Other Open Access
Wouter Bulten; Geert Litjens; Hans Pinckaers; Peter Ström; Martin Eklund; Kimmo Kartasalo; Maggie Demkin; Sohier Dane
This is the challenge design document for the "The PANDA challenge: Prostate cANcer graDe Assessment using the Gleason grading system", accepted for MICCAI 2020.
With 1.1 million new diagnoses every year, prostate cancer (PCa) is the most common cancer in men in developed countries. The biopsy Gleason grading system is the most important prognostic marker for prostate cancer but suffers from significant inter-observer variability, limiting its usefulness for individual patients. The Gleason grade is determined by pathologists on hematoxylin and eosin (H&E) stained tissue specimens based on the architectural growth patterns of the tumor.
Automated deep learning systems have shown promise in accurately grading prostate cancer. Several studies have shown that these systems can achieve pathologist-level performance. A large multi-center evaluation on diagnostic data is still missing. In this challenge, we strive to improve on these works by publishing the most extensive multi-center dataset on Gleason grading as of yet. The training set consists of up to 11.000 whole-slide images of digitized H&E-stained biopsies originating from two centers. This is the largest public whole-slide image dataset available, roughly 8 times the size of the CAMELYON17 challenge. Furthermore, in contrast to previous challenges, we do not use small tissue micro-arrays, but full diagnostic biopsy images. Using a sizeable multicenter test set, graded by expert uropathologists, we will evaluate challenge submissions on their applicability to a clinically relevant task