Longitudinal Photoacoustic Imaging of the Pharmacodynamic Effect of Vascular Targeted Therapy on Tumors

Abstract

Purpose: Photoacoustic imaging (PAI) is a novel noninvasive and nonionizing imaging technique that allows longitudinal imaging of tumor vasculature in vivo and monitoring of response to therapy, especially for vascular targeted chemotherapy agents. In this study, we used a novel high-resolution all-optical PAI scanner to observe the pharmacodynamic response to the vascular-disrupting agent OXi4503.

Experimental design: Two models of colorectal carcinoma (SW1222 and LS174T) that possess differing pathophysiologic vascularization were established as subcutaneous tumors in mice. Monitoring of response was performed over a 16-day "regrowth" period following treatment at 40 mg/kg, and at day 2 for a "dose response" study at 40 mg/kg, 10 mg/kg, 1 mg/kg, and sham dose.

Results: Qualitative and quantitative changes in PA signal are observed, with an initial decrease followed by a plateau and subsequent return of signal indicating regrowth. Both tumor types exhibited a decrease in signal; however, the more vascularized SW1222 tumors show greater response to treatment. Decreasing the dose of OXi4503 led to a decrease in PA signal intensity of 60%, 52%, and 20% in SW1222 tumors and 30%, 26%, and 4% for LS174T tumors.

Conclusions: We have shown for the first time that PAI can observe the pharmacodynamic response of tumor vasculature to drug treatment both longitudinally and at different dose levels. Assessment of differing response to treatment based on vascular pathophysiologic differences among patients has the potential to provide personalized drug therapy; we have demonstrated that PAI, which is clinically translatable, could be a powerful tool for this purpose.

Trial registration: ClinicalTrials.gov NCT02576301.

Conflict of interest statement

P. Beard has a financial interest in DeepColor SAS. However, DeepColor SAS had no involvement in this work. The other authors declare no potential conflicts of interest

©2019 American Association for Cancer Research.

Figures

Figure 1. Fabry-Perot-based photoacoustic scanner and example image of a subcutaneous tumor xenograft.

(a) scanner architecture. Excitation laser pulses are transmitted through the FP sensor head and absorbed in the tissue generating photoacoustic signals which are then detected by the Fabry-Perot polymer film ultrasound sensor. Inset: an expanded view of the sensor which comprises a polymer spacer sandwiched between a pair of dichroic mirrors that are transparent to the excitation laser wavelength but highly reflective to the sensor interrogation beam wavelength. The sensor operates by raster scanning a focused continuous wave interrogation laser beam across it and measuring the change in the power of the reflected beam produced by acoustically-induced changes in the polymer spacer thickness. (b) Photoacoustic image of SW1222 tumor xenograft acquired using the scanner and displayed (clockwise) as x-y, x-z and y-z maximum intensity projections (MIPs). An animated volume rendered representation of this data can be viewed online (Video 1. The yellow arrows on the x-y MIP indicate the tumor. The x-z and y-z MIPs show that the entire tumor can be visualised with high resolution to a depth of approximately 8mm.

Figure 2

Photoacoustic images displayed as maximum intensity projections (MIPs) showing the longitudinal response of a SW1222 tumor (mouse m4sw) to 40mg/kg IV dose of OXi4503 over 16 days. The horizontal x-y maximum intensity projections (MIP) are for z = 1 - 6 mm. The tumour region in the pre-treatment image is indicated by yellow arrows. After treatment a region in the tumor core characterised by a lack of PAI contrast can be seen. This void diminishes over the 16 time course, with PA signal returning to areas of the tumor core.

Figure 3

Photoacoustic images displayed as maximum intensity projections (MIPs) showing the longitudinal response of a second subcutaneous SW1222 tumor (mouse m3sw) to 40mg/kg IV dose of OXi4503 over 16 days. The horizontal x-y MIPs are for z = 1 - 6 mm. The pre-treatment vertical x-z MIP shows the tumor (indicated with yellow arrows) growth into the body of the mouse, rather than protruding outwards as seen in the previous example of figure 2. After treatment, the destruction and regrowth of the fine vascular signal in the tumor core can therefore be observed clearly in not only the horizontal MIPs, but also in the vertical MIPs.

Figure 4

Photoacoustic images displayed as maximum intensity projections (MIPs) showing the longitudinal response of an LS174T tumor (mouse m4LS) to 40mg/kg IV dose of OXi4503 over 16 days. The horizontal x-y MIPs are for z = 1-6 mm. The vessel network in the tumor region, indicated with yellow arrows in the pre-treatment images, is seen to consist of relatively large sparsely distributed vessels, when compared to the spatially averaged vascular PA signal seen in the more highly vascularised homogenous SW1222 tumors. After treatment, a region in the tumor core which is characterised by a lack of contrast can be seen, due to the disruption of vessels.

Figure 5

Photoacoustic images displayed as maximum intensity projections (MIPs) showing the longitudinal response of a second LS174T tumor (mouse m2LS) to 40mg/kg IV dose of OXi4503 over 16 days. The horizontal x-y maximum intensity projections (MIP) are for z = 1 - 6 mm. After treatment, the loss of signal contrast in the tumor region (indicated with yellow arrows in the pre-treatment images) is still significant even at Day 16

Figure 6

Photoacoustic (PA) signal intensity and tumor volume of the two tumor types treated with 40mg/kg of OXi4503. Photoacoustic signal at different time points for (a) SW1222 and (b) LS174T tumors. Tumor volume measured by callipers at corresponding time points for (c) SW1222 and (d) LS174T tumors. For both tumor types, the PA signal intensity reduces after treatment before increasing again almost to pre-baseline levels. This is in agreement with the post-treatment destruction of the tumor core and subsequent repopulation of the vasculature, as shown in the PA images Figure 2–5. The tumor volumes in the SW1222 predominantly show an arrest in the tumor growth after treatment, up to day 16. In the LS174 tumors, an initial reduction in volume after treatment is followed by an increase at the later time points. SW1222: the PA signal data for mice m4SW and m3SW was obtained from the images shown in figures 2 and 3 respectively. LS174T: the PA signal data for mice m4LS and m2LS was obtained from the images shown in figures 4 and 5 respectively

Figure 7

(a) Sample of photoacoustic images (x-y MIPs) showing the response of two types of human colorectal tumor (SW1222, LS174T) to different dose of OXi4503. Yellow arrows in baseline images indicate the location of the tumor. The images are shown before IV treatment and 48 hours after treatment with saline, 1mg/kg, 10mg/kg or 40mg/kg. OXi4503 can be seen to produce a vascular disrupting response in the SW1222 down to 1mg/kg. The effect of OXi4503 on the LS174T tumors is reduced at 40mg/kg and 10mg/kg and negligible at 1mg/kg. (b) Percentage change in PA signal intensity of the two tumor types after treatment for groups of n≥4 mice, except SW1222 control and LS174T 1mg/kg where n = 3. The treatment is seen to produce the greatest dose dependent response in the SW1222, compared to the LS174T. This is consistent with the extent of vascularisation being greater in the SW1222 tumors.