Rapid 3D mapping of T 1 relaxation times is valuable in diverse clinical applications. Recently, the variable flip angle (VFA) spoiled gradient recalled echo approach was shown to be a practical alternative to conventional methods, providing better precision and speed. However, the method is known to be sensitive to transmit field (B 1 ؉ ) inhomogeneity and can result in significant systematic errors in T 1 estimates, especially at high field strengths. The main challenge is to improve the accuracy of the VFA approach without sacrificing speed. In this article, the VFA method was optimized for both accuracy and precision by considering the influence of imperfect transmit fields, noise bias, and selection of flip angles. An analytic solution was developed for systematic B 1 ؉ -induced T 1 errors and allows simple correction of T 1 measurements acquired with any imaging parameters. A noise threshold was also identified and provided a guideline for avoiding T 1 biases. Finally, it was shown that three flip angles were the most efficient for maintaining accuracy and high precision over large ranges of Rapid and accurate measurement of the longitudinal, or T 1 , relaxation time has long challenged MRI scientists but remains an important goal because of its clinical relevance across a diverse range of applications. Many of these applications, however, including dynamic contrast-enhanced studies of cancer, perfusion studies of muscle, diagnosis of neurologic disorders such as multiple sclerosis and epilepsy, and enhanced tissue discrimination for image-guided procedures, require low-noise, high-resolution mapping over a large volume. These requirements cannot be achieved in a clinically acceptable time frame (Ͻ30 min) using conventional methods of inversion-or saturation-recovery (1-3). An alternative method (4) involves determining T 1 from a set of two or more spoiled gradient recalled-echo (SPGR) images acquired with varied flip angles. Advantages of this approach include low power deposition compared to spin echo techniques and low spatial distortion compared to rapid echo planar imaging. Most importantly, the accuracy has been shown to be similar to that achieved with conventional and accelerated techniques but with a significant reduction in imaging time (5,6).A dominant source of error in the variable flip angle (VFA) approach is inaccurate knowledge of the flip angles due to transmit field B 1 ϩ inhomogeneity. Although the impact on T 1 accuracy is known (7,8), correction of measured T 1 has been reported in only a few studies (9,10), mainly due to the complexity and long scan time requirements of mapping the B 1 ϩ field. A second source of systematic error is noise-induced bias. The effect on T 1 measurements is generally subtle and may be the reason it has not, to our knowledge, been previously reported. However, these errors can be appreciable below a certain signal-tonoise (SNR) threshold, particularly with multiple angles, and one must identify and image at a suitable SNR level to ensure accurate measureme...