Free
Multidisciplinary Ophthalmic Imaging  |   November 2012
Dynamic Contrast-Enhanced MRI for Assessing Therapeutic Response of Choroidal Neovascularization in a Rat Model
Author Affiliations & Notes
  • Jae-Hun Kim
    From the Departments of Radiology and
  • Geun Ho Im
    Center for Molecular and Cellular Imaging, Samsung Biomedical Research Institute, Seoul, Korea; the
    Department of Biomedical Engineering, Hanyang University, Seoul, Korea; and the
  • Jaemoon Yoon
    Opthalmology, Samsung Medical Center, Samsung Biomedical Research Institute, Sungkyunkwan University School of Medicine, Seoul, Korea; the
  • Jehoon Yang
    Center for Molecular and Cellular Imaging, Samsung Biomedical Research Institute, Seoul, Korea; the
  • Julius Juhyun Chung
    Samsung Advanced Institute for Health Sciences & Technology, Sungkyunkwan University, Seoul, Korea.
  • Ji Hoon Cha
    From the Departments of Radiology and
  • Sun I. Kim
    Department of Biomedical Engineering, Hanyang University, Seoul, Korea; and the
  • Don-il Ham
    Opthalmology, Samsung Medical Center, Samsung Biomedical Research Institute, Sungkyunkwan University School of Medicine, Seoul, Korea; the
  • Jung Hee Lee
    From the Departments of Radiology and
    Center for Molecular and Cellular Imaging, Samsung Biomedical Research Institute, Seoul, Korea; the
    Samsung Advanced Institute for Health Sciences & Technology, Sungkyunkwan University, Seoul, Korea.
  • *Each of the following is a corresponding author: Jung Hee Lee, Department of Radiology, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50, Ilwon-Dong, Gangnam-Gu, Seoul 135-710, Korea; hijunghee@skku.edu
  • Don-il Ham, Department of Ophthalmology, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50, Ilwon-Dong, Gangnam-Gu, Seoul 135-710, Korea; di.ham@samsung.com
Investigative Ophthalmology & Visual Science November 2012, Vol.53, 7693-7700. doi:10.1167/iovs.12-9805
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Jae-Hun Kim, Geun Ho Im, Jaemoon Yoon, Jehoon Yang, Julius Juhyun Chung, Ji Hoon Cha, Sun I. Kim, Don-il Ham, Jung Hee Lee; Dynamic Contrast-Enhanced MRI for Assessing Therapeutic Response of Choroidal Neovascularization in a Rat Model. Invest. Ophthalmol. Vis. Sci. 2012;53(12):7693-7700. doi: 10.1167/iovs.12-9805.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose.: We evaluated the potential of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) as a noninvasive biomarker of choroidal neovascularization (CNV) and its utility as a tool for monitoring therapeutic response in laser-induced rat CNV models.

Methods.: CNV was induced in the right eyes of 14 rats using a laser. Rats (n = 7) were treated daily for 14 days with a candidate drug (KR-31831, 50 mg/kg of body weight) having antiangiogenic effects, whereas control rats (n = 7) were treated with the vehicle alone (10% cremophor, 10% absolute ethyl alcohol, and 80% saline). DCE-MRI examinations were performed on the day before surgery (D − 1), and 3, 7, and 14 days after surgery (D + 3, D + 7, and D + 14), from which pharmacokinetic parameters (Ktrans , ve , vp ) were calculated. Angiography was performed to visualize CNV using FITC-labeled high molecular weight dextran after MRI on D + 14. The paired Wilcoxon test and Mann-Whitney U test were performed for statistical analysis.

Results.: The Ktrans and ve values of the CNV-induced right eyes were significantly higher than those of the intact eyes in control rats at D + 14 (P < 0.05). In the CNV-induced eyes, the relative Ktrans and ve values of the KR-31831–treated group were significantly lower than those of the nontreated group at D + 14 (P < 0.05). The angiography showed that decreased CNV was observed in rats treated with KR-31831.

Conclusions.: Quantitative DCE-MRI produces noninvasive biomarker of CNV, thus allowing monitoring of therapeutic response of antiangiogenic drugs in neovascular age-related macular degeneration (AMD).

Introduction
Age-related macular degeneration (AMD) is the most common cause of visual loss and blindness for elderly people. 1,2 On the basis of clinical appearance, the stages of AMD are categorized as early or late, with late AMD, in which severe visual loss is usual, being classified further as dry or wet. Dry AMD, known as geographic atrophy, is associated with the thinning of macular tissues. On the other hand, wet AMD, known as neovascular AMD, is associated with choroidal neovascularization (CNV), a pathologic process whereby abnormal new vessels grow from the choroidal capillaries through the Bruch's membrane separating the choroid and retina. 3  
Although, to our knowledge, there is no specific treatment for dry AMD to date, there have been significant advances in the treatment of wet AMD. Wet AMD can be treated by argon laser photocoagulation to destroy the extrafoveal neovascular membrane and by photodynamic therapy (PDT) with verteporfin molecules to occlude selectively active CNV. 46 CNV also can be treated with antiangiogenic drugs, such as ranibizumab (Lucentis) and bevacizumab (Avastin), which inhibit VEGF, known to have an important role in the development of angiogenesis. 79  
To assess treatment response for neovascular AMD in clinical studies, a visual acuity score of 15 letters or more, or analysis using the foveal thickness on optical coherence tomography and lesion size on fluorescent angiography were compared between pre- and post-treatment groups. 8,9 These assessment approaches are based on behavioral measurement and structural changes under the retina. 
Recent developments in magnetic resonance imaging (MRI) have focused on quantitative measurements of functional changes attributed to therapeutic effects. Among MRI techniques, dynamic contrast-enhanced (DCE) MRI can provide functional information about vascular endothelial proliferation, vascular density, and angiogenesis. 1012 In DCE-MRI data, the MR signal-time curves are measured after an intravenous injection of paramagnetic contrast agent that diffuses over time in the extravascular extracellular space (EES). For quantification of these signal-time curves, pharmacokinetic modeling methods have been developed by Brix et al., 13 Lasson et al., 14 and Tofts and Kermode 15 to estimate pharmacokinetic parameters, such as blood flow, fraction of plasma volume, the capillary permeability-surface area product, and the rate constant for the exchanges of contrast agent between plasma and EES. A set of standardized quantity names and symbols related to pharmacokinetic parameters of DCE-MRI have been described by Tofts et al. 16 Most studies have used a two-compartment model, which assumes: (1) the tissue is composed of two components, the EES (ve ) and vascular space (vp ), and (2) the transport of the contrast agent occurs between the EES and vascular space quantified by a volume transfer constant (Ktrans ). 
In our study, we hypothesized that DCE-MRI can provide a noninvasive biomarker of CNV, which can be used to monitor the therapeutic response of antiangiogenic drugs in neovascular AMD. The purpose of our study was to evaluate the usefulness of the DCE-MRI technique not only for assessing CNV, but also for monitoring therapeutic response of drugs in a laser-induced rat CNV model. 
Materials and Methods
Preparation of CNV Rat Model and Treatment
All animal work was performed with a license issued by the Institutional Animal Care and Use Committee (IACUC) and was performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The 14 Brown Norway rats (BN/Crlj, 9 weeks old, 200 g) were purchased from Charles River Laboratories (Yokohama, Japan) and housed in specific pathogen-free conditions. For surgical operation, rats were anesthetized with an intraperitoneal administration of Avertin (#T4, 840-2, 240 mg/kg; Sigma-Aldrich, St. Louis, MO). A mixture of 0.5% tropicamide and 0.5% phenylephrine hydrochloride was applied to the right eye to dilate the pupils 30 minutes before surgical laser operation. A lesion was induced by a diode pumped solid state laser (532 nm in wavelength, 100 ms in duration, 50 μm in size, 200 mW in power; Novus Varia; Lumenis Inc., Yokneam, Israel) around the disc of the retina through a slit-lamp delivery system. Only the laser spots in which a bubble was produced were considered effective and included in the study. For each rat, approximately 5 to 10 lesions were induced on the right eye, whereas the left eye was kept intact as a control. 
For treatment of CNV, a candidate drug KR-31831 was obtained from the Korea Research Institute of Chemical Technology and prepared as a suspension in a vehicle (10% cremophor [vol/vol], 10% absolute ethyl alcohol [vol/vol], and 80% normal saline) for intraperitoneal injection to rats. In the drug-treated group (n = 7), rats were treated daily with KR-31831 with a dose of 50 mg/kg of body weight for 14 days. In the control group (n = 7), rats were treated intraperitoneally not with the drug, but with the vehicle alone. DCE-MRI examinations were performed 1 day before the laser photocoagulation (D − 1), and 3, 7, and 14 days after the laser photocoagulation (D + 3, D + 7, and D + 14), where D0 is the day of laser photocoagulation. 
DCE-MRI Data Acquisition
All MR images were obtained using a horizontal 7.0 T MRI System (Bruker Biospin, Fällanden, Switzerland) equipped with a 20 cm gradient set capable of supplying up to 400 mT/m with a 100 μsec rise-time. A birdcage coil (72 mm i.d.; Bruker Biospin) was used for excitation, and an actively decoupled phased array coil was used for signal reception. The rats initially were anesthetized with isoflurane (5% for induction followed by 2% during set-up and imaging time) in a mixture of O2 and N2 gases (3:7), delivered to a nose cone for spontaneous respiration throughout the experiment. The head of the rat was fixed carefully using a bite/ear bar, and the rectal temperature was monitored and maintained at 36.5 + 0.5. For T1 mapping, MR images were acquired with a T1-weighted 2D FLASH sequence: matrix size = 128 × 128; resolution = 200 × 200 μm2; TR/TE = 60/3 ms; number of average = 1; slice thickness = 1.0 mm without interslice gaps; number of slices = approximately 5 to 7; variable flip angles = 5°, 15°, 35°, 60°, and 70° for each MR image. For DCE-MRI, baseline pre-injection coronal images were acquired with 70° flip angle during 2 minutes with a 6-second sampling interval (20 MR images). With an automatic injection of gadoterate meglumine (0.1 mmol of Gd3+/kg of body weight; Dotarem; Guerbet, Villepinte, France) over 5 seconds via a rat tail vein, postinjection dynamic images were acquired during 8 minutes (80 MR images). After acquisition of the DCE-MR images, high resolution coronal T2-weighted images were acquired to define retina areas anatomically with a 2D RARE sequence: matrix size = 256 × 256, resolution = 100 × 100 μm2, flip angle = 180°, TR/TE = 3000/30 ms, a slice thickness = 1.0 mm without interslice gaps. 
Automatic Segmentation of Retina
The retina in the rat was segmented automatically using the active contour method and morphologic imaging processing techniques. The active contour method, also known as the snake method, has been used for delineating object boundaries. 17 The basic idea behind the active contour method is evolving a curve to detect an object boundary toward minimizing a cost function that sums external and internal forces. In our method, the external force is minimized when the curve is at the object boundary position, and the internal force is minimized when the curve shape is as smooth as possible. First, the inner edge of the retina was detected using the active contour method on a high resolution T2-weighted MR image. In this step, we excluded the first and last slices to minimize the partial volume effect. The segmented inner voxels of the retina were down sampled to overlap onto low resolution T1-weighted MR images. Second, dilation image processing was performed to expand the inner edge of the retina using a disk mask with a one-voxel radius. Finally, the retina was segmented automatically by subtraction of the inner voxels of retina from the one-voxel dilated image. 
Quantification of DCE-MRI
The concentration of contrast agent was estimated by determining the difference in longitudinal relaxation rates: where T10 and T1 are the pre- and postcontrast T1 values, respectively, and r1 is the longitudinal relaxivity of the contrast agent (2.92 s−1mM−1, see report of Pedersen et al.18). The T10 values were estimated using the variable flip angle method,19 and then the T1 values were computed by where TR is the repetition time, and Spre and Spost are the MR signal intensities at pre- and postcontrast. 
The pharmacokinetic parameters then were computed using the two compartment model16: where Ct is the concentration of contrast agent in the area of interest, Cp is the concentration in the blood plasma, vp is the fractional blood plasma volume per unit volume of tissue, Ktrans is the volume transfer constant, and ve is the fractional EES per unit volume of tissue. Cp(t) was estimated using the concentration of contrast agent in blood vessels Cb(t) through the hematocrit: where the Hct is the hematocrit estimated in rats to be 0.44.20  
The arterial input function (AIF) was defined manually in the cerebral artery by an animal imaging expert (GHI) for each rat at the baseline. In our study, the AIF (Cb[t]) was determined by averaging the AIFs determined for 14 rats at D − 1; this averaged AIF then was used for calculating the pharmacokinetic parameters (Ktrans , ve , vp ) for all rats. The relative pharmacokinetic parameters also were calculated by the ratios of the pharmacokinetic parameters in the right to those in the left eye. These could minimize interindividual variations that may be attributed to cardiovascular changes, such as blood pressure, blood volume, and heart rate. Nonlinear fitting was implemented using a nonlinear least-squares fitting algorithm from the Matlab Optimization Toolbox (Mathworks Inc., Natick, MA). 
Statistical Analysis
All pharmacokinetic parameters were reported as means ± SD. The paired Wilcoxon test was performed to assess the differences in pharmacokinetic parameters between the CNV-induced (right) and intact (left) eyes in control rats. The Mann-Whitney U test was performed to test the differences in the relative pharmacokinetic parameters between the KR-31831–treated and nontreated groups in CNV-induced eyes. P < 0.05 was considered to be statistically significant. 
Angiography Using FITC-Labeled Dextran
Rats were anesthetized and perfused through the heart with 1 mL PBS containing 50 mg/mL fluorescein-labeled dextran (2 × 106 average mW; Sigma-Aldrich). One or 2 minutes later, the rats were sacrificed with 100% CO2, and the CNV-induced eyes were removed and fixed in 10% buffered formalin for 1 hour. They then were sectioned at the equator and the anterior half, vitreous, and retina were removed. The retinal pigment epithelium–choroid–sclera complex was dissected through four to five relaxing radial incisions and flat mounted on a slide for viewing under microscope. Angiography was performed on D + 14 for control and KR-31831–treated rats. 
Results
Figure 1 shows color fundus photographs of the right and left eyes in a rat at D + 3. CNV was induced in 6 spots in the right eye. Figure 2 shows the procedures of automatic segmentation of the retina in a rat using the active contour method and morphologic image processing technique. The active contour method was performed on a high resolution T2-weighted MR image with an initial mask (Fig. 2a). The initial curve (box shape) was evolved iteratively to detect the inner edge of retina (Fig. 2b). Finally, the inner edge of the retina was detected after 500 evolutions of the initial curve (Fig. 2c). The segmented image on a high resolution T2-weighted image (voxels within the green line in Fig. 2c) was interpolated linearly to overlap low resolution T1-weighted images (Fig. 2d). We expanded one-voxel outside of this overlapped image using dilation image processing (Fig. 2e). Finally, the retina was segmented automatically by image subtraction technique (voxels between green and yellow line in Fig. 2f). A quarter of the segmented retina was defined as a region of interest (ROI) for each eye. The mean concentration profile was created by averaging the concentration of contrast agent as a function of time across all voxels within the segmented retina for each eye. It was higher for the right retina than for the left retina in the control group, whereas these were similar between the left and right retinas in the KR-31831–treated group at D + 7 and D + 14 (Fig. 3). For quantification of these concentration profiles, the AIF was defined by drawing the arterial vessel for each rat manually (Figs. 4a, 4b). Nonlinear fitting was performed on the mean concentrations from the segmented retina using the population AIF to compute Ktrans , ve , and vp (Figs. 4c, 4d). The Table summarizes the mean and SD of Ktrans , ve , and vp parameters in all rats. The Ktrans and ve values of the CNV-induced right eyes were significantly higher than those of the intact eyes in control rats at D + 14 (P < 0.05), and the relative Ktrans and ve values of the KR-31831–treated eyes were significantly lower than those of nontreated eyes at D + 14 (P < 0.05). There was no significant change in the vp values at any time points (Fig. 5). Figure 6 shows FITC-dextran angiography for representative control and KR-31831–treated rats at D + 14. CNV was well visualized by fluorescein-labeled high molecular-weight dextran (Figs. 6a, 6b). The angiography showed that CNV was reduced observably in the rat treated with antiangiogenic drug compared to those of nontreated rats. 
Figure 1. 
 
Fundus photography for right (a) and left (b) eyes from a rat with laser-induced choroidal neovascularization at 3 days after surgical laser operation. Red arrows represent the laser-induced choroidal neovascularization.
Figure 1. 
 
Fundus photography for right (a) and left (b) eyes from a rat with laser-induced choroidal neovascularization at 3 days after surgical laser operation. Red arrows represent the laser-induced choroidal neovascularization.
Figure 2. 
 
Automatic segmentation of the retina in a rat using the active contour method and morphologic image processing techniques. (a) The active contour method was performed on T2-weighted MR images with an initial box boundary. The initial curve was evolved iteratively to define the inner edge of the retina at 40 (b) and 500 (c) iterations. (d) The segmented image in (c) was interpolated linearly to overlap onto T1-weighted MR images. (e) Dilation image processing was performed to expand the inner edge of the retina (yellow boundary). (f) Subtraction image processing was performed to segment the retina voxels between the green and yellow boundaries.
Figure 2. 
 
Automatic segmentation of the retina in a rat using the active contour method and morphologic image processing techniques. (a) The active contour method was performed on T2-weighted MR images with an initial box boundary. The initial curve was evolved iteratively to define the inner edge of the retina at 40 (b) and 500 (c) iterations. (d) The segmented image in (c) was interpolated linearly to overlap onto T1-weighted MR images. (e) Dilation image processing was performed to expand the inner edge of the retina (yellow boundary). (f) Subtraction image processing was performed to segment the retina voxels between the green and yellow boundaries.
Figure 3. 
 
Concentration profiles in the retina region from a representative control (a) and KR-31831–treated (b) rats at D − 1, D + 3, D + 7, and D + 14.
Figure 3. 
 
Concentration profiles in the retina region from a representative control (a) and KR-31831–treated (b) rats at D − 1, D + 3, D + 7, and D + 14.
Figure 4. 
 
Estimating pharmacokinetic parameters using the AIF. (a) Manual-defined ROI for the arterial vessel in the cerebral artery. (b) Concentration profile for the arterial input function. The blue point represents the concentration profile from each rat and the red dot-line represents the population AIF. (c) Automatic segmentation of retina in the right eye. (d) Concentration profile from the segmented retina (c). The red line represents the fitted data using a nonlinear fitting method.
Figure 4. 
 
Estimating pharmacokinetic parameters using the AIF. (a) Manual-defined ROI for the arterial vessel in the cerebral artery. (b) Concentration profile for the arterial input function. The blue point represents the concentration profile from each rat and the red dot-line represents the population AIF. (c) Automatic segmentation of retina in the right eye. (d) Concentration profile from the segmented retina (c). The red line represents the fitted data using a nonlinear fitting method.
Figure 5. 
 
The mean relative Ktrans (a), ve (b), and vp (c) parameters for the drug-treated and control groups. The red bar represents the mean of relative parameters for the KR-31831–treated group, and the white bar represents the mean of relative parameters for the control group. The error bar represents the SDs of parameters for each time point. *P < 0.05.
Figure 5. 
 
The mean relative Ktrans (a), ve (b), and vp (c) parameters for the drug-treated and control groups. The red bar represents the mean of relative parameters for the KR-31831–treated group, and the white bar represents the mean of relative parameters for the control group. The error bar represents the SDs of parameters for each time point. *P < 0.05.
Figure 6. 
 
Images of laser-induced choroidal neovascularization on D + 14 for microscopic evaluation. FITC-dextran angiography for a control rat (a) and a drug-treated rat (b). Red arrows represent the laser-induced choroidal neovascularization.
Figure 6. 
 
Images of laser-induced choroidal neovascularization on D + 14 for microscopic evaluation. FITC-dextran angiography for a control rat (a) and a drug-treated rat (b). Red arrows represent the laser-induced choroidal neovascularization.
Table. 
 
Mean and SD of the Ktrans , ve , and vp Parameters for Control and KR-31831 Treated Group
Table. 
 
Mean and SD of the Ktrans , ve , and vp Parameters for Control and KR-31831 Treated Group
Ktrans (min−1) ve vp
L R L R L R
Control group
 D + 0 0.0158 (0.0056) 0.0138 (0.0054) 0.5252 (0.2053) 0.4740 (0.1717) 0.1412 (0.0264) 0.1559 (0.0249)
 D + 3 0.0163 (0.0066) 0.0195 (0.0097) 0.5848 (0.1653) 0.6548 (0.1504) 0.1317 (0.0846) 0.1416 (0.0817)
 D + 7 0.0172 (0.0074) 0.0196 (0.0108) 0.6340 (0.2252) 0.6273 (0.1634) 0.1209 (0.0937) 0.1158 (0.0737)
 D + 14 0.0115 (0.0033) 0.0184 (0.0063)* 0.4769 (0.1992) 0.6464 (0.1479)* 0.1288 (0.0904) 0.1117 (0.0875)
Treated group
 D + 0 0.0152 (0.0052) 0.0116 (0.0014) 0.5359 (0.2100) 0.5277 (0.2272) 0.1649 (0.0563) 0.1797 (0.0558)
 D + 3 0.0120 (0.0027) 0.0144 (0.0034) 0.5322 (0.2123) 0.5756 (0.1009) 0.1127 (0.0496) 0.1037 (0.0445)
 D + 7 0.0145 (0.0070) 0.0154 (0.0086) 0.4939 (0.1954) 0.5364 (0.1489) 0.0953 (0.0781) 0.1002 (0.0688)
 D + 14 0.0128 (0.0030) 0.0142 (0.0059) 0.4863 (0.1081) 0.5489 (0.1142) 0.1037 (0.0781) 0.1007 (0.0640)
Discussion
In our study, the use of DCE-MRI to assess choroidal neovascularization and to monitor therapeutic response of antiangiogenic drug effect in experimental CNV rats was evaluated. First, we demonstrated that DCE-MRI provides noninvasive biomarkers to assess abnormal vessel growth under the retina, showing higher Ktrans and ve parameters in CNV-induced eyes than in intact eyes. Second, our data showed support that DCE-MRI could be a useful tool for the evaluation of antiangiogenic drug treatment efficacy in CNV rats, showing lower relative Ktrans and ve parameters in treated rats than untreated rats after treatment. 
Our results assessing laser-induced CNV using DCE-MRI showed that leakage of CNV is prominent at D + 14 in rat CNV models. In mouse CNV models, CNV formation reached its maximum size at D + 5, and there was significant size reduction by D + 7 using spectral domain optical coherence tomography. 21 However, progression of laser-induced CNV is different between mice and rats. In rat CNV models, CNV vessels are visualized by D + 4, and the volume of CNV vessels increases exponentially during the next 3 days using FITC-labeled dextran. By D + 7, a well-defined CNV network is established, the volume of which is preserved for several weeks. 22 The incidence of CNV in rats using FITC-labeled dextran reported that no vessels are found at D + 1, 47% of lesions at D + 3, 71% at D + 6, and 100% at D + 10. 23 From these previous studies, we found that well-defined and prominent CNV in rats can be observed from 10 days after laser exposure. Thus, our quantitative results of DCE-MRI are consistent with the previous results in laser-induced rat CNV models. 
Berkowitz et al. tested the use of DCE-MRI for in vivo measurement of passive blood–retinal barrier permeability in experimental diabetic rats. 24 They showed that the changes in vitreous MRI signal intensity on T1-weighted images are a marker of blood–retinal barrier disruption. They suggests the possibility that DCE-MRI could be a useful tool for evaluating drug treatment efficacy after VPF/VEGF treatment in experimental diabetic rats and this remains a possibility for future work. In the current study, we confirmed the power of DCE-MRI to evaluate drug treatment efficacy after antiangiogenic drug treatment in CNV rats. 
The conventional methodology for evaluating treatment effects of CNV is based on behavioral measurement (i.e., visual acuity score test) and morphologic change (i.e., the change of foveal thickness measured by optical coherence tomography, and the change of lesion size measured by fluorescent angiography). 8,9 These assessment approaches, however, may be subjective and take months to observe any structural/morphologic change after treatment. On the other hand, quantitative DCE-MRI is based on the functional change of microvasculature and provides parameters with increased objectivity. With functional information of the microvascular environment, quantitative DCE-MRI can provide complementary information to conventional assessment methods to improve the speed and accuracy of evaluating abnormal new vessels under the retina. 
For quantification of DCE-MRI data, the estimation of the AIF is required to compute pharmacokinetic parameters (Ktrans , ve , and vp ). However, as mentioned in our previous report, 25 it often is difficult to determine AIF voxels that may be distant from large vessels in small animals. In our study, we failed to define the AIF automatically in a few rats due to a problem with the small size of arterial vessels near the retina (approximately 2–5 voxels size) along with the low spatial resolution of MRI. In addition, we found that there was a large intertrial variability (D − 1, D + 3, D + 7, and D + 14) in the AIFs due to the partial volume effect even when AIF voxels were selected carefully by visual inspection of voxels' concentration profiles in some rats. Therefore, in our study, we used a mean AIF that was derived by averaging AIFs from 14 rats measured at D − 1. The mean AIF then was used for computing the pharmacokinetic parameters (Ktrans , ve , and vp ) for all rats at D − 1, D + 3, D + 7, and D + 14. 
KR-31831 is a small molecular weight drug, newly developed as an antiangiogenic candidate drug, which is known to suppress endothelial cell proliferation, tube formation, invasion, and migration in vitro, and inhibit vessel formation in vivo. 2629 In our previous study, we evaluated KR-31831 as an antiangiogenic drug using DCE-MR image in a subcutaneous mouse model bearing a human SKOV3 ovarian carcinoma cell line. 30 Previous work has demonstrated the antiangiogenic effects of KR-31831 on rat CNV models using FITC-dextran angiography. 31 In our study, we confirmed the antiangiogenic effects of KR-31831 on rat CNV model using the DCE-MRI tool and demonstrated the use of the DCE-MRI technique as a noninvasive tool to measure angiogenesis of the retina and to monitor antiangiogenic drug effects for neovascular disease in rat CNV models. 
There is a possibility that the signal from the retina was contaminated by the signal of the adjacent tissue area due to limited spatial resolution (200 × 200 μm2). To minimize partial volume effects, we defined the region of retina by excluding the first and last slices of DCE-MR images, and drawing the ROIs as small as possible (boundaries with one voxel-width). To confirm our results, we performed an additional analysis in the muscle around the eye. In this analysis, there were no significant differences in the relative Ktrans , ve , and vp parameters between KR-31831–treated and control rats (see Supplementary Material and Supplementary Fig. S1). 
Previous studies have demonstrated the Gd-DTPA leakage of the vitreous region using DCE-MRI indicating blood–retinal barrier breakdown. 32,33 In our study, however, we did not find any MR signal enhancement in the vitreous region in CNV rats. To clarify this discrepancy, we performed additional experiments on D − 1, D + 3, D + 7, and D + 14 with a higher image resolution (100 × 100 μm2), higher signal-to-noise ratio (SNR, number of average = 5) and longer imaging acquisition time (65 minutes of DCE-MRI acquisition time with a 1-minute sampling interval, 65 MR images). In these high SNR images, we found that there was leakage of Gd-DTPA into the vitreous region in CNV rats. To observe the leakage in the vitreous region, acquisition of high SNR images seems to be mandatory, which requires long temporal sampling times scarifying quantification. For quantification of DCE-MRI data using the AIF in our study, we acquired DCE-MRI data with a short temporal sampling time (6 seconds) resulting in lower SNR images (1 number of average), in which we observed the MR signal enhancements in the retina and muscle around the eye. 
In conclusion, we demonstrated the use of DCE-MRI to assess choroidal neovascularization and to monitor therapeutic response to antiangiogenic drugs in experimental CNV rats. In the future, DCE-MRI may open up the paths to personalized medicine (i.e., personalization of therapeutic method, clinical drug choice, and drug dose adjustment) in patients with neovascular AMD. 
Supplementary Materials
References
Bressler NM. Age-related macular degeneration is the leading cause of blindness. JAMA . 2004;291:1900–1901. [CrossRef] [PubMed]
Ferris FL III Fine SL Hyman L. Age-related macular degeneration and blindness due to neovascular maculopathy. Arch Ophthalmol . 1984;102:1640–1642. [CrossRef] [PubMed]
de Jong PT. Age-related macular degeneration. N Engl J Med . 2006;355:1474–1485. [CrossRef] [PubMed]
Haddad WM Coscas G Soubrane G. Eligibility for treatment and angiographic features at the early stage of exudative age related macular degeneration. Br J Ophthalmol . 2002;86:663–669. [CrossRef] [PubMed]
Argon laser photocoagulation for neovascular maculopathy. Five-year results from randomized clinical trials. Macular Photocoagulation Study Group. Arch Ophthalmol . 1991;109:1109–1114. [CrossRef] [PubMed]
Harding S. Photodynamic therapy in the treatment of subfoveal choroidal neovascularisation. Eye (Lond) . 2001;15:407–412. [CrossRef] [PubMed]
Gragoudas ES Adamis AP Cunningham ET Jr Feinsod M Guyer DR. Pegaptanib for neovascular age-related macular degeneration. N Engl J Med . 2004;351:2805–2816. [CrossRef] [PubMed]
Chakravarthy U. Bevacizumab for the treatment of neovascular age related macular degeneration. BMJ . 2010;340:c2834.
Martin DF Maguire MG Ying GS Grunwald JE Fine SL Jaffe GJ. Ranibizumab and bevacizumab for neovascular age-related macular degeneration. N Engl J Med . 2011;364:1897–1908. [CrossRef] [PubMed]
Hillman GG Singh-Gupta V Zhang H Dynamic contrast-enhanced magnetic resonance imaging of vascular changes induced by sunitinib in papillary renal cell carcinoma xenograft tumors. Neoplasia . 2009;11:910–920. [CrossRef] [PubMed]
Garbow JR Santeford AC Anderson JR Engelbach JA Arbeit JM. Magnetic resonance imaging defines cervicovaginal anatomy, cancer, and VEGF trap antiangiogenic efficacy in estrogen-treated K14-HPV16 transgenic mice. Cancer Res . 2009;69:7945–7952. [CrossRef] [PubMed]
Raatschen HJ Simon GH Fu Y Vascular permeability during antiangiogenesis treatment: MR imaging assay results as biomarker for subsequent tumor growth in rats. Radiology . 2008;247:391–399. [CrossRef] [PubMed]
Brix G Semmler W Port R Schad LR Layer G Lorenz WJ. Pharmacokinetic parameters in CNS Gd-DTPA enhanced MR imaging. J Comput Assist Tomogr . 1991;15:621–628. [CrossRef] [PubMed]
Larsson HB Stubgaard M Frederiksen JL Jensen M Henriksen O Paulson OB. Quantitation of blood-brain barrier defect by magnetic resonance imaging and gadolinium-DTPA in patients with multiple sclerosis and brain tumors. Magn Reson Med . 1990;16:117–131. [CrossRef] [PubMed]
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;17:357–367. [CrossRef] [PubMed]
Tofts PS Brix G Buckley DL Estimating kinetic parameters from dynamic contrast-enhanced T(1)-weighted MRI of a diffusable tracer: standardized quantities and symbols. J Magn Reson Imaging . 1999;10:223–232. [CrossRef] [PubMed]
Chan TF Vese LA. Active contours without edges. IEEE Trans Image Proc . 2001;10:266–277. [CrossRef]
Pedersen M Morkenborg J Jensen FT Stodkilde-Jorgensen H Djurhuus JC Frokiaer J. In vivo measurements of relaxivities in the rat kidney cortex. J Magn Reson Imaging . 2000;12:289–296. [CrossRef] [PubMed]
Fram EK Herfkens RJ Johnson GA Rapid calculation of T1 using variable flip angle gradient refocused imaging. Magn Reson Imaging . 1987;5:201–208. [CrossRef] [PubMed]
Matsuzawa T Morita N Hayashi Y A survey of the results of haematological parameters, using a common rat blood sample in Japanese laboratories. Comp Haematol Int . 1996;6:125–133. [CrossRef]
Giani A Thanos A Roh MI In vivo evaluation of laser-induced choroidal neovascularization using spectral-domain optical coherence tomography. Invest Ophthalmol Vis Sci . 2011;52:3880–3887. [CrossRef] [PubMed]
Campos M Amaral J Becerra SP Fariss RN. A novel imaging technique for experimental choroidal neovascularization. Invest Ophthalmol Vis Sci . 2006;47:5163–5170. [CrossRef] [PubMed]
Edelman JL Castro MR. Quantitative image analysis of laser-induced choroidal neovascularization in rat. Exp Eye Res . 2000;71:523–533. [CrossRef] [PubMed]
Berkowitz BA Roberts R Luan H Peysakhov J Mao X Thomas KA. Dynamic contrast-enhanced MRI measurements of passive permeability through blood retinal barrier in diabetic rats. Invest Ophthalmol Vis Sci . 2004;45:2391–2398. [CrossRef] [PubMed]
Kim JH Im GH Yang J Choi D Lee WJ Lee JH. Quantitative dynamic contrast-enhanced MRI for mouse models using automatic detection of the arterial input function. NMR Biomed . 2012;25:674–684. [CrossRef] [PubMed]
Park SY Seo EH Song HS KR-31831, benzopyran derivative, inhibits VEGF-induced angiogenesis of HUVECs through suppressing KDR expression. Int J Oncol . 2008;32:1311–1315. [PubMed]
Yi EY Park SY Song HS KR-31831, a new synthetic anti-ischemic agent, inhibits in vivo and in vitro angiogenesis. Exp Mol Med . 2006;38:502–508. [CrossRef] [PubMed]
Kim HH Paek IB Ji HY Lee S Yi KY Lee HS. Metabolism of a novel antiangiogenic agent KR-31831 in rats using liquid chromatography-electrospray mass spectrometry. J Sep Sci . 2005;28:1818–1822. [CrossRef] [PubMed]
Kim SJ Lee HI Ji HY Pharmacokinetics of a novel antiangiogenic agent KR-31831 in rats. Biopharm Drug Dispos . 2005;26:21–26. [CrossRef] [PubMed]
Yang J Kim JH Im GH Evaluation of antiangiogenic effects of a new synthetic candidate drug KR-31831 on xenografted ovarian carcinoma using dynamic contrast enhanced MRI. Korean J Radiol . 2011;12:602–610. [CrossRef] [PubMed]
Kim IT Park HY Choi JS Joo CK. Antiangiogenic effect of KR-31831 on corneal and choroidal neovascularization in rat models. Invest Ophthalmol Vis Sci . 2012;53:3111–3119. [CrossRef] [PubMed]
Sato Y Berkowitz BA Wilson CA de Juan E Jr. Blood-retinal barrier breakdown caused by diode vs argon laser endophotocoagulation. Arch Ophthalmol . 1992;110:277–281. [CrossRef] [PubMed]
Berkowitz BA Sato Y Wilson CA de Juan E. Blood-retinal barrier breakdown investigated by real-time magnetic resonance imaging after gadolinium-diethylenetriaminepentaacetic acid injection. Invest Ophthalmol Vis Sci . 1991;32:2854–2860. [PubMed]
Footnotes
 Supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science, and Technology (2011-0031520 and 2010-0023606).
Footnotes
 Disclosure: J.-H. Kim, None; G.H. Im, None; J. Yoon, None; J. Yang, None; J.J. Chung, None; J.H. Cha, None; S.I. Kim, None; D. Ham, None; J.H. Lee, None
Figure 1. 
 
Fundus photography for right (a) and left (b) eyes from a rat with laser-induced choroidal neovascularization at 3 days after surgical laser operation. Red arrows represent the laser-induced choroidal neovascularization.
Figure 1. 
 
Fundus photography for right (a) and left (b) eyes from a rat with laser-induced choroidal neovascularization at 3 days after surgical laser operation. Red arrows represent the laser-induced choroidal neovascularization.
Figure 2. 
 
Automatic segmentation of the retina in a rat using the active contour method and morphologic image processing techniques. (a) The active contour method was performed on T2-weighted MR images with an initial box boundary. The initial curve was evolved iteratively to define the inner edge of the retina at 40 (b) and 500 (c) iterations. (d) The segmented image in (c) was interpolated linearly to overlap onto T1-weighted MR images. (e) Dilation image processing was performed to expand the inner edge of the retina (yellow boundary). (f) Subtraction image processing was performed to segment the retina voxels between the green and yellow boundaries.
Figure 2. 
 
Automatic segmentation of the retina in a rat using the active contour method and morphologic image processing techniques. (a) The active contour method was performed on T2-weighted MR images with an initial box boundary. The initial curve was evolved iteratively to define the inner edge of the retina at 40 (b) and 500 (c) iterations. (d) The segmented image in (c) was interpolated linearly to overlap onto T1-weighted MR images. (e) Dilation image processing was performed to expand the inner edge of the retina (yellow boundary). (f) Subtraction image processing was performed to segment the retina voxels between the green and yellow boundaries.
Figure 3. 
 
Concentration profiles in the retina region from a representative control (a) and KR-31831–treated (b) rats at D − 1, D + 3, D + 7, and D + 14.
Figure 3. 
 
Concentration profiles in the retina region from a representative control (a) and KR-31831–treated (b) rats at D − 1, D + 3, D + 7, and D + 14.
Figure 4. 
 
Estimating pharmacokinetic parameters using the AIF. (a) Manual-defined ROI for the arterial vessel in the cerebral artery. (b) Concentration profile for the arterial input function. The blue point represents the concentration profile from each rat and the red dot-line represents the population AIF. (c) Automatic segmentation of retina in the right eye. (d) Concentration profile from the segmented retina (c). The red line represents the fitted data using a nonlinear fitting method.
Figure 4. 
 
Estimating pharmacokinetic parameters using the AIF. (a) Manual-defined ROI for the arterial vessel in the cerebral artery. (b) Concentration profile for the arterial input function. The blue point represents the concentration profile from each rat and the red dot-line represents the population AIF. (c) Automatic segmentation of retina in the right eye. (d) Concentration profile from the segmented retina (c). The red line represents the fitted data using a nonlinear fitting method.
Figure 5. 
 
The mean relative Ktrans (a), ve (b), and vp (c) parameters for the drug-treated and control groups. The red bar represents the mean of relative parameters for the KR-31831–treated group, and the white bar represents the mean of relative parameters for the control group. The error bar represents the SDs of parameters for each time point. *P < 0.05.
Figure 5. 
 
The mean relative Ktrans (a), ve (b), and vp (c) parameters for the drug-treated and control groups. The red bar represents the mean of relative parameters for the KR-31831–treated group, and the white bar represents the mean of relative parameters for the control group. The error bar represents the SDs of parameters for each time point. *P < 0.05.
Figure 6. 
 
Images of laser-induced choroidal neovascularization on D + 14 for microscopic evaluation. FITC-dextran angiography for a control rat (a) and a drug-treated rat (b). Red arrows represent the laser-induced choroidal neovascularization.
Figure 6. 
 
Images of laser-induced choroidal neovascularization on D + 14 for microscopic evaluation. FITC-dextran angiography for a control rat (a) and a drug-treated rat (b). Red arrows represent the laser-induced choroidal neovascularization.
Table. 
 
Mean and SD of the Ktrans , ve , and vp Parameters for Control and KR-31831 Treated Group
Table. 
 
Mean and SD of the Ktrans , ve , and vp Parameters for Control and KR-31831 Treated Group
Ktrans (min−1) ve vp
L R L R L R
Control group
 D + 0 0.0158 (0.0056) 0.0138 (0.0054) 0.5252 (0.2053) 0.4740 (0.1717) 0.1412 (0.0264) 0.1559 (0.0249)
 D + 3 0.0163 (0.0066) 0.0195 (0.0097) 0.5848 (0.1653) 0.6548 (0.1504) 0.1317 (0.0846) 0.1416 (0.0817)
 D + 7 0.0172 (0.0074) 0.0196 (0.0108) 0.6340 (0.2252) 0.6273 (0.1634) 0.1209 (0.0937) 0.1158 (0.0737)
 D + 14 0.0115 (0.0033) 0.0184 (0.0063)* 0.4769 (0.1992) 0.6464 (0.1479)* 0.1288 (0.0904) 0.1117 (0.0875)
Treated group
 D + 0 0.0152 (0.0052) 0.0116 (0.0014) 0.5359 (0.2100) 0.5277 (0.2272) 0.1649 (0.0563) 0.1797 (0.0558)
 D + 3 0.0120 (0.0027) 0.0144 (0.0034) 0.5322 (0.2123) 0.5756 (0.1009) 0.1127 (0.0496) 0.1037 (0.0445)
 D + 7 0.0145 (0.0070) 0.0154 (0.0086) 0.4939 (0.1954) 0.5364 (0.1489) 0.0953 (0.0781) 0.1002 (0.0688)
 D + 14 0.0128 (0.0030) 0.0142 (0.0059) 0.4863 (0.1081) 0.5489 (0.1142) 0.1037 (0.0781) 0.1007 (0.0640)
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×