TUMOR VASCULAR BIOLOGY
Cancer is characterized by an abnormal tumor vasculature that is a mediator of treatment resistance.1 As the rapidly replicating tumor outgrows its existing blood supply, the tumor secretes proangiogenic factors to spur neoangiogenesis and sustain tumor growth. These new blood vessels are dilated, tortuous, and more permeable, resulting in increased interstitial fluid pressure. As tumor cell density increases, compression of the new blood vessels occurs, causing the vasculature to be very inefficient in these parts of the tumor and stimulating further abnormal angiogenesis.2 Other areas of the tumor may be less squeezed, so there can be significant spatial heterogeneity in tumor vasculature. Areas with the most compressed and tortuous vascular network will not have enough oxygen delivery for radiation to be most effective or adequate delivery of cytotoxic chemotherapy necessary to kill tumor cells. Furthermore, hypoxia itself is a maker of more aggressive tumors as discussed below. Thus, targeting this heterogeneous tumor vasculature is a promising therapeutic strategy.
Glioblastoma (GBM), the most common malignant primary brain tumor is a particularly vascular tumor and its pathological hallmark is abnormal microvascular proliferation. In particular, vascular endothelial growth factor (VEGF) is oversecreted in GBM and is 1 of the primary drivers behind tumor angiogenesis.3 With the approval of bevacizumab, a monoclonal antibody to VEGF in 2009 for recurrent GBM, blocking VEGF has become an important part of GBM therapy. However, despite this strong biological rationale for antiangiogenic therapy (AAT) and early promising phase II data, recent phase III data failed to demonstrate an overall survival benefit for bevacizumab in GBM.4 The phase III RTOG 0825 and AVAglio studies randomized patients with newly diagnosed GBM treated with standard chemoradiation to placebo or bevacizumab.5,6 The median overall survival was the same in the 2 arms in both studies (about 16 months), suggesting a limited role for bevacizumab in unselected patient populations with newly diagnosed GBMs. Given these disappointing results, more work needs to be done to understand how this class of vessel-targeting agents works in order to determine their role in GBM management. This review covers how vascular magnetic resonance imaging (MRI) tools available for clinical use today can be applied in humans to noninvasively study tumor vascular biology and, specifically, response to AAT. With these imaging tools, we may be able to optimize therapy and identify those patients most likely to benefit from AAT.
IMPACT OF ANTIANGIOGENIC THERAPY ON TUMOR VASCULATURE
There are several hypotheses for how AAT may impact tumor biology and help a subset of patients. The original hypothesis was that these drugs work to starve the tumor of its blood supply.7 Alternately, these drugs may work to normalize the abnormal tumor vasculature and revert the vessels to a more normal and efficient state by restoring the proangiogenic and antiangiogenic factor balance, pruning immature abnormal vessels, and recruiting perivascular cells to protect mature vessels.8 By normalizing the abnormal tumor vasculature, AAT improves tumor blood flow, resulting in better delivery of cytotoxic chemotherapy and oxygen (a necessity for radiation to be effective).9,10 In addition, normalization affects the tumor microenvironment and shifts the immunosuppressive phenotype of the tumor microenvironment to a more immunostimulatory environment.11 This shift allows activated tumor-associated macrophages to target the tumor and suppress tumor growth. Likely, there is a dose effect of AAT so that lower doses result in more normalization, and higher doses result in pruning.12–14
These very complex interactions occurring in the tumor microenvironment can be studied with MRI, and changes in MRI parameters can potentially guide therapeutic decisions. For example, a patient who is responding to AAT via pruning might not benefit from the addition of chemotherapy so he/she could be spared the associated toxicity. Because the exact mechanisms by which antiangiogenic drugs remodel the vasculature are not fully understood, the identification and understanding of valid prognostic and predictive imaging biomarkers that can noninvasively track response to therapy are thus a major challenge for the field of oncology.
BEYOND STRUCTURAL IMAGING—VASCULAR MRI IN BRAIN TUMORS
Limitations of Conventional MRI in Brain Tumors
While the criterion standard for radiographic assessment of brain tumor response to chemoradiation therapy15,16 is contrast-enhanced MRI, it is well recognized that the value of conventional, contrast-enhanced MRI for AAT is limited, as reviewed elsewhere.17–20 Antiangiogenic drugs are not traditional cytotoxic agents and do not necessarily result in decreased tumor burden, either immediately or long term. Reductions in the amount of tumor contrast enhancement, the traditional metric used in response criteria, may be misleading in the setting of AAT as a decrease may not reflect a true change in tumor load and thus artificially boost conventional imaging response rates.
Measuring the mechanistic impact of AAT on vascular biology is therefore more closely linked to therapy impact than measuring changes in contrast-enhancing volume or area. Because perfusion MRI is not dependent on a disrupted blood-brain barrier, there is increasing evidence that 2 types of perfusion measurement, MRI-based dynamic susceptibility contrast (DSC) and dynamic contrast-enhanced (DCE) imaging, can provide more useful information about vascular structure and function and subsequently drug mechanism of action and impact on tumor biology.21
Dynamic Susceptibility Contrast MRI
Dynamic susceptibility contrast MRI can estimate changes in tumor vessel structure and function by measuring the passage of a contrast agent bolus through each MRI voxel. This technique is commonly referred to as perfusion MRI, and there are a variety of parameters that can be quantified using this technique, with each providing complementary information about tumor vasculature (Table 1). Cerebral blood flow (CBF)—also denoted blood perfusion—and cerebral blood volume (CBV) measurements are the most commonly reported measures of DSC in the cancer literature.22 Because of the increase in capillary density associated with tumor angiogenesis,23 a large body of work shows that perfusion MRI, and CBV in particular, is strongly correlated to brain tumor malignancy and therapy response. Regardless of contrast enhancement, an area of tumor may have elevated CBV because of increased vessel number or size but have low CBF because of slow flow through the area from increased tortuosity and vessel compression from solid stress.24,25
Changes in tumor vascular parameters are useful markers of response to AAT and shed light on how these drugs are working. For example, patients with newly diagnosed GBM and recurrent GBM treated with cediranib, a VEGF receptor tyrosine kinase inhibitor, who responded with an early increase in subnormal tumor perfusion lived 6 months longer than did those with stable or decreased perfusion.9,26 One possible explanation is that patients with increased perfusion were responding to treatment by vascular normalization of blood flow and oxygen delivery. Particularly in the cohort with newly diagnosed GBM also receiving temozolomide and radiation, the improvement in survival was likely driven by better delivery of chemotherapy and oxygen to the tumor. Strikingly, an increase in tumor perfusion was seen in only 1 of 14 patients with newly diagnosed GBM not treated with AAT, suggesting that the changes in perfusion were an effect of the antiangiogenic drug.
Findings of increased subnormal perfusion have also been found in a subset (one third) of patients with metastatic breast cancer to the brain treated with bevacizumab and carboplatin. Similar to the primary brain tumor patients, these patients had increased survival of approximately 6 months.27 In all these studies, increased perfusion was an early response to AAT but also a temporary physiological phenomenon indicating a window of vascular remodeling. This may represent an early biomarker of response to distinguish those most likely to benefit from combination therapy and potentially spare those not likely to benefit from unnecessary toxicity or ineffective therapy.
Cerebral blood volume has also shown promise in assessing response to AAT with a decrease in abnormal CBV during therapy being associated with improved survival in several human studies.28–30 Unlike the perfusion-based parameter CBF, CBV reflects the static ratio of blood to tissue in a given image voxel, and its value is influenced by the vessel density and abnormal vessel calibers in the tumor, whereas CBF reflects flow and a more dynamic process. Therefore, it is critically important to note that CBF and CBV represent different mechanistic properties of the tissue and in combination can provide very detailed biological information about the vasculature changes following AAT, and changes do not necessarily go in the same direction.
Vessel-caliber MRI provides an average measurement of the cross-sectional diameter of vessels in tissue.31,32 The most common approach to vessel-caliber MRI requires a combined DSC imaging approach using what is known as a spin-echo gradient-echo imaging readout. The spin-echo sequence is sensitive to small vessels below a radius of 10 μm, whereas the gradient echo is sensitive to small and larger vessels. The quotients of these 2 readouts form the basis for estimates of mean vessel diameter, mean vessel density, and vessel size index (Table 1). Intriguingly, while the vessel-caliber MRI technique has been around for more than 20 years, it is only now starting to be available on routine clinical scanners outside the research community, providing the opportunity to assess the added value of vessel-caliber information in human cancer trials.33
Emerging human data show vessel-caliber imaging as a particularly useful biomarker in evaluating changes in brain tumor vasculature in response to AAT. A collective reduction of abnormal tumor vessel calibers was observed in patients with newly diagnosed and recurrent GBM treated with cediranib,9,26 suggesting that reduction of abnormal tumor vessel calibers is a particularly robust effect of AAT. Vessel-caliber MRI also allows for assessment of the selective targeting of certain-sized vessels with various AATs.33 This is in line with preclinical evidence showing AATs preferentially reducing tumor vessel density at the capillary level and cleaning up the chaotic tumor vessel tree.34,35
Measuring Tissue Oxygenation With MRI
An attractive feature of vascular MRI is the presence of paramagnetic deoxyhemoglobin in venous blood that serves as a naturally occurring contrast agent. A standard gradient-echo–based perfusion MRI sequence is sensitive to this blood oxygenation level–dependent contrast and allows for relative assessments of the blood oxygen level in the vasculature, a highly relevant parameter in brain tumors undergoing AAT.36 As noted above, the overdilated, hyperpermeable, and tortuous vessels lead to a heterogeneous perfusion signature with some tumor regions starved of oxygen and nutrients.37–39 Hypoxic cancers are difficult to treat for several reasons. First, an acidic microenvironment attenuates the immune system by making macrophages protumorigenic and immunosuppressive.40 Second, hypoxia may select for more malignant cells because cells that respond to physiological cues normally undergo apoptosis under hypoxic conditions.41 Third, hypoxia is a known cause for radiotherapy resistance.41,42 Improving the tumor microenvironment by alleviating hypoxia may therefore be a promising strategy to enhance cancer treatment and slow tumor progression. To this end, mapping the hypoxic state of cancerous tissue in vivo could be a valuable adjunct in treatment monitoring.
Most studies looking at hypoxia in cancer are preclinical, but emerging data now also show exploratory assessments of cancer oxygenation status in patients with brain tumor using DSC-MRI.43 One attractive feature of DSC-MRI is that theoretical data suggest that the maximum oxygen extraction fraction is determined by blood flow and the level of heterogeneity in capillary transit times, measures that are readily available with DSC-MRI.44 Vessel architectural imaging has recently been proposed as a tool to study tissue oxygenation in brain cancers undergoing AAT.10 Vessel architectural imaging utilizes the full potential of DSC vessel size MRI to obtain functional information on relative tissue oxygen saturation levels during AAT.45 In patients with newly diagnosed GBM treated with cediranib, those who experienced an increase in tumor perfusion also had an improvement in tumor oxygenation.9 Similar results were found in patients with breast cancer with brain metastases treated with bevacizumab and carboplatin.27 There are also data supporting use of vessel architectural imaging in non–central nervous system organs without a blood-brain barrier.46 Having this measurement of relative tissue oxygenation is particularly appealing when trying to distinguish vessel pruning, resulting in decreased oxygenation, and vessel normalization, resulting in improved oxygenation. This information may eventually influence selection of concomitant therapies and drive development of hypoxia-targeting agents.
Dynamic Contrasted-Enhanced MRI
Antiangiogenic therapy has a profound impact on vascular permeability. This is evident in the remarkable decrease in contrast enhancement often seen after initiation of therapy and decrease in peritumoral edema.26 Most patients treated with AAT are able to reduce their dose of corticosteroids, which has a significant impact on quality of life.26 Dynamic contrast enhancement MRI is a technique that can measure changes in vascular permeability using pharmacokinetic models to quantify the leakage of contrast agent across the blood-brain barrier (Table 1).47 The benefit of DCE-MRI in the setting of AAT is the quantitative estimation of average numerical values of contrast extravasation into the extravascular space, whereas measurement of enhancing tumor area is subject to significant interobserver variability.48 A range of human studies have shown that pretreatment Ktrans or an early drop in Ktrans drops can predict response to AAT.26,49,50
A particularly important question in brain tumor treatment is how well drugs get into the brain. Dynamic contrasted-enhanced MRI may be a useful tool to assess how open the blood-brain barrier is and thus how easy it might be for a drug to cross the blood-brain barrier. In the scenario of AAT where there does appear to be decreased leakiness, DCE-MRI could help identify critical windows of opportunity when concomitant therapy will be most effective. Several studies on VEGF targeting tyrosine kinase inhibitors evaluated by DCE-MRI suggest an apparent linear relationship between reductions in permeability by Ktrans and the administered dose of cediranib, vatalanib, sorafenib, and pazopanib, among others.21 This effect, however, was not observed with a range of vascular disrupting agents such as fosbretabulin, where instead a therapeutically active window was observed below the maximum tolerated dose.21 Therefore, having this tool of measuring how open the blood barrier is will be critical to drug development and optimize dose with concomitant therapies.
Part of the challenge with DCE-MRI, though, is that the parameter, Ktrans, is a composite of vascular permeability, vessel surface area, and blood flow, which may complicate its interpretation with AAT data.47 In conditions of low permeability where the contrast agent cannot leak, Ktrans reflects permeability, but in conditions of high permeability, Ktrans will also reflect blood flow. Therefore, more sophisticated pharmacokinetic models for DCE imaging are being developed that accurately reflect the complicated vascular biology seen in tumors.51
Challenges With Vascular MRI of the Brain
Several challenges remain before some of these sophisticated vascular imaging techniques can gain more widespread clinical use. These challenges range from an incomplete understanding of brain tumor biology to more technical aspects of image acquisition and processing. Vascular remodeling is a dynamic process, so even something as simple as selection of imaging time points during AAT will influence how to interpret vascular changes as there is an early window of normalization after which improved perfusion may not be seen.26 Furthermore, different doses of antiangiogenic drugs may also impact tumor vasculature differentially.12 Therefore, the ability to serially measure changes in tumor vasculature is critical to better understanding an individual patient’s response to AAT.
One conspicuous challenge facing the imaging field today is determining the optimal imaging parameters or combination of parameters to study tumor vasculature—particularly as new approaches are proposed and each will require validation to determine the true added value. For example, there is no globally accepted consensus on the dose and type of contrast agents for DSC imaging and also how best to correct for leakage of the contrast agent in brain tumors.52 In order to make vascular imaging a routine tool, certain steps need to be done to optimize or, more specifically harmonize, its use. Efforts to generate consensus around brain tumor imaging are underway through groups such as the Jump Starting Brain Tumor Coalition and the National Institutes of Health–funded Quantitative Imaging Network.53 These efforts also focus on simplifying and creating an automatic start-to-end analysis pipeline that will greatly improve the utility of these methods outside the research community. Having this consensus in place will enable multicenter studies to be conducted to better understand the role of vascular imaging as well as to provide guidelines on how to best construct an optimal, multicenter image trial design.54
IMPROVING AAT EVALUATION: MAPPING TUMOR HETEROGENEITY
With recent advents in the understanding of molecularly targetable pathways and pharmacogenomic predictors, the oncologic community has started to appreciate the apparent overwhelming heterogeneity within each cancer and cancer type and the need for a “biology-to-trial” rather than a “compound-to-trial” approach for targeted therapies.55 The heterogeneity of a single primary brain tumor is particularly striking—cell-to-cell variability has a profound and direct consequence on the time to progression as well as inherent resistance to specific therapies. Although AAT is considered to be a less specific attack on the tumor, the heterogeneous gene expression within tumors dictates the need for new predictive and prognostic biomarkers that can separate responders from nonresponders beyond what can be achieved with current response criteria such sa RECIST and RANO. The lack of appropriate patient selection is an often cited reason for the failure of RTOG 0825 and AVAglio studies mentioned previously.5,6 Moreover, as systemic therapy improves, metastases to the brain are increasingly common in cancer patients. This patient group also calls for revised and specific response criteria because these patients often have multiple lesions, where each satellite lesion represents its own cancerous microenvironment and may therefore respond differently to AATs.56
Magnetic resonance imaging tools are particularly appealing for mapping of tumor heterogeneity. In addition to the noninvasive nature of the modality that allows for safe, repeated examinations, a single magnetic resonance measurement may capture physiological changes of a heterogeneous tumor microscopically—down to interactions between atoms—but can represent that information on a submillimeter scale and so captures very detailed information about tumor biology in sophisticated ways. Thus, if different parts of the tumor respond differently to a particular treatment, such as in the setting of AATs that are selectively sensitive to peripheral vessel structures compared with those of the center of the tumor, this could be detected by MRI but may be missed if only a portion of the tumor is sampled surgically.34,35 The practical implication of imaging-based analysis is the possibility to perform more advanced statistical assessments such as histogram analysis and voxel-based approaches that effectively capture tumor heterogeneity and minimize reader variability (Fig. 1).57
NEXT STEPS IN VASCULAR IMAGING
Overall, what appears to be most critical to predicting tumor response is whether these imaging parameters indicate a return from a hostile and malignant phenotype to a more favorable, efficient, and normal-appearing microenvironment. Early clinical data suggest that whether it is perfusion, CBV, Ktrans, or vessel size, patients whose values revert to what is seen in normal tissue are the ones who appear to do better.9,21 Therefore, the general approach of looking at threshold levels and how much individual imaging parameter values are increasing or decreasing may be less helpful than assessing the induced change relative to normal tissue. Given some inherent variability in imaging from time point to time point, reference to normal tissue may be more accurate than absolute or threshold values.
The ideal scenario may include a combined analysis approach such that information from each parameter could be incorporated into a single tumor vascular map.49 Although these parameters have been discussed individually up to this point, each one provides complementary information about dynamic changes in vasculature structure and function. Next steps for vascular imaging must include improved longitudinal registration of images over time or improved tools to look at regional or voxel-based changes in tumors during therapy in order to capture the known tumor heterogeneity discussed previously.58 Finally, although not clinically available yet, higher field strengths may provide more detailed information on tumor blood vessels, and studies using 7-T MRI are ongoing.59 These multimodal approaches will also enable incorporation of additional imaging techniques such as positron emission tomography imaging. Positron emission tomography provides complementary information to MRI and opens up additional avenues to study tumor metabolism. Novel tracers such as fluoromisonidazole that highlight areas of tumor hypoxia will be helpful to validate vascular MRI tools but are beyond the scope of this review.
Finally, while we have focused on vascular imaging in the setting of AAT, there may be uses beyond this scenario, in particular the clinical conundrum of treatment-related changes in which there is an increase in contrast enhancement that is indistinguishable from active tumor. Early treatment-related changes in reaction to chemoradiation or immunotherapy occur in as many as 30% of patients or later radiation necrosis in as many as 15% of patients.60,61 Changes in tumor vasculature are thought to be the underlying biology behind these processes, so being able to image changes in vasculature will improve interpretation of MRIs in these patients and improve our understanding of these pathological processes.62 Correct identification of treatment-related versus tumor-related imaging changes is critical to correctly managing patients and avoiding unnecessary therapy.
AN INDUSTRY PERSPECTIVE
The adoption of vascular MRI tools into the broad brain tumor community is both a work in progress and a complex process. As with all practices in medicine, adoption of new approaches is driven by the interplay between local preference, technique familiarity and availability, education and awareness of practitioners, and frequently reimbursement practices. Even relatively straightforward questions in perfusion MRI such as the selection, volume, and administration rate of contrast agents can be influenced by local formulary guidelines and financial questions. The availability of CBV, CBF, and Ktrans maps is no longer a question of the availability of US Food and Drug Administration–cleared software packages, but can be highly dependent on local radiology practices, workflow constraints, and practitioner demand. As with many aspects of medical practice, there remains substantial variability even in the face of promising early data.
Despite these challenges, the shifting nature of reimbursement practices and focus on “value” rather than “volume” bode well for these mechanistic techniques. As health systems and practitioners assume financial responsibility linked to outcomes, choosing the optimal therapy and switching off of ineffective therapies as quickly as possible will gain visibility and attention. This in turn will likely increase demand for accurate assessment methods, and the tools described here are particularly well suited for assessing the effects of AAT, which, because of its frequent high cost, may come in for particular scrutiny.
In conclusion, MRI has the ability to measure structural and functional changes in tumor vasculature. These tools have tremendous promise for improving outcomes for patients with brain tumors. The failure of AAT in large randomized phase III trials of GBM is puzzling given the strong biological rationale and clear benefit in certain patients. This argues for a need to better select those patients most likely to benefit. Dynamic, physiologic MRI, with its ability to evaluate vasculature, holds the potential to be a key method for optimizing therapy. Assessing changes in tumor perfusion, vessel size, and oxygenation relative to normal tissue may help direct dosing of AAT as well as whether to combine AAT with chemotherapy or radiation. These tools directly reflect changes in the tumor microenvironment and thus reflect the direct impact of drugs, improving our understanding of why these drugs may or may not be working in individual patients. The next step forward will be to standardize these techniques to allow for their wider application in both clinical and research settings.
1. Jain RK, di Tomaso E, Duda DG, et al. Angiogenesis in brain tumours. Nat Rev Neurosci
. 2007; 8: 610–622.
2. Padera TP, Stoll BR, Tooredman JB, et al. Pathology: cancer cells compress intratumour vessels. Nature
. 2004; 427: 695.
3. Plate KH, Risau W. Angiogenesis in malignant gliomas. Glia
. 1995; 15: 339–347.
4. Friedman HS, Prados MD, Wen PY, et al. Bevacizumab alone and in combination with irinotecan in recurrent glioblastoma. J Clin Oncol
. 2009; 27: 4733–4740.
5. Gilbert MR, Sulman EP, Mehta MP. Bevacizumab for newly diagnosed glioblastoma. N Engl J Med
. 2014; 370: 2048–2049.
6. Chinot OL, Wick W, Mason W, et al. Bevacizumab plus radiotherapy-temozolomide for newly diagnosed glioblastoma. N Engl J Med
. 2014; 370: 709–722.
7. Folkman J. Anti-angiogenesis: new concept for therapy of solid tumors. Ann Surg
. 1972; 175: 409–416.
8. Jain RK. Normalization of tumor vasculature: an emerging concept in antiangiogenic therapy. Science
. 2005; 307: 58–62.
9. Batchelor TT, Gerstner ER, Emblem KE, et al. Improved tumor oxygenation and survival in glioblastoma patients who show increased blood perfusion after cediranib and chemoradiation. Proc Natl Acad Sci U S A
. 2013; 110: 19059–19064.
10. Emblem KE, Mouridsen K, Bjornerud A, et al. Vessel architectural imaging identifies cancer patient responders to anti-angiogenic therapy. Nat Med
. 2013; 19: 1178–1183.
11. Huang Y, Yuan J, Righi E, et al. Vascular normalizing doses of antiangiogenic treatment reprogram the immunosuppressive tumor microenvironment and enhance immunotherapy. Proc Natl Acad Sci U S A
. 2012; 109: 17561–17566.
12. Carmeliet P, Jain RK. Molecular mechanisms and clinical applications of angiogenesis. Nature
. 2011; 473: 298–307.
13. Lorgis V, Maura G, Coppa G, et al. Relation between bevacizumab dose intensity and high-grade glioma survival: a retrospective study in two large cohorts. J Neurooncol
. 2012; 107: 351–358.
14. Levin VA, Mendelssohn ND, Chan J, et al. Impact of bevacizumab administered dose on overall survival of patients with progressive glioblastoma. J Neurooncol
. 2015; 122: 145–150.
15. Wen PY, Macdonald DR, Reardon DA, et al. Updated response assessment criteria for high-grade gliomas: response assessment in neuro-oncology working group. J Clin Oncol
. 2010; 28: 1963–1972.
16. Macdonald DR, Cascino TL, Schold SC Jr, et al. Response criteria for phase II studies of supratentorial malignant glioma. J Clin Oncol
. 1990; 8: 1277–1280.
17. Leach MO, Brindle KM, Evelhoch JL, et al. The assessment of antiangiogenic and antivascular therapies in early-stage clinical trials using magnetic resonance imaging: issues and recommendations. Br J Cancer
. 2005; 92: 1599–1610.
18. Michaelis LC, Ratain MJ. Measuring response in a post-RECIST world: from black and white to shades of grey. Nat Rev Cancer
. 2006; 6: 409–414.
19. Oxnard GR, Morris MJ, Hodi FS, et al. When progressive disease does not mean treatment failure: reconsidering the criteria for progression. J Natl Cancer Inst
. 2012; 104: 1534–1541.
20. Twombly R. Criticism of tumor response criteria raises trial design questions. J Natl Cancer Inst
. 2006; 98: 232–234.
21. O'Connor JP, Jackson A, Parker GJ, et al. Dynamic contrast-enhanced MRI
in clinical trials of antivascular therapies. Nat Rev Clin Oncol
. 2012; 9: 167–177.
22. Covarrubias DJ, Rosen BR, Lev MH. Dynamic magnetic resonance perfusion imaging of brain tumors. Oncologist
. 2004; 9: 528–537.
23. Folkman J. Angiogenesis. Annu Rev Med
. 2006; 57: 1–18.
24. Jain RK, Martin JD, Stylianopoulos T. The role of mechanical forces in tumor growth and therapy. Annu Rev Biomed Eng
. 2014; 16: 321–346.
25. Chauhan VP, Boucher Y, Ferrone CR, et al. Compression of pancreatic tumor blood vessels by hyaluronan is caused by solid stress and not interstitial fluid pressure. Cancer Cell
. 2014; 26: 14–15.
26. Batchelor TT, Sorensen AG, di Tomaso E, et al. AZD2171, a pan-VEGF receptor tyrosine kinase inhibitor, normalizes tumor vasculature and alleviates edema in glioblastoma patients. Cancer Cell
. 2007; 11: 83–95.
27. Kyrre E, Marco P, Vyshak C, et al. Brain breast metastases respond to anti-angiogenic therapy by modes of vascular normalization. Neuro Oncol
. 2014; 16( suppl 5): v143.
28. Aquino D, Di Stefano AL, Scotti A, et al. Parametric response maps of perfusion MRI
may identify recurrent glioblastomas responsive to bevacizumab and irinotecan. PLoS One
. 2014; 9: e90535.
29. Akella NS, Twieg DB, Mikkelsen T, et al. Assessment of brain tumor angiogenesis inhibitors using perfusion magnetic resonance imaging: quality and analysis results of a phase I trial. J Magn Reson Imaging
. 2004; 20: 913–922.
30. Schmainda KM, Prah M, Connelly J, et al. Dynamic-susceptibility contrast agent MRI
measures of relative cerebral blood volume predict response to bevacizumab in recurrent high-grade glioma. Neuro Oncol
. 2014; 16: 880–888.
31. Fisel CR, Ackerman JL, Buxton RB, et al. MR contrast due to microscopically heterogeneous magnetic susceptibility: numerical simulations and applications to cerebral physiology. Magn Reson Med
. 1991; 17: 336–347.
32. Tropres I, Pannetier N, Grand S, et al. Imaging the microvessel caliber and density: principles and applications of microvascular MRI
. Magn Reson Med
33. Emblem KE, Farrar CT, Gerstner ER, et al. Vessel caliber—a potential MRI
biomarker of tumour response in clinical trials. Nat Rev Clin Oncol
. 2014; 11: 566–584.
34. Dobosz M, Ntziachristos V, Scheuer W, et al. Multispectral fluorescence ultramicroscopy: three-dimensional visualization and automatic quantification of tumor morphology, drug penetration, and antiangiogenic treatment response. Neoplasia
. 2014; 16: 1–13.
35. Vakoc BJ, Lanning RM, Tyrrell JA, et al. Three-dimensional microscopy of the tumor microenvironment in vivo using optical frequency domain imaging. Nat Med
. 2009; 15: 1219–1223.
36. Ogawa S, Lee TM, Kay AR, et al. Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proc Natl Acad Sci U S A
. 1990; 87: 9868–9872.
37. Baluk P, Hashizume H, McDonald DM. Cellular abnormalities of blood vessels as targets in cancer. Curr Opin Genet Dev
. 2005; 15: 102–111.
38. Nagy JA, Chang SH, Dvorak AM, et al. Why are tumour blood vessels abnormal and why is it important to know? Br J Cancer
. 2009; 100: 865–869.
39. Yeo SG, Kim JS, Cho MJ, et al. Interstitial fluid pressure as a prognostic factor in cervical cancer following radiation therapy. Clin Cancer Res
. 2009; 15: 6201–6207.
40. Jain RK. Antiangiogenesis strategies revisited: from starving tumors to alleviating hypoxia. Cancer Cell
. 2014; 26: 605–622.
41. Wilson WR, Hay MP. Targeting hypoxia in cancer therapy. Nat Rev Cancer
. 2011; 11: 393–410.
42. Dewhirst MW, Cao Y, Moeller B. Cycling hypoxia and free radicals regulate angiogenesis and radiotherapy response. Nat Rev Cancer
. 2008; 8: 425–437.
43. Remmele S, Sprinkart AM, Muller A, et al. Dynamic and simultaneous MR measurement of R1 and R2* changes during respiratory challenges for the assessment of blood and tissue oxygenation. Magn Reson Med
. 2013; 70: 136–146.
44. Jespersen SN, Ostergaard L. The roles of cerebral blood flow, capillary transit time heterogeneity, and oxygen tension in brain oxygenation and metabolism. J Cereb Blood Flow Metab
. 2012; 32: 264–277.
45. Kiselev VG, Strecker R, Ziyeh S, et al. Vessel size imaging in humans. Magn Reson Med
. 2005; 53: 553–563.
46. Christen T, Qiu D, Wei NW, et al. Advanced analysis of USPIO injection in normal volunteers. In: Proc ISMRM
. ISMRM, Milan, Italy: 2014; 2730.
47. 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: 223–232.
48. Reuter M, Gerstner ER, Rapalino O, et al. Impact of MRI
head placement on glioma response assessment. J Neurooncol
. 2014; 118: 123–129.
49. Sorensen AG, Batchelor TT, Zhang WT, et al. A “vascular normalization index” as potential mechanistic biomarker to predict survival after a single dose of cediranib in recurrent glioblastoma patients. Cancer Res
. 2009; 69: 5296–5300.
50. Kickingereder P, Wiestler B, Graf M, et al. Evaluation of dynamic contrast-enhanced MRI
derived microvascular permeability in recurrent glioblastoma treated with bevacizumab. J Neurooncol
. 2015; 121: 373–380.
51. Ingrisch M, Sourbron S. Tracer-kinetic modeling of dynamic contrast-enhanced MRI
and CT: a primer. J Pharmacokinet Pharmacodyn
. 2013; 40: 281–300.
52. Paulson ES, Schmainda KM. Comparison of dynamic susceptibility-weighted contrast-enhanced MR methods: recommendations for measuring relative cerebral blood volume in brain tumors. Radiology
. 2008; 249: 601–613.
53. Wen PY, Cloughesy TF, Ellingson BM, et al. Report of the Jumpstarting Brain Tumor Drug Development Coalition and FDA clinical trials neuroimaging endpoint workshop (January 30, 2014, Bethesda MD). Neuro Oncol
. 2014; 16 ( suppl 7): vii36–47.
54. Ashton E, Riek J. Advanced MR techniques in multicenter clinical trials. J Magn Reson Imaging
. 2013; 37: 761–769.
55. Yap TA, Sandhu SK, Workman P, et al. Envisioning the future of early anticancer drug development. Nat Rev Cancer
. 2010; 10: 514–523.
56. Lin NU, Lee EQ, Aoyama H, et al. Challenges relating to solid tumour brain metastases in clinical trials, part 1: patient population, response, and progression. A report from the RANO group. Lancet Oncol
. 2013; 14: e396–e406.
57. Emblem KE, Nedregaard B, Nome T, et al. Glioma grading by using histogram analysis of blood volume heterogeneity from MR-derived cerebral blood volume maps. Radiology
. 2008; 247: 808–817.
58. Reuter M, Rosas HD, Fischl B. Highly accurate inverse consistent registration: a robust approach. Neuroimage
. 2010; 53: 1181–1196.
59. Radbruch A, Eidel O, Wiestler B, et al. Quantification of tumor vessels in glioblastoma patients using time-of-flight angiography at 7 Tesla: a feasibility study. PLoS One
. 2014; 9: e110727.
60. Brandsma D, Stalpers L, Taal W, et al. Clinical features, mechanisms, and management of pseudoprogression in malignant gliomas. Lancet Oncol
. 2008; 9: 453–461.
61. Ruben JD, Dally M, Bailey M, et al. Cerebral radiation necrosis: incidence, outcomes, and risk factors with emphasis on radiation parameters and chemotherapy. Int J Radiat Oncol Biol Phys
. 2006; 65: 499–508.
62. Rahmathulla G, Marko NF, Weil RJ. Cerebral radiation necrosis: a review of the pathobiology, diagnosis and management considerations. J Clin Neurosci
. 2013; 20: 485–502.