Request edit access
OSIPI open source ASL pipelines
Dear pipeline authors,
Jan Petr and Henk Mutsaerts will provide an overview of ASL pipelines at the OSIPI open source perfusion imaging symposium coming Monday, at the ISMRM. We would like to summarize the features open source ASL pipelines should have and recapitulate where we stand with the different ASL pipelines that are currently freely available.
Could you please fill in the short questionnaire below about the features your pipeline is offering (I have figured out, that having the authors answer these question will be more fair, then trying to judge this subjectively).
For each of the topics below, please select which of the options (a/b/c) best characterizes your pipeline. The options are 'cumulative' - by answering 'c' it means that the pipeline can also do 'a' and 'b'. There are no free fields at the questions, to make it easier for us to obtain stats. Please provide your comments in the free fields at the bottom.
Many thanks in advance!
1) Name of your pipeline
2) Is your pipeline publicly available and free for academical use?
Free use is restricted
3) What data structure is expected as input?
a) Fixed directory and file structure, NIFTIs.
b) Flexible directory structure, NIFTIs.
c) Any structure, import directly from DICOMs.
4) To what extent can your pipeline be configured (i.e. without changing the code)?
a) Most parameters (except for PLD, TE etc.) are fixed
b) Basic parameters such as lambda, T1blood, GM-threshold can be configured
c) All parameters can be configure included smoothing kernels, iteration numbers etc.
5) Does your pipeline process structural images as well?
a) No, processed structural images should be provided separately
b) Yes, high resolution T1w images are segmented, co-registered etc.
c) Yes, including extras such as longitudinal registration, FLAIR segmentation, or cost function (lesion)masking
6) What is the output of your pipeline?
a) CBF maps + whole brain descriptives (e.g. mean CBF for whole brain, GM, WM)
b) Also ROI descriptives for smaller regions (e.g. mean CBF for cortical regions, flow territories)
c) Also ROI descriptives for any externally provided atlas, lesions or custom ROIs from other results
7) Are there means for QC, such as PDF reports, JPG QC images to inspect?
a) No, the NIFTIs (or other derivatives) need to be visually checked
b) Yes, visual QC reports and images are provided for all the intermediate steps (registration, motion correction, segmentation etc.)
c) Yes, and also automatic QC with reports
8) What kind of labeling strategies can your pipeline quantify?
a) Single PLD of the common labeling schemes - PASL, pCASL, CASL.
b) Also multi-PLD sequences
c) Also more "advanced" sequences, such as time-encoded, QUASAR, velocity-selective
9) Does your pipeline support data from multiple vendors?
a) Data from a single MRI vendor only
b) Data from all main MRI vendors, but single vendor for a single study
c) Data from all main MRI vendors mixed in a single study.
10) Can your pipeline handle time-series?
a) No, it will process only the mean subtraction
b) Yes, it performs motion correction
c) Yes, and it also performs outlier detection/denoising
11) Does your pipeline perform partial volume correction?
b) No, but it provides GM-volume as covariate in statistics/or it normalizes by GM volume (either voxel-wise or ROI-wise)
c) Yes, partial volume correction by linear regression with GM, WM and CSF components (or similar)
12) Does you pipeline run on multiple OSes?
a) No, only a single OS
b) Yes, it supports Linux/Mac/Windows
c) Yes, and we have established the reproducibility between the three OSes
13) Is the source code available for others?
a) No, (compiled) executables only
b) Yes, the complete source code is available upon request
c) Yes, the complete source code is distributed on public GitHub (or similar)
14) Have you tested the pipeline on large clinical datasets?
a) No, < 1000 subjects
b) Yes, >1000 subjects, by third party users (not by the pipeline authors)
c) Yes, > 1000 subjects & also tested directly by the pipeline authors
Can you summarize in two sentences what are the main strong points of your pipeline, i.e. what makes it interesting compared to the other pipelines available? (RECOMMENDED - goes to the presentation)
What features should a good open source ASL pipeline have? Which are, of the above are important, which aren't, which would you add? (OPTIONAL)
Any other comments? (optional)
Which manuscript reference to your pipeline do you suggest we use?
Which website reference do you suggest we use?
Send me a copy of my responses.
Page 1 of 1
Never submit passwords through Google Forms.
This content is neither created nor endorsed by Google.
Terms of Service