Recovery of accurate T from historical 1.5 tesla proton density and T-weighted images: Application to 7-year T changes in multiple sclerosis brain.

Objective: To determine accurate quantitative transverse relaxation times (T) using retrospective clinical images and apply it to examine 7-year changes in multiple sclerosis (MS) brain.

Methods: A method for T mapping from retrospective proton density (PD) and T-weighted fast spin echo images was recently introduced, but requires measurement of flip angles. We examined whether 1.5T flip angle variation in brain can be predicted, thus enabling T analysis of historical PD and T-weighted images without a concurrent flip angle map. After method validation in healthy volunteers, retrospective longitudinal T analysis was performed in 14 MS subjects over seven years. Changes in patient T values were compared with brain atrophy, T lesion load and disability score in MS.

Results: Similar flip angle maps across volunteers enabled retrospective T from PD and T-weighted images even when different refocusing angles were used. Over seven years, significant T changes of 2-4% were observed when using T modelling and the 7-year effect size for globus pallidus T was 0.56, which was more significant than brain atrophy. No significant T results were found when using exponential fit, which cannot account for refocusing angle variation. Moreover, change is T in globus pallidus and internal capsule correlated with MS disability score over time when using T modelling.

Conclusions: Accurate quantitative T can be extracted from standard clinical 1.5T MRI exams that include PD and T-weighted imaging even when no flip angle map is available. This method was applied retrospectively to examine seven year changes in MS.