Join

This page contains the results from submissions made while the challenge was hosted on the original platform.


Task 1: Prediction of the proliferation score based on mitosis counting

The evaluation metric for this task is the quadratic weighted Cohen's kappa between the predicted and ground truth proliferation scores.


Automatic methods

Team Score Use of additional data
Lunit Inc., Korea 0.567 No
Contextvision, Sweden (SLDESUTO-BOX) 0.534 No
Sectra, Sweden 0.462 Yes; negative ROI annotations
University of Heidelberg, Germany 0.417 No
IBM Research Zurich and Brazil 0.385 Yes; ICPR 2012 and 2014 datasets
The Harker School, United States 0.367 No
Belarus National Academy of Sciences 0.321 No
Radboud UMC Nijmegen, The Netherlands 0.290 No
University of South Florida, United States 0.177 No
University of Warwick, United Kingdom 0.159 Yes; negative ROI annotations
Technical University of Munich 0.108 No

Semi-automatic methods

Team Score Manual input
Microsoft Research Asia, China 0.543 ROI selection by a pathologist


Task 2: Prediction of the proliferation score based on gene expression

The evaluation metric for this task is the Spearman's correlation coefficient between the predicted and ground truth proliferation scores.


Automatic methods

Team Score Use of additional data
Lunit Inc., Korea 0.617 No
Radboud UMC Nijmegen, The Netherlands 0.516 No
Contextvision, Sweden (SLDESUTO-BOX) 0.503 No
Belarus National Academy of Sciences 0.494 No
The Harker School, United States 0.474 No

Semi-automatic methods

Team Score Manual input
Microsoft Research Asia, China 0.710 ROI selection by a pathologist


Task 3: Mitosis detection

The evaluation metric for this task is the F1-score of the detection.


Automatic methods

Team Score Use of additional data
Lunit Inc., Korea 0.652 No
IBM Research Zurich and Brazil 0.648 Yes; ICPR 2012 and 2014 datasets
Contextvision, Sweden (SLDESUTO-BOX) 0.616 No
The Chinese University of Hong Kong 0.601 Yes; ICPR 2012 and 2014 datasets
Microsoft Research Asia, China 0.596 Yes; ICPR 2012 and 2014 datasets
Radboud UMC, The Netherlands 0.541 No
University of Heidelberg, Germany 0.481 No
University of South Florida, United States 0.440 No
Pakistan Institute of Engineering and Applied Sciences 0.424 No
University of Warwick, United Kingdom 0.396 No
Shiraz University of Technology, Iran 0.330 No
Inha University, Korea 0.251 No
Instituto Politécnico Nacional, Mexico 0.135 No
Healthcare Technology Innovation Centre, IIT Madras, India 0.017 No

Results submitted after the conclusion of the challenge


Task 1: Prediction of the proliferation score based on mitosis counting

The evaluation metric for this task is the quadratic weighted Cohen's kappa between the predicted and ground truth proliferation scores.


Automatic methods

Team Score Use of additional data
Department of Computer and Information Sciences, Pakistan Institute of Engineering and Applied Sciences Islamabad, Pakistan 0.582 No. [paper]
Radboud UMC Nijmegen, The Netherlands 0.471 Yes, see paper


Task 2: Prediction of the proliferation score based on gene expression

The evaluation metric for this task is the Spearman's correlation coefficient between the predicted and ground truth proliferation scores.


Automatic methods

Team Score Use of additional data
Radboud UMC Nijmegen, The Netherlands 0.519 Yes, see paper
Radboud UMC Nijmegen, The Netherlands 0.6315 Yes, multi-task learning


Task 3: Mitosis detection

The evaluation metric for this task is the F1-score of the detection.


Automatic methods

Team Score Use of additional data
The Chinese University of Hong Kong 0.620 Yes; ICPR 2012 and 2014 datasets
University of Alberta 0.487 No
University of Warwick, United Kingdom 0.640 Yes, mining of additional mitoses from whole-slide images
Department of Computer and Information Sciences, Pakistan Institute of Engineering and Applied Sciences Islamabad, Pakistan 0.571 Yes; ICPR 2012 and 2014 datasets
IBM Center for Open-Source Data & AI Technologies (CODAIT), San Francisco, California, United States 0.601 Yes; ICPR 2012 and 2014 datasets
Radboud UMC Nijmegen, The Netherlands 0.480 Yes, see paper
School of Electronics Information and Communications, Huazhong University of Science and Technology, China 0.669 No. [paper]
Deep Learning Lab, Centre for Mathematical Sciences from Pakistan Institute of Engineering and Applied Sciences (PIEAS), Pakistan 0.4969 No.
School of Computer Science and Technology, Huazhong University of Science and Technolog (HUST), Wuzhen, China 0.6269 No.