Search

This is an anonymized dataset consisting of 50 biopsy cores that were graded by 11 pathologists from different countries. It is associated with the publication: "An international multi-institutional validation study of the algorithm for prostate cancer detection and Gleason grading" Tolkach Y. et al. NPJ Precision Oncology 2023 Characteristics of the dataset Source: WNS (for details see original publication). Number of slides: 50, balances for Gleason grade groups (GGG), 10 Slides per GGG 1-5 (based on majority votes). Number of pathologists: 10. Country of pathologists: Germany, Austria, USA, Netherlands, Japan, Russia, Israel. File format: OME TIFF Scanner: Hamamatsu S360. Resolution: 40x, mpp = 0.2305; original scanning resolution. Compression (OME TIFF): J2K lossy Generated in software: QuPath v. 0.4.3 Grading: slide-level. How to receive the grading results? To receive the Excel sheet with grading results please fill out the accompanying form "Grading results request form.pdf" and send it to: yuri.tolkach(at)gmail.com License and rules for using the dataset: 1) The usage of the dataset is allowed for academic research purposes only. Commercial use is not allowed. 2) The transfer of grading results to other persons that have not completed the Grading results request form is not allowed. 3) In case of using the dataset in a publication, the citation of original paper (Tolkach Y. et al. NPJ Precision Oncology 2023) is mandatory. 4) The PBGG dataset (whole-slide images) is distributed under the following License: Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) https://creativecommons.org/licenses/by-nc-sa/4.0/

This is an anonymized dataset consisting of 50 biopsy cores that were graded by 10 pathologists from different countries. It is associated with the publication: "An international multi-institutional validation study of the algorithm for prostate cancer detection and Gleason grading" Tolkach Y. et al. NPJ Precision Oncology 2023 Characteristics of the dataset Source: UKK (for details see original publication). Number of slides: 50, balances for Gleason grade groups (GGG), 10 Slides per GGG 1-5 (based on majority votes). Number of pathologists: 10. Country of pathologists: Germany, Austria, USA, Netherlands, Japan, Russia, Israel. File format: OME TIFF Scanner: Hamamatsu S360. Resolution: 40x, mpp = 0.2305; original scanning resolution. Compression (OME TIFF): J2K lossy Generated in software: QuPath v. 0.4.3 Grading: slide-level. How to receive the grading results? To receive the Excel sheet with grading results please fill out the accompanying form "Grading results request form.pdf" and send it to: yuri.tolkach(at)gmail.com License and rules for using the dataset: 1) The usage of the dataset is allowed for academic research purposes only. Commercial use is not allowed. 2) The transfer of grading results to other persons that have not completed the Grading results request form is not allowed. 3) In case of using the dataset in a publication, the citation of original paper (Tolkach Y. et al. NPJ Precision Oncology 2023) is mandatory. 4) The PBGG dataset (whole-slide images) is distributed under the following License: Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) https://creativecommons.org/licenses/by-nc-sa/4.0/

These are four test datasets for the publication "Artificial intelligence for tumor detection and histological regression grading in oesophageal adenocarcinomas: a retrospective algorithm development and validation study" by Tolkach Y. et al. For dataset details, tissue classes, and number of images in each slide see original publication. All image patches were extracted using QuPath with a size 256 px at resolution MPP = 0.7813. All images are native scan quality without any normalization. For training of algorithms we reduced patches to 220 px (MPP = 0.922) corresponding to regions of 200 x 200 µm.

Poster presented as part of the Crick BioImage Analysis Symposium. As rapidly developing deep learning methods minimize laborious image processing, they are able to significantly speed up the diagnosis prediction in the field of computational pathology. However, for accurate classification or precise segmentation a robust model is crucial. Often deep learning models show excellent results from one laboratory setting, however minor differences in staining protocols or image acquisition can lead to specific artefacts. Such artefacts fall into out-of-distribution data range and thereby tangle the automated deep learning image processing causing confident failures. To address this problem, we have developed a framework for out-of-distribution detection (FrOoDo) for detecting such artefacts in an ad-hoc fashion. Furthermore, we added advanced artefacts simulations to improve our frameworks reliability and showcase the necessity for prediction quality control. Our results show that while all artefacts have an increasingly adverse effect on the prediction performance w.r.t their severity. FrOoDo is perfect for reliable detection of more severe artefacts, thus allowing for the more accurate disease prediction. As our framework builds a foundation for quality control for in computational pathology, we want to extend FrOoDo to biomedical images and explore more applications in the future. Code is available at: https://github.com/MECLabTUDA/FrOoDo. Permission has been given by authors to upload to Crick Figshare. Copyright remains with the original holders.

Although deep federated learning has received much attention in recent years, progress has been made mainly in the context of natural images and barely for computational pathology. However, deep federated learning is an opportunity to create datasets that reflect the data diversity of many laboratories. Further, the effort of dataset construction can be divided among many. Unfortunately, existing algorithms cannot be easily applied to computational pathology since previous work presupposes that data distributions of laboratories must be similar. This is an unlikely assumption, mainly since different laboratories have different staining styles. As a solution, we propose BottleGAN, a generative model that can computationally align the staining styles of many laboratories and can be trained in a privacy-preserving manner to foster federated learning in computational pathology. We construct a heterogenic multi-institutional dataset based on the PESO segmentation dataset and improve the IOU by 42\% compared to existing federated learning algorithms. An implementation of BottleGAN is available at https://github.com/MECLabTUDA/BottleGAN, Comment: Accepted for Poster at MICCAI2022

"Quality control stress test for deep learning-based diagnostic model in digital pathology" B. Schömig-Markiefka et al. Modern Pathology, 2021 https://doi.org/10.1038/s41379-021-00859-x Six datasets containing each: 50.000 patches with prostate cancer tumor tissue 50.000 patches with benign glandular prostate tissue 20.000 patches with benign non-glandular prostate tissue For details to datasets (scanner, magnification, etc.) please refer to original publication. Part 2: Datasets 5-6. By usage of Datasets please cite the original publication: &nbsp

"Quality control stress test for deep learning-based diagnostic model in digital pathology" B. Schömig-Markiefka et al. Modern Pathology, 2021 https://doi.org/10.1038/s41379-021-00859-x Six datasets containing each: 50.000 patches with prostate cancer tumor tissue 50.000 patches with benign glandular prostate tissue 20.000 patches with benign non-glandular prostate tissue For details to datasets (scanner, magnification, etc.) please refer to original publication. Part 1: Datasets 1-4.

The current study detects different morphologies related to prostate pathology using deep learning models; these models were evaluated on 2,121 hematoxylin and eosin (H&E) stain histology images captured using bright field microscopy, which spanned a variety of image qualities, origins (whole slide, tissue micro array, whole mount, Internet), scanning machines, timestamps, H&E staining protocols, and institutions. For case usage, these models were applied for the annotation tasks in clinician-oriented pathology reports for prostatectomy specimens. The true positive rate (TPR) for slides with prostate cancer was 99.7% by a false positive rate of 0.785%. The F1-scores of Gleason patterns reported in pathology reports ranged from 0.795 to 1.0 at the case level. TPR was 93.6% for the cribriform morphology and 72.6% for the ductal morphology. The correlation between the ground truth and the prediction for the relative tumor volume was 0.987 n. Our models cover the major components of prostate pathology and successfully accomplish the annotation tasks.

Supplementary Data. List of abbreviations. (PDF 29 kb)

Supplementary Material. (PDF 233Â kb)

Expression of RNYs (PCR, ΔCq expression): non-muscle-invasive bladder cancer (NMIBC) vs muscle-invasive bladder cancer (NMIBC). (DOCX 15 kb)

Correlation matrix (Spearman rho) between the expression of different YRNAs (PCR, ΔCq Expression) in tumor tissue (all p

Expression of RNYs (PCR, ΔCq expression): patients stratified according to grade of tumor differentiation. (DOCX 15 kb)

Expression of RNYs (PCR, ΔCq expression): patients with lymph node metastases (pN+) vs patients without lymph node metastases (pN0). (DOCX 15 kb)

Correlation matrix (Spearman rho) between the expression of different YRNAs (PCR, ΔCq Expression) in tumor tissue (all p