Datasets:

Formats:
parquet
Languages:
English
ArXiv:
Libraries:
Datasets
pandas
License:

Request for full version data

#3
by che111 - opened

Dear Authors,

Thanks for your impressive work and effort on building this comprehensive dataset!

May I kindly ask the full version of this dataset?
Now the data size seems only few samples, if it possible, could you release all dataset on HF?

Lab-Rasool org
edited May 17

Thank you for your interest in our project!

We are currently extracting embeddings using the UNI model for every whole slide image available in The Cancer Genome Atlas (TCGA). Additionally, we are processing and preparing to make radiology embeddings available in the coming week. In the meantime, I am pleased to inform you that the following data is already up and available for use:

  • Clinical Data (gatortron-base) for all ~11,000 patients
  • Pathology Report (gatortron-base) for all ~11,000 patients' pathology reports
  • Slide Image (UNI) for Lung Adenocarcinoma (LUAD) patients

While we continue to work on expanding the available data, I encourage you to explore our multimodal data aggregation tool called MINDS, which we built to acquire the raw data. You can find more information about MINDS in our recently published paper, available at: https://pubmed.ncbi.nlm.nih.gov/38475170/ and the codebase is open-source https://github.com/lab-rasool/MINDS. Lastly, I invite you to check out our open-source embedding processing tool, HoneyBee, which we used to generate these embeddings. You can access the HoneyBee repository on GitHub at: https://github.com/lab-rasool/HoneyBee/tree/main

Aakash-Tripathi changed discussion status to closed

Sign up or log in to comment