Related Experiment Video
Updated: Oct 3, 2025
05:30
Measurement of Tumor T2* Relaxation Times after Iron Oxide Nanoparticle Administration
Published on: May 19, 2023
1.5K
Pseudo-T2 mapping for normalization of T2-weighted prostate MRI
Kaia Ingerdatter Sørland1, Mohammed R S Sunoqrot2, Elise Sandsmark3
1Department of Circulation and Medical Imaging, Norwegian University of Science and Technology, Olav Kyrres Gate 9, 7030, Trondheim, Norway. kaia.i.sorland@ntnu.no.
Magma (New York, N.Y.)
|February 12, 2022
Summary
AutoRef, an automated method for T2-weighted MRI normalization, successfully reduced data heterogeneity across multiple centers. The femoral head/muscle reference pair (AutoRefFH) provided the most accurate pseudo-T2 values for quantitative prostate analysis.
Area of Science:
- Radiology
- Medical Imaging
- Quantitative MRI
Background:
- Quantitative analysis of multicenter T2-weighted (T2W) MRI data requires signal intensity normalization to address heterogeneity.
- AutoRef is an automated dual-reference tissue normalization technique that generates a pseudo-T2 map for transversal prostate T2W MRI.
Purpose of the Study:
- To evaluate the accuracy of AutoRef-generated pseudo-T2 values.
- To assess the multicenter standardization performance of AutoRef using three reference tissue pairs: fat/muscle (AutoRefF), femoral head/muscle (AutoRefFH), and pelvic bone/muscle (AutoRefPB).
Main Methods:
- Compared multi-echo spin echo (MESE) measured T2 values with AutoRef pseudo-T2 values in prostate regions (whole prostate, peripheral zone, transition zone/central zone/anterior fibromuscular stroma) of seven volunteers.
- Assessed AutoRef normalization on T2W images from 1186 prostate cancer patients across multiple centers.
- Measured performance by analyzing inter-patient histogram intersections of voxel intensities before and after normalization in 80 cases.
Main Results:
- AutoRefFH pseudo-T2 values closely approximated MESE T2 values in volunteers, with no significant differences observed across prostate regions.
- All AutoRef versions improved inter-patient histogram intersections in the multicenter dataset, increasing median intersections from 0.505 (original data) to approximately 0.73-0.74.
Conclusions:
- All tested AutoRef normalization methods effectively reduced data variation in multicenter T2W MRI datasets.
- The AutoRefFH method, utilizing the femoral head/muscle reference pair, yielded pseudo-T2 values closest to experimentally measured T2s, indicating its suitability for accurate quantitative analysis.
