This repository contains work-in-progress demo code for the Self-reference region method (Self-RRM) for analysing DCE-MRI. The Self-RRM relies on the reference region model, such as CERRM, where the reference tissue parameters are estimated with RRIFT. What sets Self-RRM apart from previous work is that the reference tissue is automatically identified from the tissue of interest (i.e. tumour), whereas the previous approach requires a manually identified reference region (e.g. muscle).
This demo applies the Self-RRM to DCE-MRI data acquired from a patient with glioblastoma multiforme.
- b01_process.m
- Fits the tumour data using:
- Self-RRM
- Reference region model using muscle as the reference region (conventional approach)
- Extended Tofts model using the measured arterial input function
- Fits the tumour data using:
- b02_showResults.m
- Shows parameter maps for the three fits, along with the concordance correlation coefficients
- Glibolastoma multiforme data was obtained from the Cancer Genome Atlas Glioblastoma Multiforme (TCGA-GBM) collection on the Cancer Imaging Archive (TCIA)
- Code for NMF-based hierarchical clustering was written by Nicolas Gillis