Correcting time-intensity curves in dynamic contrast-enhanced breast MRI for inhomogeneous excitation fields at 7T

Abstract

Purpose: Inhomogeneous excitation at ultrahigh field strengths (7T and above) compromises the reliability of quantified dynamic contrast-enhanced breast MRI. This can hamper the introduction of ultrahigh field MRI into the clinic. Compensation for this non-uniformity effect can consist of both hardware improvements and post-acquisition corrections. This paper investigated the correctable radiofrequency transmit ( $B_1^{+}$ ) range post-acquisition in both simulations and patient data for 7T MRI.

Methods: Simulations were conducted to determine the minimum $B_1^{+}$ level at which corrections were still beneficial because of noise amplification. Two correction strategies leading to differences in noise amplification were tested. The effect of the corrections on a 7T patient data set (N = 38) with a wide range of $B_1^{+}$ levels was investigated in terms of time-intensity curve types as well as washin, washout and peak enhancement values.

Results: In simulations assuming a common amount of T₁ saturation, the lowest $B_1^{+}$ level at which the SNR of the corrected images was at least that of the original precontrast image was 43% of the nominal angle. After correction, time-intensity curve types changed in 24% of included patients, and the distribution of curve types corresponded better to the distribution found in literature. Additionally, the overlap between the distributions of washin, washout, and peak enhancement values for grade 1 and grade 2 tumors was slightly reduced.

Conclusion: Although the correctable range varies with the amount of T₁ saturation, post-acquisition correction for inhomogeneous excitation was feasible down to $B_1^{+}$ levels of 43% of the nominal angle in vivo.

Keywords: $B_1^{+}$ mapping; 7T; DCE-MRI; RF field inhomogeneity; breast; flip-angle correction.

Figure 1

Influence of ${\mathrm{B}}_1^{+}$ on time‐intensity curves in DCE‐MRI. Low ${\mathrm{B}}_1^{+}$ induces a type II plateau curve, even in tumors that should show a type III washout curve. Each curve has been normalized to its own maximum for clarity

Figure 2

(A) Signal intensity (S) versus flip angle. In case of a nominal imaging angle above the Ernst angle (θ_Ernst), a lower flip angle can be found that produces the same signal intensity, the equivalent angle. The range between the nominal angle and the equivalent angle determines the ${\mathrm{B}}_1^{+}$ buffer in terms of signal loss. (B) Derivative of the signal intensity with respect to T₁ versus flip angle. The sensitivity to T₁ changes decreases rapidly for low flip angles. This limits the correctable ${\mathrm{B}}_1^{+}$ range further

Figure 3

True versus measured image intensity at different ${\mathrm{B}}_1^{+}$ levels, assuming a flip angle of 15° and a TR of 5.8 ms as in our patient study (see Section 2.3). This direct mapping is the basis of the proposed correction mechanism. The inset shows the nonlinear behavior of this mechanism for measured image intensity levels higher than the theoretical maximum: the blue curve shows the mapping from measured signals to corrected signals for a ${\mathrm{B}}_1^{+}$ level of 40%, and the red dot indicates the theoretical maximum signal

Figure 4

Demonstration of noise amplification in simulations at a ${\mathrm{B}}_1^{+}$ level of 50% (A) or 30% (B) of the nominal angle, assuming a flip angle of 15° and a TR of 5.8 ms as in our patient study (see Section 2.3). The time‐intensity curves shown are the mean curve of all noise instances, the shaded area indicates the standard deviation. Noise amplification increases for lower ${\mathrm{B}}_1^{+}$ levels for both the Haacke‐based and proposed methods, but the proposed method amplifies the noise less strongly

Figure 5

Occurrence and influence of the ρ estimation strategy, indicated by the colored dots. (A) Scatter plot of the minimum ${\mathrm{B}}_1^{+}$ level in the tumor (x‐axis) versus the median ${\mathrm{B}}_1^{+}$ level in the measured map (y‐axis). The median ${\mathrm{B}}_1^{+}$ level on the y‐axis is used as an indicator for the reliability of the measured map, because the linear range of the mapping technique used is limited to 50–150% ona. The fallback strategy is only used for low ${\mathrm{B}}_1^{+}$ levels in the tumor (<50% ona). The reliability of the maps did not have an influence. (B) Scatter plot of the minimum ${\mathrm{B}}_1^{+}$ level in the tumor (x‐axis) versus the estimated ρ value (y‐axis). The values estimated by the fallback strategy are in the same range as those estimated by the default strategy. (C) Visualization indicating the tumors that were corrected using the default strategy in green and the fallback strategy in red. All tumor positions are shown relative to the position of the coils, indicated in blue. % ona, percentage of the nominal flip angle

Figure 6

Typical example of an original versus a corrected time‐intensity curve. The curves shown are the mean of the top 10% most‐enhancing tumor voxels. Each curve has been normalized to its own maximum for clarity

Figure 7

Washin, washout, and peak enhancement distributions per grade, before and after DCE correction. Notice how in all distributions the number of low values are reduced by the correction, as expected