97 on the middle stages. The authors interpret the adjacent-stage error pattern as consistent with the genuinely subtle morphological boundaries between Stages I, II, and III, and conclude that deep learning can act as an assistive staging tool.

Load-bearing premise

The load-bearing premise is that a single 1.5 mm tissue core imaged in brightfield contains enough morphological information to determine the FIGO surgical stage, which is defined by the anatomic extent of tumor spread found at surgery.

Editorial extensions

If this is right

Control and Stage IV are perfectly separable in this test set (F1 = 1.00), so the practical ambiguity lies between the adjacent middle stages, where even morphologically similar cases are hard for human readers.
An automated stage prediction could be used as a triage and consistency-check tool in digital pathology, prioritizing cases with low-confidence predictions for closer pathologist review.
The accuracy was reached with only 2,418 training images, suggesting ImageNet-pretrained backbones plus augmentation and class weighting can handle small-domain histopathology staging tasks.
The model as tested sees single tissue cores; the paper argues that combining patch-level predictions across whole-slide images with clinical data may further improve separation of adjacent stages.

Reading between the lines

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

Because the independent test set was acquired with the same BX61 microscope and imaging protocol as the training data, the 97.62% figure is best read as a within-protocol ceiling; a cross-scanner or multi-stain cohort would probably lower it, and that gap is exactly what the paper's own call for multi-institutional validation targets.
If the model is truly reading stage morphology, its saliency maps should concentrate on epithelial structures, stroma, and immune infiltrate rather than on background glass, tissue edges, or stain-color patches; this is a cheap check that could be run before any clinical pilot.
A reader checking reproducibility should note that the architecture citation 'ResNet-101 [22]' has no matching entry in the reference list, which ends at [19]; the paper's bibliography therefore does not by itself supply the standard provenance for the backbone.
The near-perfect separation of the extreme classes suggests a continuous ordinal morphology score may underlie the five discrete labels; regressing on tumor burden or stage as an ordered variable could give finer-grained utility than classification alone.

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.

Desk Editor's Note

read the letter

Quick take: this is a routine fine-tuning of ResNet-101 on ovarian tissue brightfield images, and the main number — 97.62% on 126 test images — is not supported as evidence of automated staging until two things are shown. First, that the independent test set is disjoint from the training set at the patient or core level. The paper reports image counts only. Since each TMA slide holds 24 cores and scanning is automated, adjacent tiles from the same core or batch can easily straddle a nominal split. A CNN can hit near-perfect accuracy by memorizing tile- or batch-specific texture rather than stage-defining morphology. Second, that a 1.5 mm core actually contains the morphological information that defines FIGO stage. FIGO staging is defined by the extent of tumor spread found at surgery; a thin core may not show that. If either point fails, the 97.62% is not staging.

What the paper does well: it uses a sensible transfer-learning setup, with augmentation, class weighting, and a genetic algorithm for hyperparameters. It reports a confusion matrix and per-class metrics, and the errors cluster between adjacent stages, which is plausible. The authors also acknowledge the need for multi-institutional validation in their limitations paragraph. That counts.

The soft spots are proportionate to the claim. The central claim is categorical: "deep learning can achieve high accuracy in discriminating between different stages." But the evidence does not establish discriminative features of stage as opposed to correlated artifacts or patient identity. The test set is small and imbalanced — 61 of 126 are Stage I — and there are no error bars. No code or data is released, so the result is not reproducible. The citation pattern is fine; no red flags there.

This paper is for anyone working in computational pathology who wants to see a standard pipeline applied to a new task. It deserves peer review because the claim is testable and the experiment is reported with enough detail to be checked — but only if reviewers insist on patient-level splitting and external validation. The authors should be asked to show that the test set is patient-disjoint and that the model generalizes across institutions or scanners.

My recommendation: send it for review with a strong request for those revisions. It's not a desk reject, but it's not close to being accepted as is.

Referee Report

3 major / 5 minor

Summary. The paper presents a deep learning framework for automated ovarian cancer staging (classes 0, I, II, III, IV) from transmission optical microscopy bright-field images of thin tissue cores. Using a ResNet-101 initialized with ImageNet weights, the authors fine-tune on 2418 images (80/20 stratified into training/validation) and evaluate on an independent test set of 126 images, reporting an overall accuracy of 97.62% (123/126 correct). The training pipeline includes data augmentation, weighted random sampling and class-weighted cross-entropy loss, with learning rate, dropout rate, and weight decay tuned via a genetic algorithm. The paper reports per-class precision, recall, F1-scores, a confusion matrix, and describes misclassifications as primarily occurring between adjacent stages.

Significance. If the central result holds, the paper would contribute a practical automated staging tool for ovarian cancer histopathology, a domain where most prior work has focused on grading, subtyping, or survival prediction rather than direct FIGO stage classification from bright-field images. The use of a held-out test set, per-class metrics, explicit class-imbalance handling, and transparent descriptions of the augmentation and hyperparameter search are strengths. However, the significance is tempered by two load-bearing concerns: the independence of the test set from the training set at the patient and tissue-core level is not established, and the biological plausibility of deriving surgical FIGO stage from the morphology of a single 1.5 mm core is not justified. These issues must be resolved before the reported accuracy can be interpreted as evidence of genuine stage-associated morphological learning.

major comments (3)

[Section 3.1 and Appendix A] The independence of the test set is not demonstrated at the patient or tissue-core level. The manuscript reports only image counts and states that the training/validation and test images come from 'two distinct collections,' but no patient, block, core, or slide identifiers are provided. Because each TMA slide contains 24 cores and scanning is automated, images from the same core, adjacent sections, or the same scanning session could straddle the train/test boundary. Without a patient-disjoint split, the reported 97.62% accuracy (123/126) may reflect memorization of tile-specific or batch-specific texture rather than stage-associated morphology. The authors should provide patient-level metadata or re-analyze with a patient-exclusive split to support the claim of an 'independent' test set.
[Section 2.3] The text states that misclassifications 'primarily occurred between adjacent stages,' but it also reports that one actual Class 3 sample was misclassified as Class 1, which is a non-adjacent error (a jump of two stages). This internal inconsistency affects the interpretation of the confusion matrix and should be corrected, either by revising the statement or by re-checking the confusion matrix.
[Sections 3.1 and 3.3] The premise that FIGO surgical stage is learnable from the morphology of a single 1.5 mm tissue core imaged in brightfield is biologically questionable. FIGO staging is defined by the extent of tumor spread found at surgery (e.g., pelvic versus upper abdominal involvement), which is not necessarily present in a thin core. If the model is not actually capturing stage-defining morphology, the high accuracy may reflect spurious correlations with acquisition or staining covariates. The manuscript should either provide evidence that stage-discriminating morphology is present in these cores (e.g., comparison with pathologist assessment or supporting references specific to ovarian core samples) or temper the claim to 'image-based classification of biopsy tissue' rather than 'staging.' This is a correctness-risk concern that directly bears on the external validity of the central claim.

minor comments (5)

[Title] The title mentions 'grading and staging,' but the model predicts only stage (five classes, with class 0 as control); no separate grading output is reported. Consider revising the title to 'staging' or clarifying that grading is subsumed in the staging assessment.
[Section 3.1] The description of the independent test set is unclear: 'obtained from BX61' presumably refers to the same Olympus BX61 microscope used for the primary collection, but it should specify whether the test images come from different TMA slides, different patients, or different scanning sessions, and how many patients are represented in each split.
[Section 3.1 and Appendix A] The manuscript reports only image counts and does not provide the number of patients, cores, or images per core. This information is necessary for the reader to assess the effective sample size and the degree of data dependence in the splits.
[Section 3.5] The optimal dropout rate (0.303) lies at the boundary of the GA search range [0.3, 0.7]. This suggests that the optimum may lie outside the defined range, and the search space should be extended to ensure that the hyperparameter choice is not an artifact of the chosen bounds.
[Section 4 and Data Availability] The manuscript does not provide code, trained model weights, or a detailed data availability statement beyond 'available from the corresponding author upon reasonable request.' Adding a repository or explicit data-sharing plan would improve reproducibility, which is particularly important given the need to verify the patient-level independence of the splits.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the claimed 97.62% test accuracy is an empirical held-out result, not a quantity forced by construction or by a self-citation chain.

full rationale

The paper's derivation chain is empirical rather than definitional: thin-tissue brightfield images with FIGO stage labels are used to fine-tune a ResNet-101; the genetic algorithm tunes hyperparameters on validation accuracy; the final model is then evaluated on a separate 126-image test set (Section 2.2, Section 3.5, Appendix A). No equation defines the predicted stage in terms of a fitted parameter, no target metric is used as a training objective, and the reported accuracy is measured on data not used for training or hyperparameter selection. The only self-citation, Ref. [16], describes the automated scanning setup and is not load-bearing for the staging claim; it does not supply a uniqueness theorem, an ansatz, or the stage labels themselves. Concerns about patient-level independence, batch effects, and the learnability of FIGO stage from small tissue cores are external-validity or data-leakage risks, not circularity: they question whether the high accuracy would transfer, but they do not show that the reported number was preordained by its own inputs. Under the stated rules requiring a quotable reduction to a fit or self-citation, no circular step is present.

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

The central claim rests on three GA-tuned hyperparameters, two unvalidated domain assumptions, and standard supervised learning assumptions. No new physical or biological entities are introduced.

free parameters (3)

learning_rate = 2.0027e-4
Tuned by genetic algorithm on validation accuracy (Section 3.5).
dropout_rate = 0.3032
Tuned by genetic algorithm on validation accuracy (Section 3.5).
weight_decay = 3.724e-5
Tuned by genetic algorithm on validation accuracy (Section 3.5).

assumptions (3)

domain assumption FIGO surgical stage is inferable from the morphology of a single 1.5 mm tissue core brightfield image.
The dataset labels each image with a FIGO stage and the model maps image features directly to stages (Sections 3.1, 3.3). This is unvalidated and biologically questionable because staging depends on surgical-pathological extent.
domain assumption The independent test set is patient-disjoint and representative.
The paper states the test set is independent (Section 3.1) but does not describe patient-level splitting; training and test images come from the same lab and vendor, so patient overlap or batch effects cannot be excluded.
domain assumption Standard supervised learning assumptions hold, including that the training and test distributions match and that augmentations simulate realistic variability.
These assumptions are implicit in the training setup (Section 3.2) and are not verified against an external cohort.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Automated grading and staging of ovarian cancer using deep learning on the transmission optical microscopy bright-field images of thin biopsy tissue samples." pith.science (2026). https://pith.science/paper/FALPZ7CZ

@misc{pith2026250509993,
  author       = {Pith},
  title        = {Pith review of: Automated grading and staging of ovarian cancer using deep learning on the transmission optical microscopy bright-field images of thin biopsy tissue samples},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FALPZ7CZ}},
  note         = {Machine review of arXiv:2505.09993}
}

read the original abstract

Ovarian cancer remains a challenging malignancy to diagnose and manage, with prognosis heavily dependent on the stage at detection. Accurate grading and staging, primarily based on histopathological examination of biopsy tissue samples, are crucial for treatment planning and predicting outcomes. However, this manual process is time-consuming and subject to inter-observer variability among pathologists. The increasing volume of digital histopathology slides necessitates the development of robust, automated methods to assist in this critical diagnostic step for ovarian cancer. (Methods) This study presents a deep learning framework for the automated prediction of ovarian cancer stage (classified into five categories: 0, I, II, III, IV) using routine histopathological images. We employed a transfer learning approach, fine-tuning a ResNet-101 convolutional neural network pre-trained on ImageNet. The training process incorporated comprehensive data augmentation, weighted random sampling, and class weighting to address dataset characteristics. Hyperparameter optimization for learning rate, dropout rate, and weight decay was performed using a genetic algorithm to enhance model performance and generalization. (Results) Evaluated on an independent test set of ovarian thin tissue brightfield images, the developed model achieved a high overall classification accuracy of 97.62%.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

19 extracted references · 19 canonical work pages

[1]

J. W. Zhu, P. Charkhchi, and M. R. Akbari, Potential clinical utility of liquid biopsies in ovarian cancer, Molecular Cancer 21, 114 (2022)

work page 2022
[2]

Andreou, M

M. Andreou, M. Kyprianidou, C. Cortas, I. Polycarpou, D. Papamichael, P. Kountourakis, and K. Giannakou, Prognostic Factors Influencing Survival in Ovarian Cancer Patients: A 10 -Year Retrospective Study, Cancers (Basel) 15, 5710 (2023)

work page 2023
[3]

Staging for OVFTP Malignancies | Figo, https://www.figo.org/news/staging-ovftp-malignancies

work page
[4]

P. S. Ginter, R. Idress, T. M. D’Alfonso, S. Fineberg, S. Jaffer, A. K. Sattar, A. Chagpar, P. Wilson, and M. Harigopal, Histologic grading of breast carcinoma: a multi-institution study of interobserver variation using virtual microscopy, Modern Pathology 34, 701 (2021)

work page 2021
[5]

Javadi, D

S. Javadi, D. M. Ganeshan, A. Qayyum, R. B. Iyer, and P. Bhosale, Ovarian Cancer, the Revised FIGO Staging System, and the Role of Imaging, American Journal of Roentgenology 206, 1351 (2016)

work page 2016
[6]

M. Woynarowski, Inter-observer variability in histopathological assessment of liver biopsies taken in a pediatric open-label therapeutic program for chronic HBV infection treatment, World Journal of Gastroenterology 12, (2006)

work page 2006
[7]

M. G. Hanna et al., Implementation of Digital Pathology Offers Clinical and Operational Increase in Efficiency and Cost Savings, Archives of Pathology & Laboratory Medicine 143, 1545 (2019)

work page 2019
[8]

I. D. Mienye, T. G. Swart, G. Obaido, M. Jordan, and P. Ilono, Deep Convolutional Neural Networks in Medical Image Analysis: A Review, Information 16, 3 (2025)

work page 2025

Show all 19 references

[9]

Prasad, A

P. Prasad, A. M. B. Khair, M. Saeed, and N. Shetty, Artificial Intelligence in Histopathology, J Pharm Bioallied Sci 16, S4226 (2024)

2024
[10]

H. Cui, Q. Guo, J. Xu, X. Wu, C. Cai, Y . Jiao, W. Ming, H. Wen, and X. Wang, Prediction of molecular subtypes for endometrial cancer based on hierarchical foundation model, Bioinformatics 41, btaf059 (2025)

2025
[11]

Breen, K

J. Breen, K. Allen, K. Zucker, P. Adusumilli, A. Scarsbrook, G. Hall, N. M. Orsi, and N. Ravikumar, Artificial intelligence in ovarian cancer histopathology: a systematic review, Npj Precis. Onc. 7, 1 (2023)

2023
[12]

Breen, K

J. Breen, K. Allen, K. Zucker, L. Godson, N. M. Orsi, and N. Ravikumar, A comprehensive evaluation of histopathology foundation models for ovarian cancer subtype classification, Npj Precis. Onc. 9, 1 (2025)

2025
[13]

Meinhold-Heerlein and S

I. Meinhold-Heerlein and S. Hauptmann, The heterogeneity of ovarian cancer, Arch Gynecol Obstet 289, 237 (2014)

2014
[14]

D. Lim, R. Murali, M. P. Murray, E. Veras, K. J. Park, and R. A. Soslow, Morphological and Immunohistochemical Reevaluation of Tumors Initially Diagnosed as Ovarian Endometrioid Carcinoma With Emphasis on High-grade Tumors, The American Journal of Surgical Pathology 40, 302 (2016)

2016
[15]

M. H. Stoler, B. M. Ronnett, N. E. Joste, W. C. Hunt, J. Cuzick, C. M. Wheeler, and F. the N. M. H. P. R. S. Committee, The Interpretive Variability of Cervical Biopsies and Its Relationship to HPV Status, The American Journal of Surgical Pathology 39, 729 (2015)

2015
[16]

Sengupta, M

B. Sengupta, M. Alrubayan, M. Kolla, Y . Wang, E. Mallet, A. Torres, R. Solis, H. Wang, and P. Pradhan, AI -Based Detection of Optical Microscopic Images of Pseudomonas aeruginosa in Planktonic and Biofilm States, Information 16, 4 (2025)

2025
[17]

Jarkman, M

S. Jarkman, M. Karlberg, M. Pocevičiūtė, A. Bodén, P. Bándi, G. Litjens, C. Lundström, D. Treanor, and J. van der Laak, Generalization of Deep Learning in Digital Pathology: Experience in Breast Cancer Metastasis Detection, Cancers (Basel) 14, 5424 (2022)

2022
[18]

Greeley, L

C. Greeley, L. Holder, E. E. Nilsson, and M. K. Skinner, Scalable deep learning artificial intelligence histopathology slide analysis and validation, Sci Rep 14, 26748 (2024)

2024
[19]

Joshua, K

A. Joshua, K. E. Allen, and N. M. Orsi, An Overview of Artificial Intelligence in Gynaecological Pathology Diagnostics, Cancers (Basel) 17, 1343 (2025)

2025

Pith tools

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

[1] [1]

J. W. Zhu, P. Charkhchi, and M. R. Akbari, Potential clinical utility of liquid biopsies in ovarian cancer, Molecular Cancer 21, 114 (2022)

work page 2022

[2] [2]

Andreou, M

M. Andreou, M. Kyprianidou, C. Cortas, I. Polycarpou, D. Papamichael, P. Kountourakis, and K. Giannakou, Prognostic Factors Influencing Survival in Ovarian Cancer Patients: A 10 -Year Retrospective Study, Cancers (Basel) 15, 5710 (2023)

work page 2023

[3] [3]

Staging for OVFTP Malignancies | Figo, https://www.figo.org/news/staging-ovftp-malignancies

work page

[4] [4]

P. S. Ginter, R. Idress, T. M. D’Alfonso, S. Fineberg, S. Jaffer, A. K. Sattar, A. Chagpar, P. Wilson, and M. Harigopal, Histologic grading of breast carcinoma: a multi-institution study of interobserver variation using virtual microscopy, Modern Pathology 34, 701 (2021)

work page 2021

[5] [5]

Javadi, D

S. Javadi, D. M. Ganeshan, A. Qayyum, R. B. Iyer, and P. Bhosale, Ovarian Cancer, the Revised FIGO Staging System, and the Role of Imaging, American Journal of Roentgenology 206, 1351 (2016)

work page 2016

[6] [6]

M. Woynarowski, Inter-observer variability in histopathological assessment of liver biopsies taken in a pediatric open-label therapeutic program for chronic HBV infection treatment, World Journal of Gastroenterology 12, (2006)

work page 2006

[7] [7]

M. G. Hanna et al., Implementation of Digital Pathology Offers Clinical and Operational Increase in Efficiency and Cost Savings, Archives of Pathology & Laboratory Medicine 143, 1545 (2019)

work page 2019

[8] [8]

I. D. Mienye, T. G. Swart, G. Obaido, M. Jordan, and P. Ilono, Deep Convolutional Neural Networks in Medical Image Analysis: A Review, Information 16, 3 (2025)

work page 2025