Manual tumor annotations in TCGA

Zenodo (CERN European Organization for Nuclear Research)

Abstract

What is this These are manual annotations of tumor tissue on TCGA diagnostic whole slide images in major solid tumor types. The aim of this project was to enrich for regions with invasive tumor tissue for subsequent molecular prediction studies, excluding whitespace, artifacts and non-tumor tissue as efficiently as possible. The aim was not to create a perfect tumor annotation on the pixel level. Annotations were done by trained observers using QuPath v0.1.2 and were converted to CSV. "COAD" and "READ" were merged to "CRC". More resources For difference between diagnostic and frozen slides, please see: http://www.andrewjanowczyk.com/download-tcga-digital-pathology-images-ffpe/ For a list of all tumor types in TCGA, see, see: https://gdc.cancer.gov/resources-tcga-users/tcga-code-tables/tcga-study-abbreviations Legal No guarantees, no liability.

Authors 2

1. Chiara Maria Lavinia Loeffler Aachen

   RWTH Aachen University · Universitätsklinikum Aachen

   Affiliation as printed

   University Hospital RWTH Aachen

2. Jakob Nikolas Kather Aachen

   RWTH Aachen University · Universitätsklinikum Aachen

   Affiliation as printed

   University Hospital RWTH Aachen