Title: Automatic Structure Segmentation for Radiotherapy Planning Challenge 2020

Type Other Hongsheng Li, Ming Chen (2020): Automatic Structure Segmentation for Radiotherapy Planning Challenge 2020. Zenodo. Other. https://zenodo.org/record/3718885

Authors: Hongsheng Li (The Chinese University of Hong Kong) ; Ming Chen (Cancer Hospital of the University of Chinese Academy of Sciences) ;

Links

• Item record in Zenodo
• Digital object URL

Summary

This is the challenge design document for the "Automatic Structure Segmentation for Radiotherapy Planning Challenge 2020", accepted for MICCAI 2020.

We propose to hold the challenge of automatic structure segmentation for radiotherapy planning 2020 in conjunct with MICCAI 2020. We will provide data of two types of cancers, nasopharyngeal cancer and lung cancer. Four challenge tasks will be organized, including Organ-at-risk segmentation from head and neck CT scans, Organ-at-risk segmentation from chest CT scans, Gross Target Volume segmentation of nasopharynx cancer, Gross Target Volume segmentation of lung cancer. A total of 120 CT scans with more than 1520 organ or tumor annotations will be provided in the chanllenge.

Radiation therapy is one type of important cancer treatment for killing cancer cells with external beam radiation. Treatment planning is vital for the treatment, which sets up the radiation dose distribution for tumor and ordinary organs. The goal of planning is to ensure the cancer cells receiving enough radiation and to prevent normal cells in organs-at-risk (OAR) from being damaged too much. Organs-at-risk are usually the organs that are sensitive to radiation. For instance, optical nerves and chiasma in the head cannot receive too much radiation otherwise the patient risks losing his/her vision. Gross Target Volume (GTV) is the position and extent of gross tumor imaged by CT scanes, i.e. what can be seen. One important step in radiotherapy treatment planning is therefore to delineate the boundaries of tens of OARs and GTV in every slice of a patient’s CT scans, which is tedious and occupies much of oncologists’ time. Automatic OAR & GTV delineation would substantially reduce the treatment planning time and therefore reduces the overall cost for radiotherapy.

This is a re-holding of the successful StructSeg 2019 challenge with 20% more training data on GTV segmentation provided to the research community. In StructSeg 2019, 718 teams have registered and 34 teams submitted their final models for evaluation. The challenge talks and ceremony have attracted a large number of audience to learn about the top-ranking methods during MICCAI 2019.

By releasing the new training data, continuely running the online evaluation server, holding the challenge talks, the StructSeg 2020 is expected to continually attract much attention from the research community and advance the research on OAR and GTV segmentation significantly.

More information

• DOI: 10.5281/zenodo.3718885

Subjects

• MICCAI Challenges, Biomedical Challenges, MICCAI, Organ-at-risk Segmentation, Gross Target Volume segmentation, Nasopharynx cancer, Lung cancer

Dates

• Publication date: 2020
• Issued: March 20, 2020

Rights

• https://creativecommons.org/licenses/by-nd/4.0/legalcode Creative Commons Attribution No Derivatives 4.0 International
• info:eu-repo/semantics/openAccess Open Access

Format

electronic resource

Relateditems