Extended Data Fig. 5: Optimization of DCE-MRI for precise kio measurements in vivo.
a: The implementation of MGE sequence in Water-exchange DCE-MRI eliminates potential T2* artifacts caused by the contrast agent at 7 T by fitting the MGE data (Supplementary Methods Section 3, equation 5) to obtain the purely T1-weighted signal S (TE = 0 ms). b: The original data with the shortest TE (2.8 ms) still shows T2*-induced signal attenuation. c: Monto Carlo simulations demonstrate that our optimized protocol (dual-bolus injection and optimized sequence settings) shows one-fold smaller standard deviation of kio estimation than the conventional scanning protocol (single-bolus injection, TR = 10 ms, FA = 10°. Std denotes the standard deviation). Box plot specifications: box bounds mean 25th and 75th percentile, center = 50th percentile, minima/maxima = center ± 1.5 × (75th percentile – 25th percentile), no whiskers shown. d-f: An example of a kio maone-foldone-foldp overlaid on T2-weighted image in the tumor region (d) and the model fittings of the DCE-MRI data with SS model for pixels located in the core (e, low kio = 1.6 s−1) and ring (f, high kio = 10.0 s−1) of the tumor. Here the raw data and the fitting results are shown as dots and continuous curves, respectively. Scale bar, 2 mm.
