US imaging

In our study, we used a high-resolution US imaging system designed for the examination of small animals (Visual Sonics Vevo 2100 with a transducer manufactured by VisualSonics, Toronto, Canada). Transducer was placed as close as possible to animal leg or holder in way to not distort the animal during imaging. Casts and holders with fiducials were imaged using B-mode (Figure 1). The imaging transducer was used at a frequency of 21 MHz, which determines an XY field of view (FOV) up to 23 × 23mm. 3D images, each containing about 40–100 scans per imaged object, were obtained with a 0.5–0.171 mm step using a steady arm-held transducer. The step size determines the image resolution in the Z direction. For animal imaging, an MS-250 transducer at a frequency of 40 MHz for B-mode and 32 MHz (FOV: 10×10 mm) for power Doppler imaging was used. (Figure 2).

Results of EPR – CT phantom imaging and image co-registration. EPR – CT Column A: PVS open cast (blue) with plastic sleeves (white) containing glass fiducials. A dark green solution of Finland spin probe is seen. Column B: Rigid cradle/holder (blue) 3D printed from polyactic acid filament with fiducials. 1. Pictures of phantoms used. 2. A single perpendicular slice from 3D ultrasound B-mode image with segmented interior diameter of sleeves (red). 3. A single perpendicular slice from 3D EPR images of fiducials(red). 4. Ultrasound – EPR image registration using ArbuzGUI software (green – sleeves from ultrasound, magenta – EPR fiducials). For details of imaging parameters, see Methods.

Results of in vivo imaging. Two dimensional orthogonal slice of mouse leg bearing tumor from ultrasound and EPR images. Tumor segmentation (red) was performed in the ultrasound image and after registration the tumor mask was transferred into EPR image coordinate system. Ultrasound power doppler shows segmented vessels with blood flow (red). For ultrasound images pixel size for axial plane was 0.03 × 0.03 mm while for coronal and sagittal 0.03 × 0.3 mm. Pixel size was determined during image reconstruction as a field of view (FOV) divided by number of pixels in the image for each dimension. For in-vivo imaging we used FOV equal to 4×4×4cm and number of pixels 128×128×128, which correspond to 0.3125 mm symmetrical pixel size.