Dynamic contrast–enhanced MRI (DCE-MRI) is a promising technique for assessing the response of tumor vasculature to antivascular therapies. Multiagent DCE-MRI employs a combination of low and high molecular weight contrast agents, which potentially improves the accuracy of estimation of tumor hemodynamic and vascular permeability parameters. In this study, we used multiagent DCE-MRI to assess changes in tumor hemodynamics and vascular permeability after vascular-disrupting therapy. Multiagent DCE-MRI (sequential injection of G5 dendrimer, G2 dendrimer, and Gd-DOTA) was performed in tumor-bearing mice before, 2 and 24 hours after treatment with vascular disrupting agent DMXAA or placebo. Constrained DCE-MRI gamma capillary transit time modeling was used to estimate flow F, blood volume fraction vb, mean capillary transit time tc, bolus arrival time td, extracellular extravascular fraction ve, vascular heterogeneity index α−1 (all identical between agents) and extraction fraction E (reflective of permeability), and transfer constant Ktrans (both agent-specific) in perfused pixels. F, vb, and α−1 decreased at both time points after DMXAA, whereas tc increased. E (G2 and G5) showed an initial increase, after which, both parameters restored. Ktrans (G2 and Gd-DOTA) decreased at both time points after treatment. In the control, placebo-treated animals, only F, tc, and Ktrans Gd-DOTA showed significant changes. Histologic perfused tumor fraction was significantly lower in DMXAA-treated versus control animals. Our results show how multiagent tracer-kinetic modeling can accurately determine the effects of vascular-disrupting therapy by separating simultaneous changes in tumor hemodynamics and vascular permeability.

Significance: These findings describe a new approach to measure separately the effects of antivascular therapy on tumor hemodynamics and vascular permeability, which could help more rapidly and accurately assess the efficacy of experimental therapy of this class. Cancer Res; 78(6); 1561–70. ©2018 AACR.

During early development of anti-vascular cancer drugs, a number of treatment effects are commonly investigated, including the mechanism of drug action, required dose, optimal therapy scheduling, and therapeutic efficacy (1, 2). Functional imaging techniques can provide quantitative imaging biomarkers to enable early assessment of treatment effects. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI), predominantly using low molecular weight contrast agents, is widely applied in both preclinical and clinical studies for noninvasive and longitudinal monitoring of anti-vascular therapies (1–3). Tracer-kinetic modeling can be performed to calculate pharmacokinetic parameters that reflect microvascular functionality (4–6). One of the most commonly used tracer-kinetic parameters is the transfer constant Ktrans from the model introduced by Tofts and Kermode (hereafter named the Tofts model), which describes the transfer rate of contrast material from the intravascular space into the tumor interstitial space (7–9). Typically, a reduction in Ktrans is observed after anti-vascular therapy, providing useful information on therapeutic effects (1, 2). However, Ktrans is a composite measure that is influenced by both blood flow and microvascular permeability (10). These effects can often not be separately determined, because many anti-vascular therapies induce changes in both of these parameters. As an example, vascular targeting therapies can cause increased microvascular permeability due to vascular disruption along with decreased blood flow due to vascular collapse and hemorrhage (11–13). Several studies have shown that macromolecular contrast agents may be more sensitive to changes in microvascular permeability and blood volume (14, 15) and that low molecular weight and macromolecular agents could be combined to obtain an improved assessment of tracer-kinetic parameters (16–18). However, a more detailed modeling approach may be required to fully identify the various vascular changes upon treatment.

Recently, we used a multiagent DCE-MRI approach, in which contrast agents of various molecular weight [generation 5 (G5) and generation 2 (G2) poly(propylene imine) (PPI) dendrimers functionalized with Gd-DOTA moieties and low-molecular weight contrast agent Gd-DOTA] were sequentially injected within one imaging session in non-treated tumor-bearing mice (19). Simultaneous modeling of the multiagent DCE-MRI data was performed, for which a number of hemodynamic parameters, including blood flow, were constrained to be identical between the different boluses, whereas tracer-kinetic parameters related to microvascular permeability were separately determined for each agent. It was demonstrated that the hemodynamic parameters could be separately determined from vascular permeability using this approach. In addition, a contrast agent size-dependent decrease in the extraction fraction was measured, reflecting the lower degree of vascular permeability for macromolecular contrast agents. It was hypothesized that this approach can be used to evaluate the various changes in these tracer-kinetic parameters upon treatment targeting the vasculature.

The purpose of the present study was to apply this new multiagent DCE-MRI method in a therapeutic setting and investigate whether it can provide a detailed assessment of changes in both hemodynamic parameters and microvascular permeability after vascular disrupting therapy of tumors in mice with the model drug DMXAA.

Ethics statement

All animal experiments in this prospective study were performed according to the Directive 2010/63/EU of the European Commission and approved by the Animal Care and Use Committee of Maastricht University, the Netherlands.

Animal model

Twelve to 14-week-old BALB/cByJ mice (Charles River Laboratories) were inoculated with 1 × 106 CT26.WT mouse colon carcinoma cells (CRL-2638, ATCC) subcutaneously in their right hind limb. Approximately 10 days after inoculation, tumors became palpable in all animals.

Study design

Baseline MRI measurements were performed 1 day before either anti-vascular treatment (n = 13) or placebo treatment (n = 16). Subsequently, mice were randomly assigned to one of these treatment groups. Mice in the anti-vascular treatment group were intraperitoneally injected with the vascular disrupting agent 5,6-Dimethylxanthenone-4-acetic acid (DMXAA, Sigma-Aldrich), freshly prepared at a dose of 20 mg/kg DMXAA dissolved in 5% sodium bicarbonate in ultrapure water (0.5 mg DMXAA/100 μL). This agent was used as a model therapeutic agent, since the vascular shutdown by DMXAA is known to result from increased permeability (20). Mice in the placebo treatment group were intraperitoneally injected with an equivalent volume of 5% sodium bicarbonate in ultrapure water. At 2 hours after either DMXAA (n = 13) or placebo (n = 16) treatment, mice underwent post-treatment MRI measurements. After the 2 hours post-treatment MRI measurements, some of the animals (DMXAA n = 6, placebo n = 8) were sacrificed. The remaining animals additionally underwent MRI at 24 hours after DMXAA (n = 7) or placebo (n = 8) treatment and were subsequently sacrificed.

Contrast agents for multiagent DCE-MRI

Multiagent DCE-MRI measurements were performed using contrast agents with a range of molecular weights. Modified PPI dendrimers, functionalized with Gd-DOTA moieties via a polyethylene glycol (PEG) spacer, were synthesized by SyMO-Chem BV (Eindhoven, The Netherlands). These agents comprised a high molecular weight (59517 Da) generation 5 dendrimer (G5-PPI-(PEG6-GdDOTA)64) and an intermediate molecular weight (7317 Da) generation 2 dendrimer (G2-PPI-(PEG6-GdDOTA)8). In addition to these macromolecular agents, the clinically available low molecular weight (754 Da) gadoterate meglumine (Gd-DOTA, Dotarem, Guerbet) was used. Contrast agent relaxivity in Balb/c mouse serum (Innovative Research) was measured at 7 T and 37°C. Additional information on synthesis and characterization of the dendrimer-based contrast agents is provided in the Supplementary Information with additional figures and a table showing results of the dendrimer characterization (Supplementary Fig. S1: synthesis schemes of the G2 and G5 dendrimers; Supplementary Fig. S2: mass spectroscopy spectrum of the PPI building blocks; Supplementary Fig. S3: mass spectroscopy spectrum of G2 and gel permeation chromatography spectrums of the G2 and G5 dendrimers; Supplementary Fig. S4: size measurements of the G2 and G5 dendrimers using dynamic light scattering; Supplementary Table S1: longitudinal (r1) and transverse (r2) MRI relaxivity values of GdDOTA, G2 and G5). The dendrimer-based contrast agents were dissolved in PBS. For contrast agent administration, an infusion line was filled with equal volumes of the G5 dendrimer, G2 dendrimer and Gd-DOTA (at a concentration of 125 mmol/L, ∼0.1 mmol/kg), separated by equal volumes of saline. Small air bubbles between each volume prevented mixing of the agents. The gadolinium concentrations of the solutions used in the in vivo experiments were determined with inductively coupled plasma-optical emission spectrometry and the injected volumes were corrected for small differences in the actual concentration.

MRI acquisition and analysis

Animal preparation.

Mice were anesthetized using isoflurane (3% for induction and 1%–2% for maintenance) in medical air (at a flow rate of 0.4 L/min). A tail vein catheter for contrast agent injections was placed, after which the mice were positioned in a custom-made cradle, equipped with an anesthesia mask and warm water circuit to maintain body temperature at 37°C. MRI measurements were performed with a 7T Bruker BioSpec 70/30 USR (Bruker BioSpin MRI GmbH) equipped with a 1H 59/35 (outer/inner diameter in mm) circular polarized MRI transceiver volume coil (Bruker BioSpin MRI GmbH). During MRI measurements, respiration rate was monitored using a balloon pressure sensor and temperature with a rectal temperature probe.

B0 and B1 calibration.

First, local shimming of the tumor was performed. Thereafter, anatomical reference T2-weighted images were acquired using a three-dimensional turbo rapid acquisition with relaxation enhancement (RARE) sequence. Sequence parameters were: repetition time/effective echo time = 2,000/34.4 ms, flip angle = 90°, field of view = 30 × 30 × 24 mm3, acquisition matrix = 75 × 75 × 16, and RARE factor = 8. To correct for local inhomogeneities in the radiofrequency field, B1 mapping was performed with the 180° signal null method (21) using a gradient-spoiled and radiofrequency-spoiled three-dimensional T1-weighted fast low-angle shot (FLASH) sequence. The flip angles of the global excitation pulses were centered around 180° (flip angles = 120°, 150°, 180°, 210°, 240°). Other sequence parameters were: repetition time/echo time = 100/3.03 ms, field of view = 30 × 30 × 48 mm3, and acquisition matrix = 56 × 42 × 26 reconstructed to 75 × 75 × 32. Only the center slices, corresponding to the anatomical T2-weighted images, were selected for further analysis. Data were fitted to determine the nominal flip angle required for signal nulling, from which the B1-correction factors were calculated.

T1 mapping.

The longitudinal relaxation time T1 before contrast agent injection was derived from T1-weighted images that were acquired using a gradient-spoiled and radiofrequency-spoiled three-dimensional FLASH sequence with variable flip angles. Detailed sequence parameters were: repetition time/echo time = 1.38/0.69 ms, flip angles = 1°, 2°, 3°, 5°, 7°, 10°, 13°, 20°, field of view = 30 × 30 × 24 mm3, acquisition matrix = 50 × 39 × 14, reconstructed to 75 × 75 × 16, and 15 repetitions. Only repetitions 8-15, at which a steady-state signal intensity was obtained, were used for further analysis. B1 inhomogeneity correction was performed by multiplication of the flip angles by B1 correction factors in each pixel. Non-linear regression of the variable flip angle T1-mapping data was performed using the standard spoiled-gradient echo signal-equation to determine the pre-contrast T1 values for each voxel.

Multiagent DCE-MRI.

Multiagent DCE-MRI measurements were performed using the same sequence and sequence parameters as for pre-contrast T1 mapping, with a flip angle of 15° and 730 repetitions, resulting in a temporal resolution of 0.82 seconds and total acquisition time of 10 minutes. Contrast agent injections were performed at 1 (G5 dendrimer), 4 (G2 dendrimer) and 7 (gadoterate meglumine) minutes after start of the acquisition, using an infusion pump (Chemyx Fusion 100) at a rate of 2 mL/min. B1-corrected dynamic ΔR1 (=1/ΔT1) values were calculated for each voxel based on the standard spoiled gradient-echo signal-equation, using the pre-contrast T1 values and post-contrast dynamic signal intensities. Regions of interest delineating the tumor were manually drawn using the anatomical T2-weighted images.

Multiagent tracer-kinetic modeling

Details of the multiagent tracer-kinetic modeling have been described in our previously published work (19). In short, arterial input functions (AIF) were determined for each agent separately using the Monte Carlo Blind Estimation algorithm (22, 23) applied to the ΔR1 curves of tumor voxels that exhibited significant perfusion. A pixel was considered perfused if the mean ΔR1 between the first and second contrast agent injection was higher than 2 times the standard deviation of ΔR1 before contrast injection. The DCE-MRI–derived perfused fraction (PF) was calculated from this analysis. Multiagent curves in the perfused pixels were simultaneously fitted with the gamma capillary transit time (GCTT) model (24). This model characterizes contrast uptake curves with six model parameters: blood flow F, extraction fraction E [i.e., the fraction of contrast agent molecules that leak into the extravascular extracellular space (EES), reflective of vascular permeability], washout rate constant kep of the contrast agent molecules from EES to intravascular space, mean capillary transit time tc, bolus arrival time td and vascular heterogeneity index α−1, the latter representing the width of the distribution of tc within a tissue voxel. In principle, each bolus of injected contrast agent could be separately fitted, resulting in 18 model parameters. By imposing physiological constraints, four of these six parameters (F, tc, td, and α−1) were identical between the different boluses, whereas E and kep were separately determined for each contrast agent. From the estimated parameters transfer constant Ktrans = E × F, extravascular extracellular volume fraction ve = E × F/kep and blood volume fraction vb = F × tc were derived. The latter three parameters, Ktrans, ve, and vb, are all estimated parameters that are described in the conventionally used extended Tofts model (9). As indicated earlier, Ktrans is a composite measure of vascular permeability and blood flow. Only the perfused pixels were included in the DCE-MRI fitting, because the parameters cannot be accurately determined in pixels that do not exhibit significant enhancement. Median tumor parameter values for each measurement were calculated for each MRI examination.

Histologic analysis

After the MRI measurements, a subset of mice [2 hours after placebo (n = 4), 24 hours after placebo (n = 4), 2 hours after DMXAA (n = 3), 24 hours after DMXAA (n = 4)] was injected with bisBenzimide H 33342 trihydrochloride (Hoechst, Sigma-Aldrich) as a perfusion marker.

For the Hoechst staining, mice were kept anesthetized after the MRI measurement and Hoechst (8 mg/mL in saline, 32 mg/kg) was injected via the tail vein. After 3 minutes, mice were sacrificed by cervical dislocation. Further processing of the tumor tissue is described in the Supplementary Information. In short, the frozen tumor sections were subsequently stained for endothelial cells. Semi-automated analysis of the stained slides was performed to determine perfused vessel fractions (PF) from the Hoechst staining and vascular density (VD) from the endothelial staining.

Statistical analysis

Data are presented as mean ± standard deviation (SD). Statistical analysis was performed in SPSS (version 20, IBM). Normality of the tracer-kinetic DCE-MRI parameter values and histopathologic data in the different groups was confirmed using a Kolmogorov–Smirnov test (P > 0.315) and subsequently parametric tests were used for the statistical analyses. Linear mixed models were used to assess whether the multiagent DCE-MRI parameters significantly changed over time in the DMXAA-treated and placebo-treated animal groups. If a certain parameter showed significant differences over time, the parameter values were compared between the different time points using Bonferroni-corrected two-sided t tests. The quantitative histologic measurements between DMXAA-treated and placebo-treated animals were compared using two-sided Student t tests. For all tests, a P value lower than 0.05 was considered significant.

Representative multiagent DCE-MRI parameter maps of a DMXAA-treated and placebo-treated animal are shown in Figs. 1 and 2, respectively. Although the DCE-MRI parameters overall remained stable in the placebo-treated animals, substantial treatment effects were observed in the parameter maps of the DMXAA-treated animals. Both at 2 and at 24 hours after treatment a large region of the tumor was nonperfused. In the remaining perfused pixels of the tumor, substantially decreased flow was observed at 2 and at 24 hours. In addition, a pronounced increase in permeability was observed at 2 hours, as expressed by an increase in the extraction fraction E of G5 and G2 and to a lesser extent in the extraction fraction of Gd-DOTA. The increase in permeability was not reflected in the Ktrans maps. Ktrans of G2 and G5 did not show visual differences in the perfused pixels upon treatment, whereas Ktrans of Gd-DOTA showed a decrease at both time points after treatment.

Figure 1.

Multiagent DCE-MRI parameter maps before and longitudinally after DXMAA treatment. The parameter maps in the tumor tissue are superimposed on the anatomic T2-weighted image. The maps are colored according to the color bar on the right-hand side of the plot. The corresponding parameter range for this scale bar is shown on the left side of each row of panels. The white tumor pixels represent nonperfused pixels. After treatment, a large portion of the tumor was nonperfused. In the remaining perfused area, a pronounced decrease in flow was observed at both 2 and 24 hours after treatment. The extraction fraction, particularly of G5 and G2, initially increased at 2 hours after treatment and then restored again to baseline values at 24 hours after treatment. The increased permeability was not reflected in the Ktrans values due to the confounding influence of diminished blood flow. E, extraction fraction; F, blood flow; Ktrans, transfer constant.

Figure 1.

Multiagent DCE-MRI parameter maps before and longitudinally after DXMAA treatment. The parameter maps in the tumor tissue are superimposed on the anatomic T2-weighted image. The maps are colored according to the color bar on the right-hand side of the plot. The corresponding parameter range for this scale bar is shown on the left side of each row of panels. The white tumor pixels represent nonperfused pixels. After treatment, a large portion of the tumor was nonperfused. In the remaining perfused area, a pronounced decrease in flow was observed at both 2 and 24 hours after treatment. The extraction fraction, particularly of G5 and G2, initially increased at 2 hours after treatment and then restored again to baseline values at 24 hours after treatment. The increased permeability was not reflected in the Ktrans values due to the confounding influence of diminished blood flow. E, extraction fraction; F, blood flow; Ktrans, transfer constant.

Close modal
Figure 2.

Multiagent DCE-MRI parameter maps before and longitudinally after placebo treatment. The parameter maps in the tumor tissue are superimposed on the anatomic T2-weighted image. The maps are colored according to the color bar on the right-hand side of the plot. The corresponding parameter range for this scale bar is shown on the left side of each row of panels. The white tumor pixels represent nonperfused pixels. The DCE-MRI parameters did not show differences between the time points. E, extraction fraction; F, blood flow; Ktrans, transfer constant.

Figure 2.

Multiagent DCE-MRI parameter maps before and longitudinally after placebo treatment. The parameter maps in the tumor tissue are superimposed on the anatomic T2-weighted image. The maps are colored according to the color bar on the right-hand side of the plot. The corresponding parameter range for this scale bar is shown on the left side of each row of panels. The white tumor pixels represent nonperfused pixels. The DCE-MRI parameters did not show differences between the time points. E, extraction fraction; F, blood flow; Ktrans, transfer constant.

Close modal

The mean quantitative DCE-MRI parameters of the perfused pixels at the different time points are presented in Fig. 3. In the placebo-treated animals most of the parameters did not significantly change over time, except for a mild, yet significant, decrease in F and Ktrans Gd-DOTA at both time points after placebo treatment (P < 0.005 and P < 0.016, respectively) and an increase in tc at 2 hours after treatment (P = 0.008). In the DMXAA-treated animals, at 2 and 24 hours after treatment most parameters were significantly different from baseline. From the hemodynamic parameters that are non–contrast-agent specific, PF (P < 0.001), F (P < 0.001), vb (P < 0.001), and α−1 (P < 0.001) significantly decreased at both time points after the DMXAA treatment compared with baseline measurements, whereas tc significantly increased at both post-treatment time points (P < 0.040). The DMXAA treatment did not induce significant changes in ve (P > 0.082). The changes in contrast agent–specific DCE-MRI parameters after DMXAA treatment were markedly different among the contrast agents. E G5 initially strongly increased at 2 hours after DMXAA treatment (P < 0.001), whereupon it significantly dropped again at 24 hours (P < 0.001). E G2 also significantly increased at 2 hours after DMXAA treatment (P = 0.003), after which its value seemed to drop again at 24 hours, yet not significantly with respect to the 2-hour time point (P = 0.206). E Gd-DOTA did not show significant change upon DMXAA treatment. Finally, Ktrans G5 did not change upon treatment, while both Ktrans G2 and Ktrans Gd-DOTA significantly decreased at both post-treatment time points (P < 0.024 and P < 0.001, respectively).

Figure 3.

Treatment-induced changes in multiagent DCE-MRI parameters. Mean ± SD multiagent DCE-MRI parameter values over time in the perfused pixels of the DMXAA-treated (solid line) and placebo-treated (dashed line) tumors. *, P < 0.05 and **, P < 0.001, significant difference compared with baseline. ##, P < 0.001, significant difference compared with 2 hours after treatment. α−1, vascular heterogeneity index; E, extraction fraction; F, blood flow; Ktrans, transfer constant; PF, perfused fraction; tc, mean capillary transit time; vb, blood volume fraction; ve, extravascular extracellular volume fraction.

Figure 3.

Treatment-induced changes in multiagent DCE-MRI parameters. Mean ± SD multiagent DCE-MRI parameter values over time in the perfused pixels of the DMXAA-treated (solid line) and placebo-treated (dashed line) tumors. *, P < 0.05 and **, P < 0.001, significant difference compared with baseline. ##, P < 0.001, significant difference compared with 2 hours after treatment. α−1, vascular heterogeneity index; E, extraction fraction; F, blood flow; Ktrans, transfer constant; PF, perfused fraction; tc, mean capillary transit time; vb, blood volume fraction; ve, extravascular extracellular volume fraction.

Close modal

The histopathologic results from Hoechst perfusion measurements are shown in Fig. 4A and B. DMXAA-treated animals exhibited large regions of nonperfused tissue in histology. Histologic PF was significantly lower for the DMXAA-treated animals at both time points after treatment compared to the placebo-treated animals at the same time points (P < 0.005). VD from histology was not significantly different between the DMXAA-treated and control animals (P > 0.094).

Figure 4.

Treatment-induced changes in histologic measurements of the tumor vasculature. A, Histologic images of a tumor section of a mouse sacrificed 2 hours after placebo treatment (left) and a mouse sacrificed 2 hours after DMXAA treatment (right). Scale bars, 1 mm. Zoomed images of the regions indicated with the white rectangles are shown on the bottom row. The blue signal originates from the Hoechst dye and represents perfusion. The red signal originates from the endothelium staining and represents vasculature. In the DMXAA-treated animals, large regions of nonperfused tumor tissue were observed. B, Bar charts of perfusion fraction (PF) and vascular density (VD) at 2 and 24 hours after placebo and DMXAA treatment. *, significant difference in PF between placebo-treated and DMXAA-treated groups at 2 hours after treatment (P = 0.005). ##, significant difference between placebo-treated and DMXAA-treated groups at 24 hours after treatment (P < 0.001).

Figure 4.

Treatment-induced changes in histologic measurements of the tumor vasculature. A, Histologic images of a tumor section of a mouse sacrificed 2 hours after placebo treatment (left) and a mouse sacrificed 2 hours after DMXAA treatment (right). Scale bars, 1 mm. Zoomed images of the regions indicated with the white rectangles are shown on the bottom row. The blue signal originates from the Hoechst dye and represents perfusion. The red signal originates from the endothelium staining and represents vasculature. In the DMXAA-treated animals, large regions of nonperfused tumor tissue were observed. B, Bar charts of perfusion fraction (PF) and vascular density (VD) at 2 and 24 hours after placebo and DMXAA treatment. *, significant difference in PF between placebo-treated and DMXAA-treated groups at 2 hours after treatment (P = 0.005). ##, significant difference between placebo-treated and DMXAA-treated groups at 24 hours after treatment (P < 0.001).

Close modal

In this article, we presented the first application of a recently developed multi- tracer-kinetic modeling approach (19) for the assessment of tumor treatment response. Sequential injection of different contrast agents has been employed previously for evaluation of tumor vascular properties using DCE-MRI (17, 18, 25–27). However, in these previous studies the different injections were separately analyzed, whereas simultaneous modeling of the multiagent data facilitates a comprehensive and robust assessment of tissue properties by exploiting the fact that a number of parameters can be constrained to be identical between the agents based on physiological grounds (19).

The vascular disrupting agent DMXAA was chosen as model agent for the assessment of anti-vascular therapy. DMXAA (also known as Vadimezan and ASA404) induces a cascade of anti-vascular events in the tumor early after administration (13, 28). Within 30 minutes after administration, it induces endothelial cell apoptosis, leading to increased vascular permeability and decreased blood perfusion resulting from loss of plasma and associated higher blood viscosity (28). The subsequent loss of endothelial barrier causes increased vascular permeability (20). A second mechanism of action of DMXAA is the induction of TNFα among other cytokines (28). TNFα contributes to increase in vascular permeability (29). The enhanced permeability causes an increase in interstitial pressure, which further compromises tumor blood flow (30). Generally, in mouse models this cascade of events completely shuts down the tumor vasculature, after a transient period of decreased blood flow and high vascular permeability. In clinical trials with DMXAA in patients the anti-vascular cascade was observed only partly, with minor increase in TNFα after DMXAA administration (13). Because of the limited therapeutic effects in patients, the development of DMXAA has been discontinued in 2010 after unsuccessful phase III trials, in which no effects of DMXAA on overall survival could be proven (31). Nevertheless, DMXAA remains an interesting drug to investigate in the preclinical setting because of its well-defined anti-vascular effects, which can be used to investigate novel imaging methods for the characterization of tumor vasculature, as was done in the present study.

The here observed effects of treatment on DCE-MRI parameters are exactly in line with the previously described vascular events following administration of DMXAA. The rapid drop in PF, F, and vb is consistent with induction of endothelial apoptosis and subsequent vascular shutdown early after administration. The increase in tc at both time points after treatment is likely related to high blood viscosity and associated low blood flow. The significant drop in α−1 after treatment is indicative of a narrower distribution of tc in the tumor pixels and a lower vascular heterogeneity. A possible explanation for this finding is that the response of blood vessels to the vascular therapy depends on the blood flow (28). Vessels with low blood flow may be particularly sensitive to vascular disrupting agents (28). The DMXAA may thus have particularly affected these blood vessels, leading to immediate vascular shutdown. Removal of these specific capillary paths by DMXAA treatment may have resulted in the observed more homogeneous distribution of capillary transit times in the remaining perfused vasculature. Next to the overall diminished blood flow, we observed a transient increase in permeability (E G2 and E G5) at 2 hours, which dropped again at 24 hours after treatment. This finding is in agreement with endothelial barrier loss and increased vascular permeability after DMXAA treatment. The drop in vascular permeability at 24 hours after treatment could be a result from increased interstitial pressure (30).

An important result of this study is that the increase in permeability would not have been detected using only a low molecular weight contrast agent, such as Gd-DOTA, as we did not observe an increase in E Gd-DOTA at 2 hours after DMXAA treatment. This finding further advocates the use of high molecular weight contrast agents for characterization of tumor vasculature (14). Moreover, the use of differently sized contrast agents in the same acquisition can give additional insights in the vascular permeability, such as the endothelial pore size (3). Our results also further stress that a more complete insight in vascular changes can be obtained from advanced modeling of macromolecular DCE-MRI data, compared with conventional Tofts modeling. Ktrans of G2 and G5 both failed to show an increase in permeability after treatment, because of the counteracting effects of decreased flow and increased permeability on Ktrans (10).

In the placebo-treated control animals generally no changes in DCE-MRI parameters were observed over time, except for a decrease in F and Ktrans Gd-DOTA and increase in tc. The reduction in tumor perfusion may be related to the repeated anesthesia. It has previously been shown that repeated isoflurane anesthesia leads to a decreased heart rate and blood pressure in rats (32), although we have not confirmed this by monitoring of heart rate and blood pressure for the mice in this study. This systemic physiological effect of anesthesia likely also influences the local blood flow in the tumor tissue.

The histopathologic results from Hoechst staining corroborated the DCE-MRI findings. Although the number of vessels was similar between DMXAA- and placebo-treated animals, there was a significant drop in the histology-derived PF in the DMXAA-treated animals, similar as seen in DCE-MRI.

Although we used multiagent tracer-kinetic modeling for assessment of vascular effects after treatment with a vascular disrupting agent, this method could also be used for evaluation of other anti-vascular therapies. The technique could be particularly useful for characterization of the dynamic effects of anti-angiogenic therapies on the tumor vasculature. Anti-angiogenic agents may induce transient normalization of the tumor vasculature, prior to vascular shutdown (33). Vessel normalization is characterized by reduction of the abnormal tortuosity and hyperpermeability of tumor blood vessels, leading to reduced interstitial pressure and increased blood flow (34). This transient period of vessel normalization can be exploited by combination therapies of anti-angiogenic and cytotoxic agents (34). For full efficacy of these combination therapies it is important to optimize the treatment schedule and dosing. Although Tofts modeling may not be sensitive enough for detection of vessel normalization, because of the counteracting effects of higher blood flow and lower permeability on Ktrans, the multiagent approach, combined with a more sophisticated model of the tissue microvasculature (such as the GCTT model used here), is able to differentiate between these effects and is therefore probably more suitable for evaluation of anti-angiogenic therapies.

Whereas the application of the multiagent tracer-kinetic modeling is mainly focused on the preclinical evaluation of tumor vasculature, it is also relevant to humans. Recently, the feasibility of clinical application of the multiagent modeling approach has been shown in patients with (high risk of) pancreatic cancer (35). It was shown that the multiagent approach is promising for the identification of pancreatic malignancies. In that particular study, the injection protocol consisted of sequential injection of low-molecular weight gadoteridol and the iron-based nanoparticle blood pool agent ferumoxytol.

Our study had some limitations. Only one vascular therapeutic agent was evaluated. The sensitivity of the technique to the vascular effects of other therapeutic agents needs to be assessed in future studies. In addition, the method was only tested in one animal tumor model. While we expect that the method is translatable to other animal models as well as to human applications, future studies in different cancer models are needed to further validate the use of the multiagent tracer-kinetic modeling technique for assessment of anti-vascular therapies.

In conclusion, we have shown that multiagent tracer-kinetic modeling can be used to accurately determine the effects of a vascular disrupting agent on the tumor vasculature. The technique could separate simultaneous effects on tumor hemodynamics and vascular permeability, whereas conventional DCE-MRI failed to discriminate between these vascular changes.

No potential conflicts of interest were disclosed.

Conception and design: S.J. Hectors, I. Jacobs, H.M. Janssen, K. Nicolay, M.C. Schabel, G.J. Strijkers

Development of methodology: S.J. Hectors, I. Jacobs, H.M. Keizer, M.C. Schabel, G.J. Strijkers

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S.J. Hectors, I. Jacobs, J. Bussink, M.C. Schabel

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S.J. Hectors, I. Jacobs, J. Lok, J. Peters, J. Bussink, M.C. Schabel

Writing, review, and/or revision of the manuscript: S.J. Hectors, I. Jacobs, J. Lok, J. Bussink, H.M. Keizer, H.M. Janssen, M.C. Schabel, G.J. Strijkers

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S.J. Hectors, I. Jacobs, J. Lok, J. Peters, H.M. Keizer, M.C. Schabel, G.J. Strijkers

Study supervision: S.J. Hectors, I. Jacobs, K. Nicolay, G.J. Strijkers

Other (synthesis): F.J. Hoeben

Other (algorithm design, software development and testing, generation of figures): M.C. Schabel

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
O'Connor
JP
,
Jackson
A
,
Parker
GJ
,
Roberts
C
,
Jayson
GC
. 
Dynamic contrast-enhanced MRI in clinical trials of antivascular therapies
.
Nat Rev Clin Oncol
2012
;
9
:
167
77
.
2.
Zweifel
M
,
Padhani
AR
. 
Perfusion MRI in the early clinical development of antivascular drugs: decorations or decision making tools?
Eur J Nucl Med Mol Imaging
2010
;
37
:
S164
82
.
3.
Nielsen
T
,
Wittenborn
T
,
Horsman
MR
. 
Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) in preclinical studies of antivascular treatments
.
Pharmaceutics
2012
;
4
:
563
89
.
4.
Ferl
GZ
,
Port
RE
. 
Quantification of antiangiogenic and antivascular drug activity by kinetic analysis of DCE-MRI data
.
Clin Pharmacol Ther
2012
;
92
:
118
24
.
5.
Sourbron
SP
,
Buckley
DL
. 
Tracer kinetic modelling in MRI: estimating perfusion and capillary permeability
.
Phys Med Biol
2012
;
57
:
R1
33
.
6.
Sourbron
SP
,
Buckley
DL
. 
Classic models for dynamic contrast-enhanced MRI
.
NMR Biomed
2013
;
26
:
1004
27
.
7.
Tofts
PS
,
Kermode
AG
. 
Measurement of the blood-brain barrier permeability and leakage space using dynamic MR imaging. 1. Fundamental concepts
.
Magn Reson Med
1991
;
17
:
357
67
.
8.
Tofts
PS
. 
Modeling tracer kinetics in dynamic Gd-DTPA MR imaging
.
J Magn Reson Imaging
1997
;
7
:
91
101
.
9.
Tofts
PS
,
Brix
G
,
Buckley
DL
,
Evelhoch
JL
,
Henderson
E
,
Knopp
MV
, et al
Estimating kinetic parameters from dynamic contrast-enhanced T1-weighted MRI of a diffusable tracer: standardized quantities and symbols
.
J Magn Reson Imaging
1999
;
10
:
223
32
.
10.
Sourbron
SP
,
Buckley
DL
. 
On the scope and interpretation of the Tofts models for DCE-MRI
.
Magn Reson Med
2011
;
66
:
735
45
.
11.
Siemann
DW
. 
The unique characteristics of tumor vasculature and preclinical evidence for its selective disruption by tumor-vascular disrupting agents
.
Cancer Treat Rev
2011
;
37
:
63
74
.
12.
Tozer
GM
,
Prise
VE
,
Wilson
J
,
Cemazar
M
,
Shan
S
,
Dewhirst
MW
, et al
Mechanisms associated with tumor vascular shut-down induced by combretastatin A-4 phosphate: intravital microscopy and measurement of vascular permeability
.
Cancer Res
2001
;
61
:
6413
22
.
13.
Baguley
BC
,
Siemann
DW
. 
Temporal aspects of the action of ASA404 (vadimezan; DMXAA)
.
Expert Opin Investig Drugs
2010
;
19
:
1413
25
.
14.
de Lussanet
QG
,
Langereis
S
,
Beets-Tan
RG
,
van Genderen
MH
,
Griffioen
AW
,
van Engelshoven
JM
, et al
Dynamic contrast-enhanced MR imaging kinetic parameters and molecular weight of dendritic contrast agents in tumor angiogenesis in mice
.
Radiology
2005
;
235
:
65
72
.
15.
Jaspers
K
,
Aerts
HJ
,
Leiner
T
,
Oostendorp
M
,
van Riel
NA
,
Post
MJ
, et al
Reliability of pharmacokinetic parameters: small vs. medium-sized contrast agents
.
Magn Reson Med
2009
;
62
:
779
87
.
16.
Lemasson
B
,
Serduc
R
,
Maisin
C
,
Bouchet
A
,
Coquery
N
,
Robert
P
, et al
Monitoring blood-brain barrier status in a rat model of glioma receiving therapy: dual injection of low-molecular-weight and macromolecular MR contrast media
.
Radiology
2010
;
257
:
342
52
.
17.
Orth
RC
,
Bankson
J
,
Price
R
,
Jackson
EF
. 
Comparison of single- and dual-tracer pharmacokinetic modeling of dynamic contrast-enhanced MRI data using low, medium, and high molecular weight contrast agents
.
Magn Reson Med
2007
;
58
:
705
16
.
18.
Henderson
E
,
Sykes
J
,
Drost
D
,
Weinmann
HJ
,
Rutt
BK
,
Lee
TY
. 
Simultaneous MRI measurement of blood flow, blood volume, and capillary permeability in mammary tumors using two different contrast agents
.
J Magn Reson Imaging
2000
;
12
:
991
1003
.
19.
Jacobs
I
,
Strijkers
GJ
,
Keizer
HM
,
Janssen
HM
,
Nicolay
K
,
Schabel
MC
. 
A novel approach to tracer-kinetic modeling for (macromolecular) dynamic contrast-enhanced MRI
.
Magn Reson Med
2016
;
75
:
1142
53
.
20.
Zhao
L
,
Ching
LM
,
Kestell
P
,
Kelland
LR
,
Baguley
BC
. 
Mechanisms of tumor vascular shutdown induced by 5,6-dimethylxanthenone-4-acetic acid (DMXAA): Increased tumor vascular permeability
.
Int J Cancer
2005
;
116
:
322
6
.
21.
Dowell
NG
,
Tofts
PS
. 
Fast, accurate, and precise mapping of the RF field in vivo using the 180 degrees signal null
.
Magn Reson Med
2007
;
58
:
622
30
.
22.
Schabel
MC
,
Fluckiger
JU
,
DiBella
EV
. 
A model-constrained Monte Carlo method for blind arterial input function estimation in dynamic contrast-enhanced MRI: I. Simulations
.
Phys Med Biol
2010
;
55
:
4783
806
.
23.
Schabel
MC
,
DiBella
EV
,
Jensen
RL
,
Salzman
KL
. 
A model-constrained Monte Carlo method for blind arterial input function estimation in dynamic contrast-enhanced MRI: II. In vivo results
.
Phys Med Biol
2010
;
55
:
4807
23
.
24.
Schabel
MC
. 
A unified impulse response model for DCE-MRI
.
Magn Reson Med
2012
;
68
:
1632
46
.
25.
Beaumont
M
,
Lemasson
B
,
Farion
R
,
Segebarth
C
,
Remy
C
,
Barbier
EL
. 
Characterization of tumor angiogenesis in rat brain using iron-based vessel size index MRI in combination with gadolinium-based dynamic contrast-enhanced MRI
.
J Cereb Blood Flow Metab
2009
;
29
:
1714
26
.
26.
Pike
MM
,
Stoops
CN
,
Langford
CP
,
Akella
NS
,
Nabors
LB
,
Gillespie
GY
. 
High-resolution longitudinal assessment of flow and permeability in mouse glioma vasculature: Sequential small molecule and SPIO dynamic contrast agent MRI
.
Magn Reson Med
2009
;
61
:
615
25
.
27.
Su
MY
,
Muhler
A
,
Lao
X
,
Nalcioglu
O
. 
Tumor characterization with dynamic contrast-enhanced MRI using MR contrast agents of various molecular weights
.
Magn Reson Med
1998
;
39
:
259
69
.
28.
Baguley
BC
. 
Antivascular therapy of cancer: DMXAA
.
Lancet Oncol
2003
;
4
:
141
8
.
29.
Hofmann
S
,
Grasberger
H
,
Jung
P
,
Bidlingmaier
M
,
Vlotides
J
,
Janssen
OE
, et al
The tumour necrosis factor-alpha induced vascular permeability is associated with a reduction of VE-cadherin expression
.
Eur J Med Res
2002
;
7
:
171
6
.
30.
Seshadri
M
,
Spernyak
JA
,
Mazurchuk
R
,
Camacho
SH
,
Oseroff
AR
,
Cheney
RT
, et al
Tumor vascular response to photodynamic therapy and the antivascular agent 5,6-dimethylxanthenone-4-acetic acid: implications for combination therapy
.
Clin Cancer Res
2005
;
11
:
4241
50
.
31.
Lara
PN
 Jr
,
Douillard
JY
,
Nakagawa
K
,
von Pawel
J
,
McKeage
MJ
,
Albert
I
, et al
Randomized phase III placebo-controlled trial of carboplatin and paclitaxel with or without the vascular disrupting agent vadimezan (ASA404) in advanced non-small-cell lung cancer
.
J Clin Oncol
2011
;
29
:
2965
71
.
32.
Albrecht
M
,
Henke
J
,
Tacke
S
,
Markert
M
,
Guth
B
. 
Influence of repeated anaesthesia on physiological parameters in male Wistar rats: a telemetric study about isoflurane, ketamine-xylazine and a combination of medetomidine, midazolam and fentanyl
.
BMC Vet Res
2014
;
10
:
310
.
33.
Jain
RK
. 
Normalization of tumor vasculature: an emerging concept in antiangiogenic therapy
.
Science
2005
;
307
:
58
62
.
34.
Ferrara
N
,
Adamis
AP
. 
Ten years of anti-vascular endothelial growth factor therapy
.
Nat Rev Drug Discov
2016
;
15
:
385
403
.
35.
Schabel
MC
,
Gilbert
E
,
Guimaraes
A
,
Wyatt
C
. 
Physiologically-constrained multiagent DCE-MRI for pancreatic cancer imaging
.
In:
Proceedings of the ISMRM 25th Annual Meeting; 2017 Apr 22–27
;
Honolulu, HI. Concord (CA)
:
ISMRM
; 
2017
.
Abstract nr 0317
.