Published on in Vol 23, No 2 (2021): February

Preprints (earlier versions) of this paper are available at https://preprints.jmir.org/preprint/23436, first published .
Hidden Variables in Deep Learning Digital Pathology and Their Potential to Cause Batch Effects: Prediction Model Study

Hidden Variables in Deep Learning Digital Pathology and Their Potential to Cause Batch Effects: Prediction Model Study

Hidden Variables in Deep Learning Digital Pathology and Their Potential to Cause Batch Effects: Prediction Model Study

Max Schmitt   1 * , MSc ;   Roman Christoph Maron   1 * , MSc ;   Achim Hekler   1 , MSc ;   Albrecht Stenzinger   2 , MD ;   Axel Hauschild   3 , MD ;   Michael Weichenthal   3 , MD ;   Markus Tiemann   4 , MD ;   Dieter Krahl   5 , MD ;   Heinz Kutzner   6 , MD ;   Jochen Sven Utikal   7, 8 , MD ;   Sebastian Haferkamp   9 , MD ;   Jakob Nikolas Kather   10 , MD ;   Frederick Klauschen   11 , MD ;   Eva Krieghoff-Henning   1 , PhD ;   Stefan Fröhling   12 , MD ;   Christof von Kalle   13 , MD ;   Titus Josef Brinker   1 , MD

1 Digital Biomarkers for Oncology Group, National Center for Tumor Diseases, German Cancer Research Center (DKFZ), Heidelberg, Germany

2 Institute of Pathology, University Hospital Heidelberg, University of Heidelberg, Heidelberg, Germany

3 Department of Dermatology, University Hospital Kiel, University of Kiel, Kiel, Germany

4 Institute for Hematopathology Hamburg, Hamburg, Germany

5 Private Institute of Dermatopathology, Heidelberg, Germany

6 Private Institute of Dermatopathology, Friedrichshafen, Germany

7 Skin Cancer Unit, German Cancer Research Center (DKFZ), Heidelberg, Germany

8 Department of Dermatology, University Medical Center Mannheim, University of Heidelberg, Mannheim, Germany

9 Department of Dermatology, University Hospital of Regensburg, Regensburg, Germany

10 Department of Medicine III, RWTH University Hospital Aachen, Aachen, Germany

11 Institute of Pathology, Charité University Hospital Berlin, Berlin, Germany

12 National Center for Tumor Diseases, German Cancer Research Center (DKFZ), Heidelberg, Germany

13 Department of Clinical-Translational Sciences, Charité and Berlin Institute of Health, Berlin, Germany

*these authors contributed equally

Corresponding Author:

  • Titus Josef Brinker, MD
  • Digital Biomarkers for Oncology Group
  • National Center for Tumor Diseases
  • German Cancer Research Center (DKFZ)
  • Im Neuenheimer Feld 460
  • Heidelberg, 69120
  • Germany
  • Phone: 49 6221-3219304
  • Email: titus.brinker@dkfz.de