MMLiSeg

Multi-modality Liver Segmentation Challenge

Motivation

The liver, being the largest solid organ in the human body, is crucial for metabolism and digestion. Accurate and precise segmentation of liver is essential for diagnosing cancer, planning treatment, and monitoring treatment response. However, current deep learning practice has poor generalizability facing domain shifts, that is, when deep learning models are evaluated on unseen datasets with different imaging modalities, the accuracy of segmentation can significantly decrease. This challenge, Multi-modality Liver Segmentation Challenge (MMLiSeg), aims to contribute to the effort of motivate participants to create generalizable automatic segmentation algorithms that can accurately segmentation across imaging modalities.

Task

MMLiSeg targets to generalizable segmentation methods for Liver segmentation in different modalities, including T1, T2 and DWI MRIs. There are three phases, i.e., training, validation and test phases. In the training phase, T1 and T2 MRI data are provided. In the validation phase, T1, T2 and DWI data are used for hyperparameters selection. While in the test phase, the submitted algorithms of participants are finally evluated on T1, T2 and DWI data.

Data

Data acquisition information

We include 180+ multi-modalities MRIs (.nii.gz) with manual segmentation of the liver region. All these clinical data have got institutional ethic approval and have been anonymized (please follow the data usage agreement).
The details of the dataset are summarized as below:

Modality Center Num. studies Num. slices
T1 1 62 3031
T2 1 62 1675
DWI 1 64 1540

Data split

The dataset has been divided into three main parts: training, validation, and test sets:

Metrics and Award

Metrics

The performance of scar and edema segmentation results will be evaluated by：

Award

The performance will be evaluated based on the of the submitted models. The best work will be selected with awards. Specifically, we will prepare one Champion Winner Award.

Registration

To access the dataset, please sign up to participate in the challenge and get access to the dataset.

Submission Guidance

Model Submission

After registration, we will assign the participant an account to login into our evaluation platform. Participants can directly upload your predictions on the validation data (in nii.gz format) via the website. Note that evaluation of validation data will be allowed up to 10 times for each task per team. For fair comparison, the test dataset will remain unseen. Participants need to submit their predicted segmentation results to care_acait@163.com.

Timeline

The schedule for this track is as follows. We have extended the DDL of test phase to October 31th. All deadlines are on 12:00 pm in Pacific Standard Time.

Training Data Release June 1, 2024
Validation Phase August 1, 2024 to September 15, 2024 (DDL)
Test Phase September 16, 2024 to October 31, 2024 (DDL)
Notification November 5
Workshop (Half-Day) November 8, 2024

Citations

Please cite these papers when you use the data for publications:

@inproceedings{gao2023reliable,
  title={A reliable and interpretable framework of multi-view learning for liver fibrosis staging},
  author={Gao, Zheyao and Liu, Yuanye and Wu, Fuping and Shi, Nannan and Shi, Yuxin and Zhuang, Xiahai},
  booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention},
  pages={178--188},
  year={2023},
}

@misc{liu2024merit,
  title={MERIT: Multi-view Evidential learning for Reliable and Interpretable liver fibrosis sTaging}, 
  author={Yuanye Liu and Zheyao Gao and Nannan Shi and Fuping Wu and Yuxin Shi and Qingchao Chen and Xiahai Zhuang},
  year={2024},
  archivePrefix={arXiv},
}

Organizer

Contact

If you have any questions regarding the MMLiSeg challenge, please feel free to contact: