Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Cancer Res. Author manuscript; available in PMC 2010 October 15.
Published in final edited form as:
PMCID: PMC2782788

Magnetic resonance imaging defines cervico-vaginal anatomy, cancer, and VEGF Trap antiangiogenic efficacy in estrogen-treated K14-HPV16 transgenic mice


Non-invasive detection of dysplasia provides a potential platform for monitoring the efficacy of chemopreventive therapy of premalignancy, imaging the tissue compartments comprising dysplasia: epithelium, microvasculature, and stromal inflammatory cells. Here, using respiratory-gated magnetic resonance imaging (MRI), the anatomy of premalignant and malignant stages of cervical carcinogenesis in estrogen-treated K14-HPV16 transgenic mice was noninvasively defined. Dynamic contrast enhanced (DCE) MRI was used to quantify leakage across premalignant dysplastic microvasculature. Vascular permeability as measured by DCE-MRI, Ktrans, was similar in transgenic (0.053 ± 0.020 min−1, n=32 mice) and nontransgenic (0.056 ± 0.029 min−1, n=17 mice) animals, despite a two-fold increase in microvascular area in the former compared with the latter. DCE-MRI did detect a significant decrease in vascular permeability accompanying diminution of dysplastic microvasculature by the anti-angiogenic agent, VEGF Trap (Ktrans =0.052 ± 0.013 min−1 pre-treatment; n=6 mice, vs. 0.019 +/− 0.008 min−1 post-treatment; n = 5 mice). Thus, we determined that the threshold of microvessel leakage associated with cervical dysplasia was below 17 kDa, and highlighted the potential of DCE-MRI to non-invasively monitor the efficacy of anti-angiogenic drugs or chemoprevention regimens targeting the vasculature, in premalignant cervical dysplasia.

Keywords: DCE-MRI, cervix, K14-HPV16, angiogenesis, dysplasia


Cervical cancer is the second most common malignancy affecting woman world-wide (1). While effective screening has markedly diminished cervical cancer incidence in the United States, dysplasia remains a common clinical challenge requiring surgical extirpation for high-grade lesions (1, 2). As such, noninvasive therapy for premalignant, but high risk, dysplasias would be a tremendous boon to gynecological care.

Dysplastic lesions are characterized by disorganized epithelial differentiation. The uterine cervical squamous epithelium can harbor dysplasia initiated by oncogenic human papilloma viruses (HPV), most frequently types 16 and 18 (1). Normal cervical epithelium is composed of a basal layer that is immature and proliferative, and multiple suprabasal layers of differentiating squamous cells that are ultimately shed from the top of the epithelium into the cervical canal or vaginal lumen. Increasing grades of dysplasia are characterized by progressive occupation of the suprabasal layer by basaloid squamous cells (1). High grade dysplasia, if untreated, has an extremely high incidence of conversion to malignancy (2). Cervical dysplasia and dysplastic lesions in general, activate an angiogenic switch that both increases the subepithelial microvasculature and produces stromal inflammation (3, 4). Angiogenesis and inflammation are features that could be exploited for imaging by a technique measuring tissue vascularity and leakage, such as dynamic contrast enhanced (DCE) magnetic resonance imaging (MRI). DCE MRI, studied extensively in solid malignancies (57), uses bolus contrast agent administration to non-invasively visualize tumor microvasculature and quantify leakage across microvessels (8, 9). Time-resolved images, collected as contrast agent enters and exits tissue (10) are pharmacokinetically modeled (11, 12), providing quantitative measures of important physiologic parameters, such as tissue microvascular integrity (vascular permeability).

Here, we used respiratory-gated MRI to obtain high resolution images of the entire mouse reproductive tract and delineated an invasive squamous carcinoma in a late-stage (1315) estrogen-treated K14-HPV16 transgenic mouse. We also interrogated the microvascular biology associated with premalignant dysplasia using DCE-MRI with a 17 kDa contrast agent, Gadomer™. No differential dysplastic microvascular leak was detectable, consistent with experiments showing undetectable physical leakage of an intravenously injected marker protein, suggesting that dysplastic microvessels have a maximal permeability pore size below 17 kDa. However, DCE-MRI detected a significant decrease in dysplastic vascular leak following treatment with the anti-angiogenesis agent, VEGF Trap (16, 17). Thus, DCE-MRI can serve as a platform for non-invasively monitoring anti-angiogenic drug efficacy, potentially deployed with chemoprevention programs, for high-risk pre-malignant disease.

Materials and Methods

Transgenic Mice

One-month-old female K14-HPV16:FVB/n congenic transgenic mice and nontransgenic littermate controls were treated for 3 or 6 months with subcutaneous insertion of 17β-estradiol pellets, 0.05 mg/60 day release (Innovative Research of America, Sarasota, FL) (`estrogen') (13, 15). The Animal Studies Committee of Washington University in Saint Louis approved all procedures in this study.


MR images were collected on a Varian NMR Systems (Palo Alto, CA) 4.7T INOVA scanner described previously (18) using Stark Contrast (Erlangen, Germany) 2.0 or 2.5 cm birdcage RF coils. Mice were anesthetized and maintained on isoflurane/O2 (1–1.25 % v/v), and a 3” length of PE-10 tubing (Becton Dickinson and Company, Sparks, MD) was inserted via the urethra into the bladder for drainage throughout data collection. Core body temperature was maintained at 37 ±1 °C by warm air circulation through the magnet bore.

For anatomic imaging, 500 μl of Omniscan™ (gadodiamide, GE Healthcare, Princeton, NJ) contrast agent, diluted 1:10 in saline to yield a 50 mM solution, was administered i.p. immediately before placing the animal into the magnet. High-resolution images were collected using a respiratory-gated, 2-D, multi-slice spin-echo sequence (18), and all images were collected during post-expiratory periods. Imaging parameters were TR ~3 s, TE = 20 ms, FOV 2.5 × 2.5 cm2 (transaxial) / 4.0 × 4.0 cm2 (coronal), 128 × 128 data points, slice thickness = 0.5 mm.

Dynamic contrast enhanced (DCE) MRI data were collected using a T1-weighted, gradient spoiled, multi-slice gradient-echo sequence. Imaging parameters were: flip angle, α=30°; repetition time, TR = 0.06 s; echo time, TE = 0.002 s; field of view, FOV = 2.5 × 2.5 cm2; number of slices, ns = 15; in-plane resolution = 195 μm; slice thickness = 0.50 mm. The temporal resolution was approximately 15 s. Early DCE-MRI data were collected following i.v. injection of Omniscan. However, values of vascular permeability (Ktrans) were all derived from experiments using Gadomer (Bayer Schering Pharma AG; Berlin, Germany). Pre-contrast, T1 maps were produced using a variable flip-angle, 3-D gradient-echo sequence (19, 20) with flip angles of 2.5°, 5.0°, 7.5°, 10.0° and 15.0°. After one minute of scanning (4 images at 15 s/image), 60 μl of a solution of Gadomer that is 50 mM Gd, (a dose of 0.12 mmol Gd/kg body weight for a 25 g mouse) was injected over 10 s using a Harvard 2 dual syringe pump (Harvard Clinical Technology, Inc.; South Natick, MA), via either a tail or jugular venous catheter.

Modeling DCE Data

The starting point for our analysis was the Patlak model, containing independent parameters vp (fractional blood-plasma volume) and Ktrans (volume transfer constant between blood plasma and extravascular extracellular space (EES)) (21, 22). Significant deviation from linearity in the Patlak plot led us to include a third independent parameter, ve (fractional volume of EES), to account for contrast agent efflux from EES to blood plasma (21, 23). Conversion of signal intensity to concentration of contrast agent (CA) was achieved by standard methods (10). Negligible T2*-weighting of the images and the fast-exchange limit were both assumed (10, 12). Values of T10 were 1.6 s for cervix (24) and 1.4 s for leg muscle (vide infra) (25); r1 of Gadomer was taken to be 9.1 mM−1 s−1 (26).

The determination of vp, ve, and Ktrans requires knowledge of CA concentration in the blood plasma (arterial input function (AIF)) throughout the experimental time course. Traditionally, the AIF is measured from a large vessel within the image, carefully avoiding inflow and partial volume artifacts (10, 27, 28). However, direct AIF determinations in mice are often difficult due to small spatial dimensions and motion effects. Instead, we derived the AIF from leg muscle reference tissue (29, 30), using the following parameter values for muscle tissue: Ktrans(Gadomer) = 0.002 min−1, vp = 0.021, and ve = 0.085 (31). To convert CA concentration data into physiological parameters, a region of interest (ROI) was drawn over the transformation zone of the cervix and the data were modeled in Matlab (Mathworks Inc, Natick, MA) using a variable projection (VARPRO) non-linear least squares approach (32).

Detection of antiangiogenesis mediated by VEGF Trap

K14-HPV16 transgenic mice were scanned by DCE-MRI prior to treatment, treated biweekly for 2 weeks with 500μg of VEGF Trap (Regeneron Pharmaceuticals Inc., Tarrytown, NY) or vehicle (5mM Phosphate, 5mM Citrate, 100mM NaCl, 0.005% Tween 20) via i.p. injection, then rescanned to test therapy-associated microvascular permeability alterations.

Determination of microvessel density

Isoflurane anesthetized mice were i.v. injected with 50 μg FITC-conjugated Lycopersicon esculentum lectin (Vector Laboratories, Burlingame, CA, #FL-1171) and after 3 min the left ventricle was perfused with 10% formalin (Fisher Scientific International Inc, Hampton, NH) for 3 minutes followed by 10% sucrose for 1 minute (Perfusion One Rodent System, McCormick Scientific, St, Louis, MO). The entire reproductive tract was removed, the vaginal cavity filled with OCT freezing media, embedded in OCT (posterior-side down), flash frozen using liquid nitrogen, and stored at −80°C. Sixty micron cryosections were mounted using SlowFade Gold with 4',6-diamino-2-phenylindole (InVitrogen, Carlsbad, CA), and viewed under appropriate filter sets using an Olympus BX61 microscope equipped with a Fire Wire Colorview II camera (Olympus, Center Valley, PA). Images of lectin-perfused vessels in the cervical transformation zone taken at 40× magnification were analyzed using Olympus MicroSuite Biological Suite software. For each image, 4 equally sized rectangular ROIs were identified along the epithelial-stromal border of the transformation zone. Subepithelial microvasculature was delineated by creating RGB color detection profiles to increase signal to noise and identify as many vessels as possible. These profiles were used for all images.

Determination of physical molecular leakage

ChromPure sheep IgG, Fc fragment, 50μg in 50μL PBS, (Jackson ImmunoResearch Laboratories, Inc, West Grove, PA, #013-000-008) was injected i.v. and allowed to circulate for 2 hours, followed by FITC-lectin injection, formalin perfusion, OCT whole organ embedding, and 60 micron cryosectioning, as described above. Air-dried sections were rinsed in PBS ×3, blocked for 3 hr with Dako Protein block (DAKO #X0909, Carpentina, CA), incubated overnight at 4°C with anti-sheep Cy3-conjugated AffiniPure Donkey IgG, (Jackson Immuno Research Laboratories, INC #713-165-147), diluted 1:100 in Dako Antibody Diluent (DAKO, #S3022), and mounted using SlowFade Gold with DAPI. For visualizing Fc Fragment leakage, images were captured using the Cy3 filter from the sample with the highest signal, which was used to determine the optimal camera gain settings. A control image from a non-injected mouse was used to correct for the Cy3 background signal. An ROI obtained from the control image was used for background subtraction for analysis of signal intensity of Fc-injected experimental tissue sections using the MicroSuite software.

Statistical analysis

Data are mean ± S.D. Mann-Whitney U, paired or unpaired Student's t-tests were used to determine statistical significance (GraphPad Prism, San Diego, CA).


Cervical transformation zone MRI imaging and histopathological correlation

First, we developed MRI techniques to visualize the entire mouse female reproductive tract, including the vagina, cervix, and lower uterus (Figure 1B). High-resolution, in vivo respiratory-gated spin-echo coronal and transaxial MR images were obtained of a 3-month-old, estrogen-treated, nontransgenic mouse with an in-plane resolution of 150 μm (Figure 1, Panels A, and C). The coronal MRI image (Figure 1, Panel A) was a striking reproduction of the actual organ anatomy delineating the cervical isthmus, canal, outer cervix, and upper vagina (Figure 1, Panel B). Transaxial images also delineated all three zones of the cervix: the upper cervical-uterine junction (data not shown) the mid-cervix with the transformation zone and isthmus division septum leading to the two uterine horns (Figure 1, Panel C, upper), and the lower cervix, here containing a single central canal and laterally bounded by the adjacent vaginal walls (Figure 1, Panel C, middle), and the vagina (Figure 1, Panel C, lower).

Figure 1
Development and histological validation of magnetic resonance imaging (MRI) of the mouse female reproductive tract

Anatomic MRI also detected cervical cancer that occurs in 80% of K14-HPV16 transgenic mice treated with estrogen for six months (13, 15) (Figure 2, see multiple delineating black arrowheads in Panels A and B). The malignancy was first evident in the mid-cervix, both histologically and in the MRI (data not shown). Histological analysis further demonstrated spread of the cancer to the lower cervix, as determined by the single “X”-shaped lumen (Figure 2, Panel A, green arrows indicate upper vagina), which invaded almost through the anterior cervical wall adjacent to the bladder (Figure 2, Panel A, six o'clock position), and effaced the right portion of the cervical canal. MRI delineated the same extent of tumor invasiveness (Figure 2, Panel B, red arrowheads), but with the additional feature of obliteration of the right cervical lumen, not evident on the histopathology (Figure 2, Panel A). This slight discordance of the MRI vis-à-vis histology is due either to volume averaging in the former or, more likely, tissue shrinkage during processing in the latter.

Figure 2
MRI detection of an invasive squamous cervical cancer

Differential histology and microvascular area in transgenic premalignant high-grade dysplasia

As high grade dysplasia, usually arising from the cervical transformation zone, is the source of subsequent invasive cervical cancer in both humans (1) and this transgenic model (13) and is a potential target for chemoprevention, we focused our histopathological (Figure 3, Panels A–D), microvascular (Figure 4, Panels A–D), and MRI imaging analyses on this anatomic region and this stage of progression in the transgenic mice treated with estrogen for three months (13). Moderate to high-grade cervical dysplasia (Figure 3, Panels B and D), but no cervical cancer, was evident, similar to our previous experience (13). Nontransgenic littermates treated with the same dosage and duration of estrogen evidenced hyperplasia without dysplasia (Figure 3, Panels A and C).

Figure 3
Histopathology of mid-high grade cervical dysplasia in estrogen treated K14-HPV16 transgenic mice
Figure 4
Microvascular anatomy and density in estrogen treated cervices

Induction of angiogenesis and increased microvascularity has been documented previously in patients with cervical dysplasia and also in dysplastic skin lesions in K14-HPV16 transgenic mice (33). Thus, we first determined differences in microvascular morphology in transgenic cervical dysplasia Figure 4, Panel A, right upper and lower, compared to nontransgenic estrogen-induced hyperplasia (Figure 4, Panel A, left upper and lower). There were two distinct zones of microvasculature in the mouse cervix, the subepithelial region and the deep stromal area (Figure 4, Panel A, left upper, white arrowheads and arrow). The subepithelial region was most affected by mouse genotype, with the transgenic microvasculature forming tufts and projections into the overlying dysplastic epithelium (Figure 4, Panel A, right upper, arrowhead, and higher magnification image in right lower). In contrast, nontransgenic microvasculature was flattened, and linearly arrayed in the stroma immediately beneath the hyperplastic epithelium (Figure 4, Panel A, left upper and lower). Image analysis revealed that the area of the subepithelial microvasculature in K14-HPV16 mice was 42.6 ± 8.8% compared to 24.3 ± 8.5% in nontransgenic cervices (Figure 4, Panel B). These data were similar to immunohistochemical analysis of dysplastic microvessel density in this model (14).

Using DCE-MRI to detect vascular permeability changes in premalignant cervical dysplasia

The increased subepithelial microvascularity in three-month estrogen-treated transgenic cervices led us to investigate microvessel biology using DCE-MRI, particularly since K14-HPV16 transgenic mice were known to accumulate activated stromal inflammatory cells that could, via chemo/cytokine release, induce leakage, even in premalignant dysplasia (14). Initially, we conducted these DCE-MRI experiments using Omniscan, a low-molecular weight, Gd-based contrast agent. Visual analysis of sequential temporal images from a DCE-MRI experiment using a nontransgenic mouse revealed a distinct pattern and distribution of Omniscan contrast agent over time (Figure 5, Panel A, and movie in Supplemental Data). Initial pre-injection images were dark (Figure 5, Panel A, left). Immediately following injection, the pelvic branches of the internal iliac artery were visualized, and coincidently the luminal lining of the isthmus and cervical canal brightly enhanced (Figure 5, Panel A, center). This compartment was presumably the subepithelial microvasculature. Finally, contrast was evenly distributed in the cervix, consistent with permeation throughout the deep cervical stroma (Figure 5, Panel A, right). We observed a swift, large signal enhancement following Omniscan injection in both transgenic and nontransgenic animals, consistent with rapid extravasation, making it very difficult to collect high-resolution DCE-MRI data rapidly enough for determination of contrast-agent kinetics.

Figure 5
T1-weighted, gradient-echo images of the mouse female reproductive tract (Panel A), showing the time course for contrast illumination of the pelvic internal iliac artery branches, the presumed epithelial/subepithelial tissue, followed by stromal permeation ...

As such, we switched to Gadomer, a 17 kDa, dendrimer-based contrast agent whose permeability was significantly lower than that of Omniscan, for DCE-MRI. Following data collection, ROI's were drawn in the cervical transformation zone (Figure 5, Panel B, left) and image intensity vs. time curves were derived from within these ROI's (Figure 5, Panel C, left). Intensity vs. time data were converted to concentration vs. time curves, and physiologic parameters were derived, as described above. The red circles in Panel C are experimental data points, while the red curve represents the modeling of this concentration vs. time data using an AIF derived from a reference tissue, muscle (29, 30) (Figure 5, Panel B, right). A representative contrast-agent concentration vs. time curve for muscle is plotted as the blue circles while the derived AIF is shown as the dashed blue curve (Figure 5, Panel C, left). Based upon our analysis of the DCE data, we determined the transfer constant (Ktrans) for an ROI within the cervical transformation zone in 32 transgenic and 17 nontransgenic mice (Figure 5, Panel C, right). Average Ktrans values for the nontransgenic, 0.056 ± 0.029 min−1, and transgenic mice, 0.053 ± 0.020 min−1, were indistinguishable.

To investigate the possibility of low-level microvascular leak to which DCE-MRI was insensitive, we determined the physical leakage (Figure 5, Panel D). Injection of Fc fragments of 50 kDa molecular weight demonstrated a robust microvascular leakage both within an invasive cancer (Figure 5, left Panel D, left), and within a dysplasia adjacent to the malignancy (Figure 5, Panel D, middle). In contrast, we did not detect microvascular leak in dysplastic microvessels in three-month, estrogen-treated transgenic mice (Figure 5, Panel D, right). Thus, despite a two-fold increase in microvascular density (Figure 4, Panel B), the switch to leaky vessels occurred later in this model, possibly at the 4.5 month point wherein carcinoma in situ, and microinvasive cancer first appear (15), or was restricted to frank invasive malignant lesions.

Anatomic response of microvasculature associated with premalignant cervical dysplasia to anti-angiogenic therapy

Noninvasive detection of the response of dysplastic lesions to antiangiogenic or antineoplastic therapies would be a tremendous boon to assess efficacy of cancer prevention. We determined the morphological response of the transgenic cervical microvasculature to VEGF Trap, and then the sensitivity of DCE-MRI to detect alterations of microvascular leakage. We used VEGF Trap (16, 17), because VEGF has also been shown to be incrementally up-regulated in both human dysplastic and malignant cervix (34), and in the cervical dysplasias and malignancies of estrogen-treated, K14-HPV16 transgenic mice (14).

Following a two-week course of VEGF Trap, there was an obvious pruning of the epithelial tufting and a marked overall reduction in microvascular density in the treated mice (Figure 6, Panel A, vehicle, left, and Trap-treated, right). Subepithelial microvascular area decreased 50% in VEGF Trap-treated mice compared to vehicle-treated transgenic mice (Figure 6, Panel B), to a level below that of estrogen-treated nontransgenic controls (Figure 4, Panel B). Despite the marked subepithelial microvascular pruning, there was no difference in the extent or grade of epithelial dysplastic histopathology in VEGF Trap versus vehicle-treated transgenic mice (data not shown).

Figure 6
Effect of VEGF Trap on the transgenic cervical microvasculature

DCE-MRI detects antiangiogenic efficacy in premalignant dysplasia

Next, we tested the ability of DCE-MRI to detect a permeability response that potentially accompanied the VEGF Trap-mediated decrease in microvascular area in cervical dysplasia. Pre- and post-VEGF Trap DCE MRI demonstrated a significant 63% decrease in Ktrans in all of the treated transgenic mice (0.052 ± 0.013 min−1, pre-treatment vs. 0.019 +/− 0.008 min−1, post-treatment; n = 6 mice) (Figure 6, Panel D). DCE-MRI data in vehicle control transgenic mice were heterogeneous and bivariant (Figure 6, Panel C), but all values fell within the range that we previously determined for this group (Figure 5 Panel C, right). Overall, there was no statistical difference in the pre-/post-treatment values, either for each vehicle treated mouse in a paired Student's t-test (Figure 6, Panel D) or within the entire vehicle-treated group (0.048 +/− 0.015 min−1 baseline and 0.045 +/− 0.021 min−1 after two weeks; n = 5 mice). The uniformity of Ktrans reduction in Trap-treated mice compared to the variable response of vehicle-treated transgenic mice suggest that at a 17 kDa cutoff, DCE-MRI was detecting the marked reduction in microvascular area in Trap-treated mice, rather than an inherent effect on the microvessel stability of neoplastic vessels.


The estrogen-treated K14-HPV16 transgenic mouse model of cervical carcinogenesis, has been extensively studied since inception (14, 3537), and its relevance for human disease has been validated by both detailed histopathological and genome-wide expression analysis (15, 36, 38). Other work has demonstrated angiogenesis induction coincident with high-grade dysplasia, similar to our findings (14). Moreover, dysplastic and malignant angiogenesis in this model has been linked to both macrophage and neutrophil expression of proteases and angiogenic factors (3, 14). Thus, the emerging importance of the K14-HPV16 transgenic mouse as a preclinical platform for testing drugs that target both malignant and dysplastic angiogenesis (3, 14) motivated us to undertake a detailed MRI-based analysis to both noninvasively determine cervico-vaginal anatomy, and to also interrogate microvascular biology associated with premalignant dysplasia.

Here, in vivo MRI at 4.7T clearly distinguished epithelium and subepithelial microvasculature from the relatively avascular deep cervical stroma, and detected cervical cancers in transgenic mice. The next challenge was determination of microvascular leak in premalignant cervical dysplasia using dynamic contrast MRI. DCE-MRI of the lower female reproductive tract poses several unique data acquisition and analysis challenges, including the relatively small size of the target organs (only a few pixels in many images), bladder proximity, and respiratory motion effects. Nonetheless, we successfully determined the transfer constant (Ktrans), describing vascular permeability/leak in the cervical transformation zone. Within each of these groups of animals, we observed a wide range of Ktrans, though the average values for transgenic and nontransgenic mice were indistinguishable. The DCE-MRI data were supported, in part, by the lack of detection of immunofluorescent analysis of Fc fragment leakage, though the molecular mass of this protein (50 kDa) was larger than Gadomer (17 kDa). Thus, the DCE-MRI data suggest that microvessels associated with dysplastic lesions at the mid-point of carcinogenic progression do not elaborate fenestrations or other leakage-associated structures (39), despite their increased density, abnormal morphology, and associated stromal inflammation (3, 4, 14).

In contrast, we were clearly able to detect and monitor the decrease in cervical microvasculature due to VEGF-trap (16, 17) in transgenic dysplasias. The antiangiogenic potency of VEGF Trap was highlighted by the marked 50% reduction in Ktrans in Trap-treated compared to vehicle-treated transgenic mice. VEGF Trap also decreased dysplastic microvascular area to a level that was 25% lower than the estrogen-treated nontransgenic controls. These data suggested that VEGF was the predominant coordinator of angiogenesis of cervical dysplasia, despite the documented contribution of PDGF-C and FGF-2 to this process in this model (4). These data also highlighted a correlation between Ktrans and microvascular area within the 17 kDa pore size cutoff.

The lack of efficacy of VEGF Trap in ameliorating cervical dysplasia, despite its pronounced microvascular potency, suggests the need for combining antiangiogenic therapy with agents targeting dysplastic epithelium, as in clinical scenarios for established malignancies (40). However, our study also highlights the ability of DCE-MRI to potentially detect reduction in the microvasculature that would likely accompany effective therapy for dysplasia. As a prominent goal of future prevention therapies is noninvasive ablation of high-risk dysplasia, DCE-MRI could become a valuable monitoring tool for treatment efficacy.


We thank Justin Halder, University of Illinois, Urbana for help with development of the Matlab-based analysis code, Krista Olsen and Aaron Lee for technical assistance, and the Bayer Schering Pharma AG; Berlin, Germany for the gift of Gadomer contrast agent. This work was supported by an NIH/NCI Small Animal Imaging Resource Program (SAIRP) grant (U24 CA83060), the Alvin J. Siteman Cancer Center at Washington University in St. Louis (P30 CA91842), and funds from the Gynecological Cancer Division, Washington University in St Louis School of Medicine.


1. Schiffman M, Castle PE, Jeronimo J, Rodriguez AC, Wacholder S. Human papillomavirus and cervical cancer. Lancet (N Am Ed) 2007;370:890–907. [PubMed]
2. Marshall K. Cervical dysplasia: early intervention. Altern Med Rev. 2003;8:156–70. [PubMed]
3. Pahler JC, Tazzyman S, Erez N, et al. Plasticity in tumor-promoting inflammation: impairment of macrophage recruitment evokes a compensatory neutrophil response. Neoplasia (N Y) 2008;10:329–40. [PMC free article] [PubMed]
4. Pietras K, Pahler J, Bergers G, Hanahan D. Functions of paracrine PDGF signaling in the proangiogenic tumor stroma revealed by pharmacological targeting. PLoS Med. 2008;5:e19. [PMC free article] [PubMed]
5. Glunde K, Pathak AP, Bhujwalla ZM. Molecular-functional imaging of cancer: to image and imagine. Trends Mol Med. 2007;13:287–97. [PubMed]
6. Herholz K, Coope D, Jackson A. Metabolic and molecular imaging in neuro-oncology. Lancet Neurol. 2007;6:711–24. [PubMed]
7. Marcus CD, Ladam-Marcus V, Cucu C, Bouche O, Lucas L, Hoeffel C. Imaging techniques to evaluate the response to treatment in oncology: Current standards and perspectives. Crit Rev Oncol Hematol. 2008 [PubMed]
8. Barrett T, Brechbiel M, Bernardo M, Choyke PL. MRI of tumor angiogenesis. J Magn Reson Imaging. 2007;26:235–49. [PubMed]
9. Jackson A, O'Connor JP, Parker GJ, Jayson GC. Imaging tumor vascular heterogeneity and angiogenesis using dynamic contrast-enhanced magnetic resonance imaging. Clin Cancer Res. 2007;13:3449–59. [PubMed]
10. Jackson A, Buckley DL, Parker GJM, editors. Dynamic contrast-enhanced magnetic resonance imaging in oncology. Springer; Berlin: 2005.
11. Buckley DL. Uncertainty in the analysis of tracer kinetics using dynamic contrast-enhanced T1-weighted MRI. Magn Reson Med. 2002;47:601–6. [PubMed]
12. Yankeelov TE, Rooney WD, Li X, Springer CS., Jr. Variation of the relaxographic “shutter-speed” for transcytolemmal water exchange affects the CR bolus-tracking curve shape. Magn Reson Med. 2003;50:1151–69. [PubMed]
13. Elson DA, Riley RR, Lacey A, Thordarson G, Talamantes FJ, Arbeit JM. Sensitivity of the cervical transformation zone to estrogen-induced squamous carcinogenesis. Cancer Res. 2000;60:1267–75. [PubMed]
14. Giraudo E, Inoue M, Hanahan D. An amino-bisphosphonate targets MMP-9-expressing macrophages and angiogenesis to impair cervical carcinogenesis. J Clin Investig. 2004;114:623–33. [PMC free article] [PubMed]
15. Riley RR, Duensing S, Brake T, Munger K, Lambert PF, Arbeit JM. Dissection of human papillomavirus E6 and E7 function in transgenic mouse models of cervical carcinogenesis. Cancer Res. 2003;63:4862–71. [PubMed]
16. Rudge JS, Thurston G, Davis S, et al. VEGF trap as a novel antiangiogenic treatment currently in clinical trials for cancer and eye diseases, and VelociGene- based discovery of the next generation of angiogenesis targets. Cold Spring Harbor Symp Quant Biol. 2005;70:411–8. [PubMed]
17. Wachsberger PR, Burd R, Cardi C, et al. VEGF trap in combination with radiotherapy improves tumor control in u87 glioblastoma. Int J Radiat Oncol, Biol, Phys. 2007;67:1526–37. [PubMed]
18. Garbow JR, Zhang Z, You M. Detection of primary lung tumors in rodents by magnetic resonance imaging. Cancer Res. 2004;64:2740–2. [PubMed]
19. Christensen KA, Grant DM, Shulman EM, Walling C. Optimal determination of relaxation times of Fourier transform nuclear magnetic resonance. Determination of spin-lattice relaxation times in chemically polarized species. J Phys Chem. 1974;78:1971–7.
20. Fram EK, Herfkens RJ, Johnson GA, et al. Rapid calculation of T1 using variable flip angle gradient refocused imaging. Magn Reson Imaging. 1987;5:201–8. [PubMed]
21. Buckley DL, Shurrab AE, Cheung CM, Jones AP, Mamtora H, Kalra PA. Measurement of single kidney function using dynamic contrast-enhanced MRI: comparison of two models in human subjects. J Magn Reson Imaging. 2006;24:1117–23. [PubMed]
22. Tofts PS, Brix G, Buckley DL, et al. Estimating kinetic parameters from dynamic contrast-enhanced T(1)-weighted MRI of a diffusable tracer: standardized quantities and symbols. J Magn Reson Imaging. 1999;10:223–32. [PubMed]
23. Tofts PS. Modeling tracer kinetics in dynamic Gd-DTPA MR imaging. J Magn Reson Imaging. 1997;7:91–101. [PubMed]
24. de Bazelaire CM, Duhamel GD, Rofsky NM, Alsop DC. MR imaging relaxation times of abdominal and pelvic tissues measured in vivo at 3.0 T: preliminary results. Radiology. 2004;230:652–9. [PubMed]
25. Bottomley PA, Foster TH, Argersinger RE, Pfeifer LM. A review of normal tissue hydrogen NMR relaxation times and relaxation mechanisms from 1-100 MHz: dependence on tissue type, NMR frequency, temperature, species, excision, and age. Med Phys. 1984;11:425–48. [PubMed]
26. Rohrer M, Bauer H, Mintorovitch J, Requardt M, Weinmann HJ. Comparison of magnetic properties of MRI contrast media solutions at different magnetic field strengths. Investig Radiol. 2005;40:715–24. [PubMed]
27. Parker GJ, Roberts C, Macdonald A, et al. Experimentally-derived functional form for a population-averaged high-temporal-resolution arterial input function for dynamic contrast-enhanced MRI. Magn Reson Med. 2006;56:993–1000. [PubMed]
28. Rijpkema M, Kaanders JH, Joosten FB, van der Kogel AJ, Heerschap A. Method for quantitative mapping of dynamic MRI contrast agent uptake in human tumors. J Magn Reson Imaging. 2001;14:457–63. [PubMed]
29. Kovar DA, Lewis M, Karczmar GS. A new method for imaging perfusion and contrast extraction fraction: input functions derived from reference tissues. J Magn Reson Imaging. 1998;8:1126–34. [PubMed]
30. Yankeelov TE, Luci JJ, Lepage M, et al. Quantitative pharmacokinetic analysis of DCE-MRI data without an arterial input function: a reference region model. Magn Reson Imaging. 2005;23:519–29. [PubMed]
31. Faranesh AZ, Kraitchman DL, McVeigh ER. Measurement of kinetic parameters in skeletal muscle by magnetic resonance imaging with an intravascular agent. Magn Reson Med. 2006;55:1114–23. [PMC free article] [PubMed]
32. Golub GH, Pereya V. Separable nonlinear least squares: the variable projection method and its applications. Inverse Probl. 2003;19:R1–R26.
33. Smith-McCune K, Zhu YH, Hanahan D, Arbeit J. Cross-species comparison of angiogenesis during the premalignant stages of squamous carcinogenesis in the human cervix and K14-HPV16 transgenic mice. Cancer Res. 1997;57:1294–300. [PubMed]
34. Smith-McCune KK, Weidner N. Demonstration and characterization of the angiogenic properties of cervical dysplasia. Cancer Res. 1994;54:800–4. [PubMed]
35. Arbeit JM, Howley PM, Hanahan D. Chronic estrogen-induced cervical and vaginal squamous carcinogenesis in human papillomavirus type 16 transgenic mice. Proc Natl Acad Sci U S A. 1996;93:2930–5. [PubMed]
36. Lu ZH, Wright JD, Belt B, Cardiff RD, Arbeit JM. Hypoxia-inducible factor-1 facilitates cervical cancer progression in human papillomavirus type 16 transgenic mice. Am J Pathol. 2007;171:667–81. [PubMed]
37. Shai A, Pitot HC, F LP. p53 Loss synergizes with estrogen and papillomaviral oncogenes to induce cervical and breast cancers. Cancer Res. 2008;68:2622–31. [PMC free article] [PubMed]
38. Brake T, Connor JP, Petereit DG, Lambert PF. Comparative analysis of cervical cancer in women and in a human papillomavirus-transgenic mouse model: identification of minichromosome maintenance protein 7 as an informative biomarker for human cervical cancer. Cancer Res. 2003;63:8173–80. [PubMed]
39. Nagy JA, Dvorak AM, Dvorak HF. VEGF-A and the Induction of Pathological Angiogenesis. Annu Rev Pathol. 2007;2:251–75. [PubMed]
40. Jain RK, Duda DG, Clark JW, Loeffler JS. Lessons from phase III clinical trials on anti-VEGF therapy for cancer. Nat Clin Pract Oncol. 2006;3:24–40. [PubMed]