Pixel-by-pixel ΔR ₂* maps are obtained from the formula: ΔR ₂ * = (1/TE)log(S _o ^bef/S _o ^aft), where TE is the echo time and S _o the signal amplitude pre-USPIO (S _o ^bef) and post-USPIO (S _o ^aft), in the gradient-echo images (see Note 6). The same algorithm is used to generate ΔR ₂ maps from the spin-echo images. The algorithm for calculating ΔR ₂* (and ΔR ₂) from the formula reported above is implemented as a script written in MATLAB (see Note 7).

After generating ΔR ₂* and ΔR ₂ maps, ΔR ₂* and ΔR ₂ values are measured in manually segmented ROIs within the region of interest. The algorithm for generating ROIs is also implemented as a script written in MATLAB.

For each ROI, mean ΔR ₂* and ΔR ₂ are calculated by averaging the values of all pixels within the ROI.

The Mel57-VEGF-A₁₆₅ metastases show very high values of ΔR ₂* and ΔR ₂ (ΔR ₂* = 280 ± 90 s⁻¹ and ΔR ₂ = 40 ± 10 s⁻¹, mean ± standard deviation, assessed in four mice and for each mouse five lesions with the highest ΔR ₂* values are considered).

The U87 tumors are characterized by relatively low ΔR ₂* and ΔR ₂ values in the core (ΔR ₂* = 17 ± 6 s⁻¹ and ΔR ₂ = 5 ± 3 s⁻¹, assessed in four mice) and moderately high ΔR ₂* and ΔR ₂ values in the peritumoral region (ΔR ₂* = 54 ± 15 s⁻¹ and ΔR ₂ = 13 ± 5 s⁻¹).

It is also of interest, for reference, to measure the values of ΔR ₂* and ΔR ₂ in healthy brain regions (cortex: ΔR ₂* = 21 ± 9 s⁻¹, ΔR ₂ = 5 ± 2 s⁻¹; striatum: ΔR ₂* = 36 ± 10 s⁻¹, ΔR ₂ = 9 ± 3 s⁻¹). An example of the macrovascular blood volume map (i.e., ΔR ₂* map) of Mel57-VEGF-A₁₆₅ and U87 tumors is shown in Fig. 3a and c, respectively. As explained in point 1 of this section, the algorithm for calculating ΔR ₂* (and ΔR ₂) is implemented as a script written in MATLAB.

In Mel57-VEGF-A₁₆₅ tumors, a high threshold for ΔR ₂* is used in the tumor areas (region inside the box, Fig. 3a; note the ΔR ₂* scale up to 300 s⁻¹). Outside the tumor region, a low cutoff threshold (ΔR ₂* scale up to 100 s⁻¹) is used to maximize the contrast between striatum (ROI indicated on the corresponding gradient-echo image by the horizontal arrow, Fig. 3b) and cortex (ROI indicated by the vertical arrow, Fig. 3b).

In Fig. 3d, the gradient-echo image corresponding to the U87 tumor macrovascular blood volume map is shown.

The T ₂ relaxation time is the time constant that governs the loss of transverse magnetization (i.e., of signal) in spin-echo imaging. This relaxation process is due to the dephasing of the individual magnetic dipoles because of the existence of non-stationary random local magnetic fields. On the other hand, the loss of transverse magnetization (i.e., of signal) in gradient-echo imaging is characterized by a relaxation time T ₂*. Two factors contribute to the T ₂* decay: the non-stationary random local magnetic fields (which generate the T ₂ decay) and the static (i.e., constant in time) inhomogeneities in the magnetic field. Static distortions in the magnetic field are particularly dominant near interfaces between regions with different magnetic susceptibilities (air–tissue interface, for example). It should be noted that the dephasing of spins resulting from static field inhomogeneities is a reversible process, which is accomplished by applying a 180° pulse (which is present in the spin-echo imaging). The dephasing due to fluctuating magnetic fields, however, is not reversible.