Skip to content

TF6.2 — DCE/DSC Challenge

Task Force 6.2 compares quantification pipelines for DSC/DCE-MRI in clinical applications and establishes benchmarks for perfusion imaging.

The aim of this task force is to compare quantification pipelines for DSC/DCE-MRI in clinical applications. Through these challenges, the performance of DSC-/DCE-MRI perfusion analysis tools developed in-house by participating groups, or the available software packages, are tested and evaluated according to technical performance metrics (e.g. bias and precision on DROs, agreement with reference methods in-vivo, reproducibility on in-vivo data, processing time). The aim is to establish a set of benchmarks for perfusion imaging in different applications.

The first challenge by this task force focused on benchmarking DCE software. As the use of artificial intelligence grows, the second challenge focuses on deep learning techniques. The goal is to encourage researchers to put their quantitative methods to the test, to stimulate collaboration, and to chart the heterogeneity of DCE analysis software. In this challenge, participants use deep learning techniques to estimate perfusion parameters in DCE-MRI of the uterus. The task force provides repeated in vivo data to assess the algorithm's precision, and simulated DCE-data to test the accuracy.

First DCE challenge

This repository contains all the documentation and datasets provided for the first ISMRM-OSIPI DCE Challenge.

📘 Go to the first challenge

Second DCE challenge

Work-in-progress: topic will be deep learning for DCE analysis. Coming soon.

Leads

Ongoing projects

The clinical data for this challenge were previously used for studying endometrial hypoxia and kindly shared for this challenge. The dataset includes the MR images of the uterus of 12 healthy female volunteers with regular menstrual cycles. Each volunteer was scanned twice: during days 1–3 of menstruation and the early/mid-secretory phase of their cycle. The two relevant MRI sequences for this challenge are DCE-MRI and IR-TrueFISP.

The task force is currently working on the design of the morphological phantom. This numeric phantom will be generated by assuming a series of perfusion parameters, T1 values, and a population-based Arterial Input Function (AIF). A two-compartment exchange model and the spoiled gradient-echo steady-state signal model will be used to create the voxel-wise signal. The assumed T1 values for blood and tissue will be provided to convert the signal to concentration.

More details are available in the two-year roadmap.

Publications

  • ISMRM abstract 2021 — cds.ismrm.org/protected/21MProceedings/PDFfiles/1094.html
  • ISMRM abstract 2022 — cds.ismrm.org/protected/22MProceedings/PDFfiles/0047.html
  • Shalom ES, Kim H, van der Heijden RA, Ahmed Z, Patel R, Hormuth DA 2nd, DiCarlo JC, Yankeelov TE, Sisco NJ, Dortch RD, Stokes AM, Inglese M, Grech-Sollars M, Toschi N, Sahoo P, Singh A, Verma SK, Rathore DK, Kazerouni AS, Partridge SC, LoCastro E, Paudyal R, Wolansky IA, Shukla-Dave A, Schouten P, Gurney-Champion OJ, Jiřík R, Macíček O, Bartoš M, Vitouš J, Das AB, Kim SG, Bokacheva L, Mikheev A, Rusinek H, Berks M, Hubbard Cristinacce PL, Little RA, Cheung S, O'Connor JPB, Parker GJM, Moloney B, LaViolette PS, Bobholz S, Duenweg S, Virostko J, Laue HO, Sung K, Nabavizadeh A, Saligheh Rad H, Hu LS, Sourbron S, Bell LC, Fathi Kazerooni A. The ISMRM Open Science Initiative for Perfusion Imaging (OSIPI): Results from the OSIPI-Dynamic Contrast-Enhanced challenge. Magn Reson Med. 2023 Dec 19. doi: 10.1002/mrm.29909.