Acute changes in liver tumour perfusion measured non-invasively with arterial spin labelling

Abstract

Background: Non-invasive measures of tumour vascular perfusion are desirable, in order to assess response to vascular targeting (or modifying) therapies. In this study, hepatic arterial spin labelling (ASL) magnetic resonance imaging (MRI) was investigated to measure acute changes in perfusion of colorectal cancer in the liver, in response to vascular disruption therapy with OXi4503.

Methods: SW1222 and LS174T tumours were established in the liver of MF1 nu/nu mice via intrasplenic injection. Perfusion and R2(*) MRI measurements were acquired with an Agilent 9.4T horizontal bore scanner, before and at 90 min after 40 mg kg(-1) OXi4503.

Results: A significant decrease in SW1222 tumour perfusion was observed (-43±33%, P<0.005). LS174T tumours had a significantly lower baseline level of perfusion. Intrinsic susceptibility MRI showed a significant increase in R2(*) in LS174T tumours (28±25%, P<0.05). An association was found between the change in tumour perfusion and the proximity to large vessels, with pre-treatment blood flow predictive of subsequent response. Histological evaluation confirmed the onset of necrosis and evidence of heterogeneous response between tumour deposits.

Conclusions: Hepatic ASL-MRI can detect acute response to targeted tumour vascular disruption entirely non-invasively. Hepatic ASL of liver tumours has potential for use in a clinical setting.

Figure 1

Baseline perfusion of tumour metastases. Average baseline estimates of blood perfusion from ASL (A) and R₂* from IS-MRI (B) in normal-appearing liver (n=6), SW1222 tumours (n=6) and LS174T tumours (n=3) (** indicates P<0.01, *** indicates P<0.005, **** indicates P<0.001 significant differences). Fluorescence microscopy of blood vessels (CD31, red) and perfusion (Hoechst 33342, blue) in SW1222 tumours (C) and LS174T tumours (D). Arrows indicate areas of normal liver tissue and stars indicate areas of tumour tissue.

Figure 2

Functional response of the tumour vasculature to disruption with OXi4503. Example high-resolution, axial images (A) with perfusion (top row) and R₂* (bottom row) maps overlaid on tumour ROIs (SW1222 on the left, and LS174T on the right, pre- and post OXi4503 therapy). A heterogeneous spatial distribution can be seen in both parameters, both within individual tumour deposits and across different tumours within the same liver. Histological staining by H&E is shown for saline control (B), and OXi4503-dosed SW1222 (C) and LS174T (D) tumours. The arrows highlight the centre of liver tumour deposits (B, viable and C and D, necrotic).

Figure 3

Correlation of baseline and post-OXi4503 measurements for perfusion (top row) and R₂* (bottom row). Graphs show the change in the value of each parameter in individual tumour deposits vs the baseline value. The change in perfusion measured in SW1222 tumours (A) displayed a significant negative correlation with baseline measurements (R²=0.76, P<0.001 Pearson's R). Better-perfused tumours at baseline exhibited a greater loss of perfusion following OXi4503 administration. LS174T tumour perfusion (B) was low at baseline, and no similar trend was found. For both SW1222 (C) and LS174T (D) tumours, the R₂* baseline data showed no correlation with baseline values.

Figure 4

The dependency of tumour location within the liver and the magnitude of perfusion changes. Analysis of SW1222 tumours following dosing with OXi4503. (A) A three-dimensional visualisation of an example mouse liver at week 5 post implantation, showing liver parenchyma (semi-transparent red), and major vasculature (purple). Tumours have been coloured according to their proximity to either a large vessel (>0.1 mm², red), small vessel (green) or no visible vessel (blue). (B) The percentage change in perfusion and R₂* at 90 min following administration of 40 mg kg⁻¹ of OXi4503, categorised according to proximity to vessels and location within the liver. Large perfusion decreases were observed in tumours close to either large (LV) or small (SV) blood vessels, although tumours near smaller vessels demonstrated a more variable response (as evidenced by larger error bars). Tumours at the periphery (P) of the liver did not display a significant change in perfusion. No significant changes were measured in R₂* in any of the liver locations. (C) Example H&E slice of liver containing three different tumours located at a range of position to a blood vessel, demonstrating a heterogeneous response to OXi4503 within the same liver.