In vivo assessment of prostate cancer response using quantitative ultrasound characterization of ultrasonic scattering properties

Abstract

Background: The study here investigated quantitative ultrasound (QUS) parameters to assess tumour response to ultrasound-stimulated microbubbles (USMB) and hyperthermia (HT) treatment in vivo. Mice bearing prostate cancer xenografts were exposed to various treatment conditions including 1% (v/v) Definity microbubbles stimulated at ultrasound pressures 246 kPa and 570 kPa and HT duration of 0, 10, 40, and 50 min. Ultrasound radiofrequency (RF) data were collected using an ultrasound transducer with a central frequency of 25 MHz. QUS parameters based on form factor models were used as potential biomarkers of cell death in prostate cancer xenografts.

Results: The average acoustic concentration (AAC) parameter from spherical gaussian and the fluid-filled spherical models were the most efficient imaging biomarker of cell death. Statistical significant increases of AAC were found in the combined treatment groups: 246 kPa + 40 min, 246 kPa + 50 min, and 570 kPa + 50 min, in comparison with control tumours (0 kPa + 0 min). Changes in AAC correlates strongly (r² = 0.62) with cell death fraction quantified from the histopathological analysis.

Conclusion: Scattering property estimates from spherical gaussian and fluid-filled spherical models are useful imaging biomarkers for assessing tumour response to treatment. Our observation of changes in AAC from high ultrasound frequencies was consistent with previous findings where parameters related to the backscatter intensity (AAC) increased with cell death.

Keywords: Cell death; Hyperthermia; Prostate cancer; Quantitative ultrasound spectroscopy.

Fig. 1

Effects of USMB and HT on tumour cell death. High magnification light microscope images of A TUNEL and B H&E stained histological sections of PC3 xenografts. A Representative images of PC3 xenograft treated with 1% microbubble stimulated at varying US pulses 246 kPa & 570 kPa and varied HT (10, 40, and 50 min). An increase in HT duration combined with USMB demonstrated higher cell death (brown color). Scale bar: 50 μm. B H&E-stained PC3 xenografts captured with a high magnification light microscope revealed a significantly higher number of distorted nuclei in the combined USMB and HT treated group compared to the untreated group or USMB alone. Combined treated cells exhibited morphological characteristics of apoptosis as well as necrosis. Scale bar: 50 μm. The image represents a representative experiment (from 3 to 4 independent experiments)

Fig. 2

Parametric maps of the AAC and changes in quantitative ultrasound parameters with different treatments. A Representative post-treatment (24-h) parametric images of AAC obtained using the SGM, overlaid on B-mode images, B Bar plots of averaged changes in AAC from SGM, C bar plots of averaged changes in AAC from FFSM, D Bar plots of averaged changes of ASD from SGM, and E Bar plots of averaged changes in ASD from FFSM from different treatment groups. A We observed increases in AAC from the combination of treatment groups: 246 kPa + 40 min, 246 kPa + 50 min, 570 kPa + 40 min, and 570 kPa + 50 min, relative to that of the control group (0 kPa + 0 min). These indicate an increase in acoustic back-scattering that reflects nuclear structure alteration as the tumour cells undergo cell death, as a result of a combination of cancer treatment: USMB-induced vascular disruption and HT. B Bar plots of changes in AAC from SGM from all treatment groups. C Bar plots of changes in AAC from FFSM from all treatment groups. The height of the bar represents the mean-value from each group. Error bars represent the standard error of the mean. Increases in AAC are apparent, particularly in the combination of treatment groups: 246 kPa + 40 min, 246 kPa + 50 min, and 570 kPa + 50 min. These changes are statistically significant (p < 0.05), as indicated by the asterisks, compared to that of the control group (0 kPa + 0 min). We used a one-way ANOVA followed by Bonferroni’s selected comparisons test to investigate for any statistical difference between the treatment groups. D Bar plots of changes in ASD from SGM for all treatment groups. E Bar plots of changes in ASD from FFSM for all treatment groups. We observed a decreasing trend in the mean-value of ASD as a function of HT duration from the combination of treatment groups. Nonetheless, the observed trend was not statistically significant. In B, C, D, and E ΔAAC-SGM, ΔAAC-FFSM, ΔASD-SGM, and ΔASD-FFSM indicates changes in AAC and ASD parameter from the SGM and FFSM models, respectively

Fig. 3

Representative plot of the estimated BSCs pre-and post-treatment, along with their respective best-fit form factor from the SGM model. These are shown from two combination of treatment groups 246 kPa + 50 min (graph depicting plot of the estimated BSCs pre-and post-treatment 570 kPa + 50 min is shown in Supplementary Figure 2). The BSCs are estimated from 50 individual attenuation-corrected normalized power spectra from two-dimensional tumour ROI. An increase in the acoustic backscattering resulting from cell deaths 24-h post-treatment, relative to the pre-treatment, is apparent from the BSC plot

Fig. 4

Representative dot plot for AAC and cell death. Correlation between changes in AAC and cell death fraction from TUNEL staining. Scatter plots of cell death fraction versus changes in AAC from all treatment groups. Apoptotic cell death percentile (brown area) was quantified from TUNEL images. We observed an increase in cell death as a function of HT duration for each combination of the treatment group. Increases in AAC due to cell deaths may have been caused by nuclear fragmentation and filling of the extracellular space with cell debris as tumour cells were treated with a combination of treatments of US-mediated vascular disruption and HT. Regression analysis of cell death fraction versus changes in AAC showed a strong correlation (r² = 0.62). The best quadratic curve-fitting to the data is shown in blue, with the goodness-of-fit function shown in the inset legend