Find research datasets worth reusing

opencc-by-4.0Apr 2024View details →

Histopathological Evaluation of Abdominal Aortic Aneurysms with Deep Learning: The HistAAA Dataset

This dataset accompanies our manuscript "Histopathological Evaluation of Abdominal Aortic Aneurysms with Deep Learning" (<a title="Histopathological evaluation of abdominal aortic aneurysms with deep learning" href="https://www.medrxiv.org/content/10.1101/2024.04.23.24306178v1">1</a>) and comprises feature vectors extracted from digital whole-slide images (WSI) of abdominal aortic aneurysm (AAA) wall samples from 369 patients treated at three European centers and corresponding expert pathologist annotations. This dataset is intended to be used for computational modelling tasks including automated prediction of pathology-related variables such as inflammation, degradation of elastic fibers, and fibrosis (<a title="Histopathological evaluation of abdominal aortic aneurysms with deep learning" href="https://www.medrxiv.org/content/10.1101/2024.04.23.24306178v1">1</a>).  Code for preprocessing is available at <a href="https://github.com/KatherLab/STAMP">https://github.com/KatherLab/STAMP</a>. Code for modeling is available at <a href="https://github.com/KatherLab/marugoto">https://github.com/KatherLab/marugoto</a>. Code for spatial heatmaps and top-attention tiles is available at <a href="https://github.com/KatherLab/highres-WSI-heatmaps/tree/AAA_heatmaps">https://github.com/KatherLab/highres-WSI-heatmaps/tree/AAA_heatmaps</a>.  <h3>Patient Cohort</h3> Data from a total of 369 patients (84.6% male, mean age 69.1 ± 8.0 years, average maximum diameter 62.9 ± 15.7 mm) undergoing open AAA repair at the Technical University Munich (TUM, n = 287), the University Hospital Würzburg (UHW, n = 36) and the Medical University Vienna (MUV, n = 46), between 2005 and 2019, are included in this dataset.  To generate annotations, aneurysm samples from the left anterior wall were independently analyzed by three pathologists as described previously (<a title="Abdominal aortic aneurysms harbor different histomorphology not associated with classic risk factors &ndash; the HistAAA study" href="https://www.medrxiv.org/content/10.1101/2024.04.16.24305904v1">2</a>). In brief, the following histopathological parameters were evaluated: Grade of inflammation in tunica media [none, minor, intermediate, major], grade of inflammation in adventitia [none, minor, intermediate, major], type of inflammation in adventitia and tunica media [none, acute, chronic], Histological Inflammation Scale of Aneurysms (HISA) grade (<a title="Inflammation of the abdominal aortic aneurysm wall" href="https://www.sciencedirect.com/science/article/abs/pii/S0950821X05801185">3</a>) [0, 1, 2, 3, 4], angiogenesis in tunica media [present, not present], calcification in tunica media [present, not present], grade of fibrosis in adventitia [minor, intermediate, major], remaining elastic fibers in tunica media [< 25%, > 25%]. In case of disagreement, consensus was reached through discussion. Hematoxylin and Eosin (HE)- and Elastica van Gieson (EvG)-stained slides were digitized using an Aperio AT2 (Leica, Wetzlar, Germany) slide scanner. Patient sex and smoking history were gathered as binary clinical parameters.  <h3>Preprocessing and Feature Extraction</h3> The STAMP protocol was utilized to process the WSIs (<a title="From Whole-slide Image to Biomarker Prediction: A Protocol for End-to-End Deep Learning in Computational Pathology" href="https://arxiv.org/abs/2312.10944">4</a>). In brief, WSI were preprocessed by tessellation into 224 x 224 pixel patches at a magnification of 256 µm per pixel, followed by computational background rejection, and color normalization of HE-stained slides. EvG-stained slides were not color-normalized as no color normalization protocols exist for this staining. Subsequently, features were extracted using CTranspath (<a title="Transformer-based unsupervised contrastive learning for histopathological image classification" href="https://www.sciencedirect.com/science/article/abs/pii/S1361841522002043">5</a>), a pre-trained histology image encoder.    <h3>References</h3> 1. Kolbinger, F. R. et al. Histopathological Evaluation of Abdominal Aortic Aneurysms with Deep Learning. medRxiv 2024.04.23.24306178 (2024) doi:10.1101/2024.04.23.24306178.  2. Nackenhorst, M. C. et al. Abdominal aortic aneurysms harbor different histomorphology not associated with classic risk factors – the HistAAA study. medRxiv 2024.04.16.24305904 (2024) doi:10.1101/2024.04.16.24305904.  3. Rijbroek, A., Moll, F. L., von Dijk, H. A., Meijer, R. & Jansen, J. W. Inflammation of the abdominal aortic aneurysm wall. Eur. J. Vasc. Surg. 8, 41–46 (1994). 4. El Nahhas, O. S. M. et al. From Whole-slide Image to Biomarker Prediction: A Protocol for End-to-End Deep Learning in Computational Pathology. arXiv [cs.CV] (2023). 5. Wang, X. et al. Transformer-based unsupervised contrastive learning for histopathological image classification. Med. Image Anal. 81, 102559 (2022).

opencc-by-4.0May 2024View details →

Histopathology pairs of hematoxylin-eosin and Ki67 stainings of testicular seminoma

About Dataset contains 77 pairs of stained sections of testicular seminoma. Each pair is produced from the neighbouring sections of the tissue to guarantee the most similarity possible of the stained tissue. HE and Ki67 images are registered and aligned. When used, please cite the following work: Petríková, D.; Cimrák, I.; Tobiá&scaron;ová, K. and Plank, L. (2024), Dataset of Hematoxylin-eosin and Ki67 histopathological image pairs complemented by algorithms providing Ki67 index, DATA, submitted. Data composition: Dataset is comprised of 154 PNG files. 77 files with prefix id_HE contain images of HE staining, and 77 files with prefix id_Ki67 contain images of Ki67 staining. Both files in each pair id_HE_string.png and id_Ki67_string.png have the same resolution with avarage value 35000px x 35000px. Data is available in 39 ZIP files each containig two HE. images and two of their Ki67 counterparts. For example, 03.zip contains the following files: <ul> <li>05_HE_slide-011-A1-S1-B11.png,</li> <li>05_Ki67_slide-012-A1-S2-B12.png,</li> <li>06_HE_slide-013-A1-S1-B13.png,</li> <li>06_Ki67_slide-014-A1-S2-B14.png.</li> </ul> Image names contain an additional random string.  An example of a HE - Ki67 pair is in file  HE-Ki67_pair.png. Data preprocessing: Original scans of tissues contained pairs of HE and Ki67 staining with rotated and shifted tissues. Therefore a preprocessing of image registration was performed to find an affine transformation for rotation and translation. The concrete steps of image registration was described in [Petrikova 2024]. Provided images are already aligned. Data usage: This dataset is aimed for AI models to predict the Ki67 index or even to generate Ki67 staining. Since one pair of HE and Ki67 images contain two physically different sections of tissue (although the neighbouring), there is no one-to-one correspondence on cellural level. However patches, created on the same locations from the images in one pair, containe similar quantitative characteristics, such as number of cells, number of Ki67 positive cells, averaged cell size, etc. In [Petrikova2024B] we elaborate the usage of the dataset by computing Ki67 index of HE patches evaluated from corresponding Ki67 patch.     References: [Petrikova2024] Petríková, D.; Cimrák, I.; Tobiá&scaron;ová, K. and Plank, L. (2024). Ki67 Expression Classification from HE Images with Semi-Automated Computer-Generated Annotations. In Proceedings of the 17th International Joint Conference on Biomedical Engineering Systems and Technologies - BIOINFORMATICS; ISBN 978-989-758-688-0; ISSN 2184-4305, SciTePress, pages 536-544. DOI: 10.5220/0012535900003657 [Petrikova2024B] Petríková, D.; Cimrák, I.; Tobiá&scaron;ová, K. and Plank, L. (2024), Dataset of Hematoxylin-eosin and Ki67 histopathological image pairs complemented by algorithms providing Ki67 index, DATA, submitted.

opencc-by-sa-4.0Mar 2018View details →

Histopathology data of bone marrow biopsies (HistBMP or HistMNIST)

Data information We prepared a dataset basing on histopathological images freely available on-line (http://www.enjoypath.com/). We selected 16 patients (patient IDs: 272, 274, 283, 289, 290, 291, 292, 295, 297, 298, 299). Each histopathological image represents a bone marrow biopsy. Diagnoses of the chosen cases were associated with different kinds of cancer (e.g., lymphoma, leukemia) or anemia. All original images were taken using HE, 40×, and each image was of size 336 × 448. Data preparation The original RGB representation was transformed to gray scale. Further, we divided each image into small patches of size 28 × 28. Eventually, we picked 10 patients for training, 3 patients for validation and 3 patients for testing, which resulted in 6,800 training images, 2,000 validation images and 2,000 test images. The selection of patients was performed in such a fashion that each dataset contained representative images with different diagnoses and amount of fat. Since the small patches resemble a widely-used benchmark in machine learning/AI community called MNIST, the dataset is referred to as HistMNIST.  First usage The dataset was used to train deep generative models (VAEs): <ul> <li>Tomczak, J. M., & Welling, M. (2016). Improving variational auto-encoders using householder flow. arXiv preprint arXiv:1611.09630.</li> </ul>

opencc-by-4.0Aug 2019View details →

Representative Sample Dataset for Resolution-Agnostic Tissue Segmentation in Whole-Slide Histopathology Images

This is a representative sample from the dataset that was used to develop resolution-agnostic convolutional neural networks for tissue segmentation1 in whole-slide histopathology images. The dataset is composed of two parts: development set and dissimilar set. Sample images from the development set: <ul> <li>breast_hne_00.tif</li> <li>breast_lymph_node_hne_00.tif</li> <li>tongue_ae1ae3_00.tif</li> <li>tongue_hne_00.tif</li> <li>tongue_ki67_00.tif</li> </ul> Sample images from the dissimilar set: <ul> <li>brain_alcianblue_00.tif</li> <li>cornea_grocott_00.tif</li> <li>kidney_cab_00.tif</li> <li>skin_perls_00.tif</li> <li>uterus_vonkossa_00.tif</li> </ul>

Mice infected with High shedder S. mansoni parasites from cross A - cage 1 - Liver histopathology data.

The present dataset contains all the histopathology images used to quantify fibrotic areas, parasite egg counts and to quantify granuloma areas in liver of mice infected with S. mansoni High shedder line. These data are presented in the manuscript entitled "No evidence for schistosome parasite fitness trade-offs in the intermediate and definitive host" (dataset # 2/11). Each folder corresponds to one mouse sample and contain, along with a readme file, all the files used to quantify fibrotic area and egg counts (_TRICH.czi), to quantify granuloma area (_HE.czi), and the annotation file (.annotations) containing all the annotated granuloma areas.

Mice infected with High shedder S. mansoni parasites from cross A - cage 2 - Liver histopathology data.

The present dataset contains all the histopathology images used to quantify fibrotic areas, parasite egg counts and to quantify granuloma areas in liver of mice infected with S. mansoni High shedder line. These data are presented in the manuscript entitled "No evidence for schistosome parasite fitness trade-offs in the intermediate and definitive host" (dataset # 3/11). Each folder corresponds to one mouse sample and contain, along with a readme file, all the files used to quantify fibrotic area and egg counts (_TRICH.czi), to quantify granuloma area (_HE.czi), and the annotation file (.annotations) containing all the annotated granuloma areas.

Mice infected with High shedder S. mansoni parasites from cross B - cage 1 - Liver histopathology data.

The present dataset contains all the histopathology images used to quantify fibrotic areas, parasite egg counts and to quantify granuloma areas in liver of mice infected with S. mansoni High shedder line. These data are presented in the manuscript entitled "No evidence for schistosome parasite fitness trade-offs in the intermediate and definitive host" (dataset # 6/11). Each folder corresponds to one mouse sample and contain, along with a readme file, all the files used to quantify fibrotic area and egg counts (_TRICH.czi), to quantify granuloma area (_HE.czi), and the annotation file (.annotations) containing all the annotated granuloma areas.

Mice infected with Low shedder S. mansoni parasites from cross A - cage 2 - Liver histopathology data.

The present dataset contains all the histopathology images used to quantify fibrotic areas, parasite egg counts and to quantify granuloma areas in liver of mice infected with S. mansoni Low shedder line. These data are presented in the manuscript entitled "No evidence for schistosome parasite fitness trade-offs in the intermediate and definitive host" (dataset # 5/11). Each folder corresponds to one mouse sample and contain, along with a readme file, all the files used to quantify fibrotic area and egg counts (_TRICH.czi), to quantify granuloma area (_HE.czi), and the annotation file (.annotations) containing all the annotated granuloma areas.

Mice infected with Low shedder S. mansoni parasites from cross A - cage 1 - Liver histopathology data.

The present dataset contains all the histopathology images used to quantify fibrotic areas, parasite egg counts and to quantify granuloma areas in liver of mice infected with S. mansoni Low shedder line. These data are presented in the manuscript entitled "No evidence for schistosome parasite fitness trade-offs in the intermediate and definitive host" (dataset # 4/11). Each folder corresponds to one mouse sample and contain, along with a readme file, all the files used to quantify fibrotic area and egg counts (_TRICH.czi), to quantify granuloma area (_HE.czi), and the annotation file (.annotations) containing all the annotated granuloma areas.

Control mice (non-infected with S. mansoni parasites) - cage 2 - Liver histopathology data (mouse ID 2C.1 / 2C.2 / 2C.3).

The present dataset contains all the histopathology images used to quantify fibrotic areas in liver of mice (non-infected with S. mansoni parasite). These data are presented in the manuscript entitled "No evidence for schistosome parasite fitness trade-offs in the intermediate and definitive host" (dataset # 11.1/11). Each folder corresponds to one mouse sample and contain, along with a readme file, all the files used to quantify fibrotic area (_TRICH.czi) and (_HE.czi) files.

Control mice (non-infected with S. mansoni parasites) - cage 2 - Liver histopathology data (mouse ID 2C.4 / 2C.5).

The present dataset contains all the histopathology images used to quantify fibrotic areas in liver of mice (non-infected with S. mansoni parasite). These data are presented in the manuscript entitled "No evidence for schistosome parasite fitness trade-offs in the intermediate and definitive host" (dataset # 11.2/11). Each folder corresponds to one mouse sample and contain, along with a readme file, all the files used to quantify fibrotic area (_TRICH.czi) and (_HE.czi) files.

Mice infected with High shedder S. mansoni parasites from cross B - cage 2 - Liver histopathology data.

The present dataset contains all the histopathology images used to quantify fibrotic areas, parasite egg counts and to quantify granuloma areas in liver of mice infected with S. mansoni High shedder line. These data are presented in the manuscript entitled "No evidence for schistosome parasite fitness trade-offs in the intermediate and definitive host" (dataset # 7/11). Each folder corresponds to one mouse sample and contain, along with a readme file, all the files used to quantify fibrotic area and egg counts (_TRICH.czi), to quantify granuloma area (_HE.czi), and the annotation file (.annotations) containing all the annotated granuloma areas.

Mice infected with Low shedder S. mansoni parasites from cross B - cage 2 - Liver histopathology data.

The present dataset contains all the histopathology images used to quantify fibrotic areas, parasite egg counts and to quantify granuloma areas in liver of mice infected with S. mansoni Low shedder line. These data are presented in the manuscript entitled "No evidence for schistosome parasite fitness trade-offs in the intermediate and definitive host" (dataset # 9/11). Each folder corresponds to one mouse sample and contain, along with a readme file, all the files used to quantify fibrotic area and egg counts (_TRICH.czi), to quantify granuloma area (_HE.czi), and the annotation file (.annotations) containing all the annotated granuloma areas.

Mice infected with Low shedder S. mansoni parasites from cross B - cage 1 - Liver histopathology data.

The present dataset contains all the histopathology images used to quantify fibrotic areas, parasite egg counts and to quantify granuloma areas in liver of mice infected with S. mansoni Low shedder line. These data are presented in the manuscript entitled "No evidence for schistosome parasite fitness trade-offs in the intermediate and definitive host" (dataset # 8/11). Each folder corresponds to one mouse sample and contain, along with a readme file, all the files used to quantify fibrotic area and egg counts (_TRICH.czi), to quantify granuloma area (_HE.czi), and the annotation file (.annotations) containing all the annotated granuloma areas.

Integrative in situ mapping of single-cell transcriptional states and tissue histopathology in an Alzheimer disease model

Amyloid-β plaques and neurofibrillary tau tangles are the neuropathologic hallmarks of Alzheimer’s disease (AD), but the spatiotemporal cellular responses and molecular mechanisms underlying AD pathophysiology remain poorly understood. Here we introduce STARmap PLUS to simultaneously map single-cell transcriptional states and disease marker proteins in brain tissues of AD mouse models at a voxel size of 95  95  350 nm. This high-resolution spatial transcriptomics map revealed a core-shell structure where disease-associated microglia (DAM) closely contact amyloid-β plaques, whereas disease-associated astrocyte-like cells (DAA-like) and oligodendrocyte precursor cells (OPC) are enriched in the outer shells surrounding the plaque-DAM complex. Hyperphosphorylated tau emerged mainly in excitatory neurons in the CA1 region accompanied by infiltration of oligodendrocyte subtypes into the axon bundles of hippocampal alveus. The integrative STARmap PLUS method bridges single-cell gene expression profiles with tissue histopathology at subcellular resolution, providing an unprecedented roadmap to pinpoint the molecular and cellular mechanisms of AD pathology and neurodegeneration.