Localization of spherical lesions in tumor-mimicking phantoms by 3D sparse array photoacoustic imaging
The experimental setup. (a) Schematic of the imaging system. (b) Phantom details. Lesions of different types (see panel c) were immersed in a tissue phantom made of 1% agarose gel mixed with 1% Intralipid™. The lesions were imaged at varying depths inside the tissue phantom and were, in general, undetectable visually. A reference line-pointer was used to indicate a fixed plane in space to which the bottom of the phantom was aligned prior to imaging. Coordinate system is indicated on the right. (c) Lesion types. Top left—Homogeneous lesions of sizes 1.5, 3, 6, and 9 mm (left to right) were prepared at different blood concentrations (shown 12.5%). Top right—Internal vascular lesion. Bottom left—“External” vascular lesion with vessel diameter of 1.2 mm. Bottom right—External vascular lesion with vessel diameter of 0.5 mm. Details of lesion preparation are given in the text.
Photoacoustic imaging of homogeneous lesions, at blood concentration of 50%. (a) Pressure profiles measured from the 6 mm sphere on three representative transducers, showing sharp peaks (at ) originating from phantom surface followed by broader peaks (at ) originating from the lesion itself. (b) PA images and cross-sectional photographs of three tumor lesions sizes 3, 6, and 9 mm, embedded in tissue phantoms at depths of 15–20 mm from the surface. The PA image on the left is of a point source referenced to a line-pointer that indicated the bottom surface of the phantoms. Imaging volume is with voxels.
PA depth-localization of spherical blood-filled lesions embedded in homogeneous phantoms compared to depth (to center of sphere) measured postimaging during dissection of the phantoms. The panels correspond to increasing blood concentration in the phantom (top to bottom) and increasing lesion size (left to right; see legend). At the highest blood concentration (75%), only lesions at the smallest size were fabricated due to preparation difficulties at the larger sizes. Symbols indicate the expected -location computed from all the voxels as described in the methods. Error bars represent ±1 standard deviation. The line of identity is plotted on each graph to indicate ideal correspondence between PA and measured depth estimates. Note: The data point at a measured depth of 8 mm for the 12.5% blood and 1.5 mm lesion was considered an outlier and was excluded during estimation of the threshold depth. With this point included the threshold depth was estimated as 7–8 mm.
PA imaging of vascular lesions. (a) Internal lesion. (b) 0.5 mm external lesion. (c) 1.2 mm external lesion. Each lesion is represented by four images. Left—Bottom view photograph of the lesion during phantom preparation, from the same perspective as the laser beam; second from left—Bottom view PA image; third from left—Side view photograph of the lesion cross-section; right—Side view PA image.
PA depth-localization of the vascular lesions compared to depth measured postimaging after dissection of each phantom. (a) Internal vessel lesion with 0.5 mm diameter blood-filled extrusion [see Fig. 4(a)]. (b) External lesion with 0.5 mm diameter blood-filled extrusion [see Fig. 4(b)]. (c) External lesion with 1.2 mm diameter blood-filled extrusion [see Fig. 4(c)]. Error bars represent one standard deviation.
Summary of the threshold depth (in mm) at which PA images of blood-filled lesions were able to accurately predict actual depth.
Comparison between homogeneous and heterogeneous lesions in terms of the effective hemoglobin concentration and the corresponding detection depth.
Article metrics loading...
Full text loading...