Ultrasound and photoacoustic methods for anatomic and functional imaging in image guided radiation therapy

Justin T Sick, Purdue University

Abstract

(MATERIAL and METHODS) First, we define the physical principals and optimal protocols that provide contrast when imaging with US and the transducer properties contributing to resolution limits. The US field of view (FOV) was characterized to determine the optimal settings with regard to imaging depth, focal region, with and without harmonic imaging, and artifact identification. This will allow us to determine the minimum errors expected when registering multimodal volumes (CT, US, CBCT). Next, we designed an in-house integrated US manipulator and platform to relate CT, 3D-US and linear accelerator coordinate systems. To validate our platform, an agar-based phantom with measured densities and speed-of-sound consistent with tissues surrounding the bladder was fabricated. This phantom was rotated relative to the CT and US coordinate systems and imaged with both modalities. These CT and 3D-US images were imported into the treatment planning system, where US-to-US and US-to-CT images were co-registered and the registration matrix used to re-align the phantom relative to the linear accelerator. The measured precision in the phantom setup, which is defined by the standard deviation of the transformation matrix components, was consistent with and exceeding acceptable clinical patient re-alignments (2 mm). Statistical errors from US-US registrations for different patient orientations ranged from 0.06-1.66 mm for x, y, and z translational components, and 0.00-1.05 degrees for rotational components. Statistical errors from US-CT registrations were 0.23-1.18 mm for the x, y and z translational components, and 0.08-2.52 degrees for the rotational components. The high precision in the multimodal registrations suggest the ability to use US for patient positioning when targeting abdominal structures. We are now testing this on a dog patient to obtain both inter and intra-fractional positional errors. The objective of this experiment is to confirm Hill’s equation describing the relationship between hemoglobin saturation (SaO2) and the partial pressure of dissolved oxygen (pO2). The relationship is modeled as a sigmoidal curve that is a function of two parameters – the Hill coefficient, n, and the net association constant of HbO2, K (or pO2 at 50% SaO2). The goal is to noninvasively measure SaO2 in breast tumors in mice using photoacoustic computed tomographic (PCT) imaging and compare those measurements to a gold standard for pO2 using the OxyLite probe. First, a calibration study was performed to measure the SaO2 (co-oximeter) and pO2 (Oxylite probe) in blood using Hill’s equation (P50=23.2 mmHg and n=2.26). Next, non-invasive localized measurements of SaO2 in MDA-MD-231 and MCF7 breast tumors using PCT spectroscopic methods were compared to pO 2 levels using Oxylite probe. The fitted results for MCF7 and MDA-MD-231 data resulted in a P50 of 17.2 mmHg and 20.7 mmHg and a n of 1.76 and 1.63, respectively. The lower value of the P50 is consistent with tumors being more acidic than healthy tissue. Current work applying photon fluence corrections and image artifact reduction is expected to improve the quality of the results. In summary, this study demonstrates that photoacoustic imaging can be used to monitor tumor oxygenation, and its potential use to investigate the effectiveness of radiation therapy and the ability to adapt therapeutic protocols.

Degree

Ph.D.

Advisors

Stantz, Purdue University.

Subject Area

Health sciences|Medical imaging|Oncology

COinS