Welcome to the website of PAIP 2019 Challenge.
This competition is part of the MICCAI 2019 Grand Challenge for Pathology.

Announcement

  • First training dataset released!

Important Dates

  •    April 15, 2019               Release of the training data (Phase 1) 
  •    May (TBA) 2019             Release of the training data (Phase 2)
  •    August 15, 2019            Release of the validation and test data
  •    September 15, 2019       Submission deadline for results

Overview

The goal of the challenge is to evaluate new and existing algorithms for automated detection of liver cancer in whole-slide images (WSIs). There are two tasks and therefore two leaderboards for evaluating the performance of the algorithms. Participants can choose to join both or either tasks according to their interests.

Task 1: Liver Cancer Segmentation
Task 2: Viable Tumor Burden Estimation

Background


The liver is a visceral organ most often involved in the metastatic spread of cancer. For the best practice, early diagnosis of liver cancer is important but many more people don't even know that they have hepatitis. Hepatocellular Carcinoma(HCC) represents about 90% of primary liver cancers and constitutes a major global health problem. The incidence of HCC is increasing both in Korea and worldwide; it is amongst the leading causes of cancer mortality globally. Between 1990 and 2015 newly diagnosed HCC cases increased by 75%, mainly due to changing age structures and population growth.

A tumor is composed of various cellular and stromal components, eg tumor cells, inflammatory cells, blood vessels, acellular matrix, tumor capsule, fluid, mucin, or necrosis. The viable tumor burden is defined as the ratio of viable tumor area to the whole area of the tumor. The need for evaluation of viable tumor burden is increasing, as an assessment of response rates for chemoradiotherapy or proportion of tumor cells in genetic testing using tissue samples. Traditional pathologists use a semiquantitative grading system for residual tumor burden or report portion of necrosis indirectly indicating viable tumor burden.

Dataset

In the challenge, participants will be provided with 2 levels of data set extracted from Whole Slide Images.

  1. Tumor with prominent peritumoral reaction
  2. Tumor with minimal peritumoral or intratumoral reaction

The data and segmentation of the entire tumor area and viable tumor area are provided by Seoul National University Hospital, South Korea. The data are fully annotated by expert pathologists and are divided into 3 groups of data sets. 

  • The training data set contains 50 WSIs
  • The validation data set contains 10 WSIs
  • The test data set contains 40 WSIs

All WSIs were scanned at 20X magnification and all cases are randomly selected irrespective of the institutions.

The ground truth information of both tasks such as tumor segmentation is given to participants for the training set. For the validation and test set the ground, however, the truth information is reserved to the challenge committee and will be used to evaluate the performance of participant's AI learning models. (See the Detailed Data Description)

How to Participate

  1. Read the challenge rules carefully
    1. Register a grand-chllenge.org account
    2. Join the PAIP2019 challenge
    3. To download the dataset, visit here to fill and sign the DATA USE AND CONFIDENTIALITY AGREEMENT (It will be reviewed within 1 business day)
    4. An e-mail with the link and access credentials to the dataset will be sent to your contact information
    5. Submissions will be enabled when the testing data has been released on August 15, 2019


    If you have questions or comments, post a message on the forum or mail to wisepaip@gmail.com



    This research project is funded by
    the Ministry of Health and Welfare, Republic of Korea.