Introduction

Prostate cancer is the most common non-skin cancer in men and can reappear in almost a third of high-risk patients after radiation treatment¹.

Magnetization transfer (MT)-prepared pulse sequences are sensitive to magnetization exchange between semisolid macromolecules and water, such as the changes in membrane structure that occur in lipid bilayers during apoptosis². The magnitude of this effect can be quantified with the magnetization transfer ratio (MTR)³. Chemical exchange saturation transfer (CEST) and relayed nuclear Overhauser effect (rNOE) are contrast mechanisms that are also measured using MT sequences. These mechanisms can be used as a non-invasive tool to detect tumour metabolism⁴.

This study investigated the MT and CEST characteristics of active and necrotic/apoptotic regions of prostate tumour xenografts in relation to their known response to radiation treatment, with the goal of determining the characteristics that may predict treatment response. This could lead to the creation of a novel, non-invasive way to evaluate a tumour's response to radiation, and in turn, treatment efficacy.

Methods

Approximately 3 x 10⁶ cells of DU145 human prostate adenocarcinoma were injected into the hind limbs of female athymic nude mice (n = 15). To simulate a standard human dose of radiation treatment, tumours were treated with a single dose of radiation at 6.3 Gy for 3.5 minutes after 4-6 weeks of growth.

MRI was performed at 7T (BioSpec 70/30 USR) 0-3 days pre-treatment and 8-20 days post-treatment. T₂-weighted structural images, T₁ and T₂ maps, and saturation transfer images with B₁ = 0.5, 2, 3 and 6 μT were acquired.

After the final MRI, tumours were excised and the imaged slices were stained for structure (H+E) and necrosis/apoptosis (TUNEL).

Images were automatically segmented into 5 regions (active tumour, necrosis/apoptosis, muscle/connective tissue, muscle and blood/edema) according to methods developed by our group⁵. MT and CEST spectra of regions deemed to be either active tumour or necrosis/apoptosis were fitted to a two-pool quantitative MT model⁶.

Results

Tumours were differentiated into three categories (responder, partial responder and non-responder) according to their response to the radiation treatment (Figure 1).

In the “responders” (n = 3), the segmentation showed an increase in relative areas of necrosis/apoptosis after treatment (Figure 2b), and the TUNEL images showed staining on the majority of the section. This indicated that the tumour was mostly necrotic.

“Partial responders” (n = 8) also had an increase in relative necrosis/apoptosis post-treatment in the segmentation and had large areas of TUNEL staining in the histology, but still showed areas of active tumour in both segmentation and histology.

“Non-responders” (n = 4) were those with a decrease in necrosis/apoptosis post-treatment, and did not show significant areas of TUNEL staining in the histology images.

MTR at 48 ppm and quantitative MT (qMT) parameters were averaged for each category of response (Figures 3-4). CEST and rNOE contributions (Figure 5) were calculated using a modification of the apparent exchange-dependent relaxation (AREX) formula⁷.

Discussion

Although it was unlikely to see the overall volume of the tumour decrease within the time frame of this experiment, possible early markers of treatment efficacy were observed in MTR, qMT parameters and CEST contributions of the studied tumours, mostly within the necrotic regions.

The average MTR with B₁ = 6 μT for the responding tumours prior to treatment was significantly less than for the non-responding tumours in both active (p < 0.05) and necrotic/apoptotic (p < 0.001) regions (Figure 3b, d). However, the only difference observed between pre- and post-treatment time points within a group was the decrease in the necrosis/apoptosis of the partial responders at 6 μT (Figure 3d). This may be driven by the decrease in R_MTM_0,MT for necrosis/apoptosis (p < 0.05) in the partial responders (Figure 4g). No significant changes were observed within groups in the active regions.

A similar phenomenon can be seen in Figure 4, where significant changes were also mostly observed in the necrotic/apoptotic regions of the tumour (Figure 4e-h). After treatment, both T_1,L and T_2,L for necrosis/apoptosis in the non-responders were significantly less than those for responders (p < 0.05) and partial responders (p < 0.01), yet no differences were seen between groups prior to treatment. This shows that T_1,L and T_2,L in necrotic/apoptotic regions of the tumour may be a good indication of how the tumour is responding to radiation treatment.

At -3.3 ppm, the CEST and rNOE contributions of active regions in partial responders prior to treatment were significantly less than those in responders and non-responders (p<0.05, Figure 5c). Contributions in the necrosis/apoptosis of partial responders decreased post-treatment at 3.5 ppm, and decreased for responders at -3.3 ppm (p < 0.05; Figure 5d, f). However, these contributions did not significantly change for non-responders over the course of treatment since metabolism slows as more cells in responding tumours become necrotic after radiation therapy.

Conclusion

The CEST and MT characteristics of prostate tumour xenografts with known responses to radiation therapy may help to identify characteristics that can indicate a tumour’s response to treatment in a non-invasive way. The differences in these characteristics, particularly in the necrosis/apoptosis, show that it may be possible to determine efficacy early on in the course of treatment.

Acknowledgements

The authors thank the Canadian Institutes of Health Research (grant number PJT148660) and Terry Fox Research Institute (grant number 1083) for funding.