Oxygen Imaging for Non-Invasive Metastasis Detection

Abstract

Sentinel lymph node (SLN) biopsy is an integral part of treatment planning for a variety of cancers as it evaluates whether a tumor has metastasized, an event that significantly reduces survival probability. However, this invasive procedure is associated with patient morbidity, and misses small metastatic deposits, resulting in the removal of additional nodes for tumors with high metastatic probability despite a negative SLN biopsy. To prevent this over-treatment and its associated morbidities for patients that were truly negative, we propose a tissue oxygen imaging method called Photoacoustic Lifetime Imaging (PALI) as an alternative or supplementary tool for SLN biopsy. As the hyper-metabolic state of cancer cells significantly depresses tissue oxygenation compared to normal tissue even for small metastatic deposits, we hypothesize that PALI can sensitively and specifically detect metastases. Before this hypothesis is tested, however, PALI's maximum imaging depth must be evaluated to determine the cancer types for which it is best suited. To evaluate imaging depth, we developed and simulated a phantom composed of tubing in a tissue-mimicking, optically scattering liquid. Our simulation and experimental results both show that PALI's maximum imaging depth is 16 mm. As most lymph nodes are deeper than 16 mm, ways to improve imaging depth, such as directly delivering light to the node using penetrating optical fibers, must be explored.

Keywords: head and neck cancer; imaging; oxygen imaging; photoacoustic; sentinel lymph node biopsy.

Figure A1

Two state model of methylene blue triplet state dynamics.

Figure 1

(a) Overview of PALI theory. The beige line shows the expected exponential decay of the triplet state. This decay is sampled by varying the delay between the pump and probe lasers and measuring the photoacoustic signal generated by the probe. (b) Implementation of PALI begins with the computer where a MATLAB script is used to program a delay into the synchronization box. This box then triggers the pump, the probe, and the Verasonics (VSX) system. Triggering VSX initiates data acquisition by the ultrasound (US) probe. The recorded signal is then transferred to the computer for further processing.

Figure 2

(a) Overview of PALI Simulation. MC: Monte Carlo simulation (b). Computational phantom to measure PALI imaging depth.

Figure 3

(a) Setup for Noise Equivalent Pressure (NEP) measurements. A black disc is illuminated and the photoacoustic signals (P(A) are first measured with an ultrasound (US) transducer and then replaced with a hydrophone (H). (b) Example signals from the transducer and the hydrophone. Signals were normalized according to the laser pulse energy.

Figure 4

(a) Overview of phantom setup. Fluid from O₂ controlled reservoir was pumped into a box where oxygen measurements were collected with PALI using an ultrasound (US) transducer and laser. These measurements were verified using a dissolved oxygen probe (DO). (b) Picture of PALI chamber. Phantom consisted of 3D printed box with holders for the laser and ultrasound transducer. Tubing containing 400 µM MB was placed at the focal point of the transducer. (c) Measurement of the effective attenuation coefficient of scattering fluid. Dots represent the measured transmission of 660 nm light through the medium. The dashed line represents the exponential fit whose rate is in units of $\frac{1}{cm\ast f}$ where $f$ is the fractional volume of the scattering fluid relative to the final fluid volume. Since the fluid was not diluted further (i.e., the fractional volume is 1) the effective attenuation is 2.34 $\frac{1}{cm}$.

Figure 5

Setup for flash photolysis experiment. Photodetector (PD) measures the change in transmission of the continuous wave probe laser following a pulse excitation by the pump laser. When the pump excites the methylene blue (M(B), the transmission of the probe laser decreases and transmission recovers as the triplet state density goes to 0.

Figure 6

Mean and standard deviation of simulated PALI decay rate measurements at varying depths for deoxygenated (a) and oxygenated (b) solutions. Black dashed lines represents the decay rate corresponding to 10 mm Hg error. Error from the true decay rate for low and high oxygenations is shown in (c,d).

Figure 7

Low (a) and high (b) oxygen decay rate measured by flash photolysis. The blue line is the photodetector (PD) signal, red is the double exponential fit, and yellow is the single exponential fit. (c) Decay rate measured by PALI for tubing in water. Stars are the data points and the lines are the fit. The red and blue lines correspond to high and low oxygen, respectively.

Figure 8

(a,b) show the measured decay rate at different depths for deoxygenated and oxygenated solutions, respectively. The red line represents the decay rate measured in water while the blue line is the decay rate measured in the scattering medium. (c,d) show the SNR and the error for the oxygenated (dark red) and deoxygenated (dark yellow) solutions, respectively.