Dataset is available here: https://data.mendeley.com/datasets/kztymsrjx9/1
\nI found in the rules that \"C. External Data. You may use data other than the Competition Data (“External Data”) to develop and test your Submissions. However, you will ensure the External Data is publicly available and equally accessible to use by all participants of the Competition for purposes of the competition at no cost to the other participants. The ability to use External Data under this Section 7.C (External Data) does not limit your other obligations under these Competition Rules, including but not limited to Section 11 (Winners Obligations).\"
\nNow that all participants can obtain this dataset for free, does it mean that we can use it for training? If yes, I will try to find more external dataset.
", "messages": [ { "id": 2563642, "postDate": "2023-12-16T13:36:25.747Z", "content": "Dataset is available here: https://data.mendeley.com/datasets/kztymsrjx9/1
\nI found in the rules that \"C. External Data. You may use data other than the Competition Data (“External Data”) to develop and test your Submissions. However, you will ensure the External Data is publicly available and equally accessible to use by all participants of the Competition for purposes of the competition at no cost to the other participants. The ability to use External Data under this Section 7.C (External Data) does not limit your other obligations under these Competition Rules, including but not limited to Section 11 (Winners Obligations).\"
\nNow that all participants can obtain this dataset for free, does it mean that we can use it for training? If yes, I will try to find more external dataset.
", "rawMarkdown": "Dataset is available here: https://data.mendeley.com/datasets/kztymsrjx9/1\n\nI found in the rules that \"C. External Data. You may use data other than the Competition Data (“External Data”) to develop and test your Submissions. However, you will ensure the External Data is publicly available and equally accessible to use by all participants of the Competition for purposes of the competition at no cost to the other participants. The ability to use External Data under this Section 7.C (External Data) does not limit your other obligations under these Competition Rules, including but not limited to Section 11 (Winners Obligations).\"\n\nNow that all participants can obtain this dataset for free, does it mean that we can use it for training? If yes, I will try to find more external dataset.", "votes": 24 }, { "id": 2565935, "postDate": "2023-12-18T11:08:47.973Z", "content": "I've found 3 additional sources that could be used. their links were dead about a week ago so I contacted their support, and now it's accessible again. The datasets below are H&E datasets containing whole slide images. They might be a little big to handle, but they are at least in svs/tif formats so it should be easier than png.
\nP.s: Perhaps you can turn this into a general external data thread?
\nPTRC hgsoc
\nDescription: High grade serous cancer (HGSC label) dataset.
\nMagnification: 20x
\nmicron per pixel (um): (0.496,0.496)
\nlicense: CC BY 4.0
\ndownload link: https://www.cancerimagingarchive.net/collection/ptrc-hgsoc/
Bevacizumab response
\nDescription: Study on the use of Bevacizumab in patients with advanced ovarian cancer. I could not find anything about subtypes here. Perhaps someone can help. Otherwise, it could be used for unsupervised/self-supervised pre-training without labels on ovarian tissue.
\nMagnification: 20x
\nmicron per pixel (um): (0.5, 0.5)
\nlicense: CC BY 4.0
\nhttps://www.cancerimagingarchive.net/collection/ovarian-bevacizumab-response/
CPTAC-OV
\nDescription: Clinical Proteomic Tumor Analysis Consortium CPTAC Ovarian Serous Cystadenocarcinoma cohort. Again serous cancer, but I could not find distinctions between low grade (lgsc) and high grade (hgsc) subtypes. Any help is appreciated.
\nMagnification: 40x
\nmicron per pixel (um): (0.250,0.250)
\nlicense: CC BY 3.0
\nhttps://www.cancerimagingarchive.net/collection/cptac-ov/
Can attest. This was shared (see caveat here https://www.kaggle.com/competitions/UBC-OCEAN/discussion/457249)
\nI am able to access, download, and use data at https://data.mendeley.com/datasets/kztymsrjx9/1 without payment.
\nThe data has a CC license.
\nThe dataset is ~ 500 images with most sized between 400X650 to 550x880 px.
\nThey are not magnified consistently.
@ZZNZNB thank you for sharing.
\nLet us know if you find more.
Thank you for sharing. Similarly, after seeing your post, I found this https://data.mendeley.com/datasets/w39zgksp6n/1 from the same website. The contributor is the same so I guess they are somewhat related although the number of classes are fewer here.
", "rawMarkdown": "Thank you for sharing. Similarly, after seeing your post, I found this [https://data.mendeley.com/datasets/w39zgksp6n/1](https://data.mendeley.com/datasets/w39zgksp6n/1) from the same website. The contributor is the same so I guess they are somewhat related although the number of classes are fewer here. ", "votes": 5 }, { "id": 2569503, "postDate": "2023-12-21T09:53:38.947Z", "content": "TCGA has also many slides, but most of them are frozen, not sure how usable those are.
\nhttps://portal.gdc.cancer.gov/projects/TCGA-OV
I've found 3 additional sources that could be used. their links were dead about a week ago so I contacted their support, and now it's accessible again. The datasets below are H&E datasets containing whole slide images. They might be a little big to handle, but they are at least in svs/tif formats so it should be easier than png.
\nP.s: Perhaps you can turn this into a general external data thread?
\nPTRC hgsoc
\nDescription: High grade serous cancer (HGSC label) dataset.
\nMagnification: 20x
\nmicron per pixel (um): (0.496,0.496)
\nlicense: CC BY 4.0
\ndownload link: https://www.cancerimagingarchive.net/collection/ptrc-hgsoc/
Bevacizumab response
\nDescription: Study on the use of Bevacizumab in patients with advanced ovarian cancer. I could not find anything about subtypes here. Perhaps someone can help. Otherwise, it could be used for unsupervised/self-supervised pre-training without labels on ovarian tissue.
\nMagnification: 20x
\nmicron per pixel (um): (0.5, 0.5)
\nlicense: CC BY 4.0
\nhttps://www.cancerimagingarchive.net/collection/ovarian-bevacizumab-response/
CPTAC-OV
\nDescription: Clinical Proteomic Tumor Analysis Consortium CPTAC Ovarian Serous Cystadenocarcinoma cohort. Again serous cancer, but I could not find distinctions between low grade (lgsc) and high grade (hgsc) subtypes. Any help is appreciated.
\nMagnification: 40x
\nmicron per pixel (um): (0.250,0.250)
\nlicense: CC BY 3.0
\nhttps://www.cancerimagingarchive.net/collection/cptac-ov/
Can attest. This was shared (see caveat here https://www.kaggle.com/competitions/UBC-OCEAN/discussion/457249)
\nI am able to access, download, and use data at https://data.mendeley.com/datasets/kztymsrjx9/1 without payment.
\nThe data has a CC license.
\nThe dataset is ~ 500 images with most sized between 400X650 to 550x880 px.
\nThey are not magnified consistently.
@ZZNZNB thank you for sharing.
\nLet us know if you find more.
Thank you for sharing. Similarly, after seeing your post, I found this https://data.mendeley.com/datasets/w39zgksp6n/1 from the same website. The contributor is the same so I guess they are somewhat related although the number of classes are fewer here.
", "votes": 5, "replies": [] }, { "id": 2569503, "author_name": "crunch01", "author_url": "", "post_date": "2023-12-21T09:53:38.947000", "content": "TCGA has also many slides, but most of them are frozen, not sure how usable those are.
\nhttps://portal.gdc.cancer.gov/projects/TCGA-OV