Dynamic Contrast Enhanced MRI

Dynamic contrast-enhanced MRI (DCE-MRI) is a functional imaging technique performed by rapidly acquiring multiple T1-weighted MR images from a given tissue during injection of a paramagnetic contrast agent. The consequential tissue signal changes are quantified, and parameters with biological linkage are extracted. A review of the experiment with a focus on the analysis required is given by Parker and Padhani [127]. Further, standardized terminology and definitions of DCE-MRI-derived parameters have been defined by Tofts et al. [ 128]. Parameters typically extracted and used in the study of pharmacological response include:

• ^trans (min-1): the volume transfer constant between plasma and the extra-vascular extracellular space. The parameter is related to both flow and vascular permeability and, as such, change cannot be interpreted precisely. However, successful pharmacological intervention is expected to reduce this parameter by either flow and/or permeability change.

• ve (unitless): the volume fraction of the extravascular extracellular space. This is assumed to relate to interstitial space, although the relationship remains complex and posttreatment changes are not predictable.

• kep (min-1): the rate constant between the extravascular extracellular space and the plasma.

• IAUC: the initial area under the curve, typically over the first 60s. This is a semiquantitative expression of the initial contrast enhancement, which requires no data modeling. The relationship to physiology is complex [129] , although both flow and permeability are likely to affect the parameter, and successful antiangiogenic treatment is expected to reduce the value.

In the recommendations by Leach et al. [130] both Ktrans and IAUC should be used as primary DCE-MRI endpoints, with ve and kep as secondary endpoints.

As emphasized in the review by O'Connor et al. [131], DCE-MRI has been used widely for the study of angiogenesis inhibitors. Since this class of agents has received considerable attention across industry, interest in DCE-MRI has become widespread. The review by Jayson and Waterton [132] outlines the progress in use of the technique for drug development, focusing on fundamental aspects such as design of the protocol, measurement validity, and reproduc-ibility. Work focusing on DCE-MRI as a pharmacodynamic endpoint has been complemented by developing our understanding of the technique for more routine clinical use: for prognostication, for example.

The use of DCE-MRI offers significant advantages over other methods able to interrogate aspects of tissue vascularity. With careful implementation, multisite standardization can be achieved and sufficient reproducibility relative to typical pharmacodynamic effects can be realized. However, it could be argued that angiogenesis inhibitors either currently in development or marketed may have proceeded without DCE-MRI within the development program. So what does DCE-MRI add to the development paradigm, and why, if such a localized indicator of antiangiogenic activity exists does it not form a fundamental step in the development of such agents? First, there are technical challenges to successful implementation that form impediments to broader use. Second, there is still insufficient understanding of the relationship between given DCE-MRI parameters and both histopathology and clinical outcome. These uncertainties lead to difficulties in interpreting DCE- MRI studies. For example, how much of a change in a DCE-MRI parameter is enough to predict benefit?

Method Standardization and Deployment Challenges

Site Limitations DCE-MRI remains a specialized technique with limited routine application, and therefore many radiology departments will have limited experience with its use. It is important to consider the practicalities of consistent DCE-MRI implementation across multiple sites required for a given clinical trial. The first step requiring consideration is the selection of sites capable of performing DCE-MRI by adhering to a common acquisition for the study. One factor to consider is what experience site personnel have in performing DCE-MRI. However, this is not to say that sites with few experienced personnel are not capable of undertaking DCE-MRI, but the users must be committed to adhering to a prescribed protocol.

Statistically, it may be better to have fewer subjects scanned at a small number of sites where the imaging can be tightly controlled versus many sites with greater diversity of equipment and experience. A pharmaceutical company considering multiple studies could consider developing a network of specialized imaging sites able to perform consistent imaging and where good communication between sites can help optimize imaging.

Implementing DCE - MRI Within the Clinical Protocol A great variety of options exist regarding incorporating DCE-MRI within the clinical protocol. Some of the most important factors to consider are:

• Is a repeat baseline examination necessary to establish methodological reproducibility?

• When should posttherapeutic administration imaging be performed?

• What are the primary DCE-MRI endpoints?

These questions should be answered by considering what is known about the expected pharmacology of the drug and what has been done previously with similar agents. Important practical parameters include how many examinations will be acceptable to the subject (with a careful review of other required procedures) and whether close time points present scheduling challenges to the radiology department.

It has been recommended that reproducibility examinations be built into the study where possible in the form of a repeat baseline [130]. This would typically be done by performing two examinations within a week prior to beginning of therapy. An example of analysis of a repeat baseline study includes that by Galbraith et al. [133], where repeat examinations performed one week apart provided confidence intervals for Ktrans and IAUC90 (64% and 61%, respectively). Other examples include that by Roberts et al. [134], where investigators assessed repeat examinations from data across two centers in multiple tumor types, comparing compartmental modeling versus model-free analysis. The 95% confidence intervals for K*™8 and IAUC60 were calculated to be 35% and 55%, respectively. Clear differences in the reproducibility are apparent from the published examples and are expected to arise from differences in acquisition, modeling techniques, and assumptions in the model. This gives credence to the idea that a reproducibility assessment is often important, and that if performed it must be representative of subsequent study conditions. Although changes following administration of drug are often large relative to the methodological variability, in many instances it may be important to minimize the variability as much as possible. In a publication by Parker et al. [135] it is demonstrated that significant improvement in reproducibility of Ktians and other parameters can be attained by using a population-based arterial input function. Such improvements in reproducibility are likely to be important. In a study where dose is varied, using DCE-MRI to understanding the lower limit of biological effectiveness could be important, and the measurement reproducibility will have a direct impact on the ability to define subtle change at low doses.

Published DCE- MRI studies show a range of selected posttherapy measurement time points [131] . Between different drugs there will be significant dosing differences, pharmacokinetic properties, and differences in mechanism likely to lead to varied time scales of DCE-MRI response. Therefore, there is no consistent formula to select posttreatment time points. However, where studies do include an early time point (such as one or two days after therapy has begun), a large response is generally seen and is similar in magnitude to subsequent time points. The data by Morgan et al. [136], Mross et al. [137], and Thomas et al. [ 138] emphasize the early response (day 2) that can be measured consistently between studies, which then persists (out to day 28). It would thus seem that in most instances an early time point provides significant data to evaluate vascular response.

Choosing a Standardized Protocol A number of acquisition and analysis options exist to perform DCE-MRI. For any one study, consistent application for patients being investigated is an obvious requirement to keep variability minimized. However, for a given company there are clear benefits in maintaining a consistent acquisition and analysis protocol across multiple studies: reduced operational burden to develop acquisition guidelines and the ability to compare one study to the next potentially to compare the magnitude of response between different drugs, doses, or dosing schedules. There are also benefits to developing consistent practices for both academic- and industry-sponsored DCE-MRI-based studies:

• Increased awareness of a standard protocol to minimize the complexity encountered by radiology departments in dealing with many different DCE-MRI protocols for different studies.

• Potential to perform metaanalyses on data from multiple studies to query relationships between DCE-MRI response and clinical outcome.

Leach et al., on behalf of Cancer Research UK, proposed recommendations for performing DCE-MRI for studying antiangiogenic and antivascular therapies [130]. Their technique provides an academic and industry perspective on the key attributes of a standardized DCE-MRI acquisition. Owing to the differences in the way the technique is implemented by different scanners, different acquisition protocols will often be required from one center to the next. However, provided that cross-site implementation is undertaken carefully, a DCE-MRI measurement at one center should be comparable with that taken at another.

MRI technology changes at a significant pace, evidenced by multichannel technology and availability of scanners operating at static magnetic field strengths of 3T and above. Standardizing measurements from one study to the next is a challenge with a background of technological evolution. When a new technology (e.g., 3T MRI) is introduced, a careful comparison with existing technology is required to ensure equivalence of measurement.

Responsibilities for Deployment and Centralized Analysis Ideally, multiple sites with quantitative imaging expertise will be selectable to ensure that significant recruitment goals can be achieved in a reasonable period of time. Often, this may not be the case, and sites with limited imaging experience are required. For this reason, deployment of the DCE-MRI acquisition is complex, costly, and time consuming, requiring the following activities: review of the adequacy of scanner specifications, implementation of the MRI sequences required, scanner performance evaluation, training of site technologists, instructions to upload the data in a timely manner on scan completion, robust quality control steps, and centralized analysis. Often, a third-party company, typically a contract research organization with imaging expertise, can be given responsibility for these activities. Alternatively, one clinical site in a network can be responsible for defining the acquisition protocol and communicating imaging details with partnering sites. The latter is particularly applicable where all sites have expertise in quantitative imaging. Whatever the approach selected, it is important that data acquired be transported readily (ideally via electronic transfer) to a central analysis lab in order that any important deviations from the acquisition protocol can be identified to ensure future consistency.

Owing to the diversity of software platforms able to analyze DCE-MRI and the options available for analysis, it is imperative that data sets be reviewed centrally using a consistent process. This will ensure that variability in the analysis is kept to an absolute minimum. Typically, each analyzable lesion has an associated DCE-MRI parameter created (Ktians, IAUC, and others) following pixel-by-pixel analysis. A mean value associated with each lesion is typically assumed to provide the simplest output with more advanced analysis (such as heterogeneity analyses) potentially able to provide further insight into the data.

Toxicity of Gadolinium-Based Contrast Agents Data show that administration of gadolinium-based contrast media is associated with a risk of developing nephrogenic systemic fibrosis (NSF) [139-141]. Although understanding is still limited, it appears that risk factors for NSF include reduced kidney function; multiple exposures to gadolinium-based chelates; the use of linear chelates, indicating a potential relationship to thermodynamic stability [142]; and the presence of major tissue injury, termed pro-inflammatory conditions [143] .

It is clear that much is still to be understood about NSF and risk factors involved. However, it is prudent to minimize the risk of NSF for all subjects being exposed to gadolinium as part of a DCE-MRI examination or other MRI study based on existing data. DCE-MRI studies used for drug development differ from clinical practice and should approach the issue with particular caution, owing to the following factors:

• The need for multiple injections of gadolinium-based contrast agents within a short period of time, particularly if a repeat baseline and early posttreatment follow-up is required.

• For clinical diagnostic purposes the justification of gadolinium in high-risk subjects can be clearly defined, but for a patient in a clinical drug study the benefits of receiving multiple examinations are less clear.

The safety of the patient is paramount at the same time as acquisition of data to assist the development of novel therapies remains important. A conservative approach is recommended to minimize risk. This could include:

• Minimizing the number of DCE-MRI examinations while still being able to provide an understanding of the pharmacology

• Use of a single-dose gadolinium agent

• Careful monitoring of renal function and setting boundaries below which subjects should not be studied

• Use of contrast agents that, based on current data, appear to have the lowest association with NSF

• Continual review of emerging data with necessary regular revision of guidelines

DCE-MRI for Drug Development Decision Making As reviewed by O'Connor et al. [131] DCE-MRI has been used extensively to study multiple antiangiogenic and antivascular therapeutics. Many of the DCE-MRI experiments performed to date have been undertaken on drugs still in development. DCE-MRI certainly appears to be a methodology applied consistently within antiangiogenic development programs. Examples include the evaluation of AG-013736 by Liu et al. [ 144] . where a relationship was observed between drug exposure and change in both Ktrans and IAUC from baseline to day 2. DCE-MRI was also incorporated into phase I investigations of the kinase inhibitor (VEGFR, PDGFR, FGFR) BIBF 1120 [145,146]. Both studies reported DCE-MRI findings after baseline, early (2 days [145] and 3 days [146]), and later time points (28 days [145] and 30 days [146]). Although DCE-MRI was obtained at multiple drug doses, it is not clear whether there was any dose-response relationship to support a dose decision for future studies.

A methodology able to define dose versus vascular response for antiangio-genic therapies should in theory play a significant role in many drug development programs. There do remain, however, a number of significant challenges. First, most published studies where DCE- MRI has been applied to novel therapies have been undertaken on subjects with a range of tumor types, typical of the phase I setting. This creates methodological challenges likely to add to measurement variability (and measurement failure in a percentage of patients), and biological response will be less uniform than a study of a given tumor type. Another challenge is interpretation of data. How much of a change in ^trans is sufficient to predict subsequent treatment success? Without data linking DCE-MRI parameter change with metrics of clinical outcome in large studies, with different tumor types, and different therapies, it is not possible to establish a broad threshold of, say, the Ktrans change required to predict therapeutic success. If this could be realized, a firm understanding of radiological effectiveness from DCE-MRI could be used to inform dose decisions for a drug development program. Until such data are available, DCE-MRI is likely to be used as a supportive function contributing at least to dose decisions, evaluating early signals of efficacy, and understanding scheduling. The value that DCE-MRI brings, however, would increase significantly, and implementation cost would clearly be justified if there were greater understanding of different levels of DCE-MRI response.

Collins [147] has critiqued how imaging (both PET and MRI) contributed to decision making in the development of combretastatin A4 phosphate. Such analyses are important in order to understand how imaging has actually contributed to decision making, the many instances when it has not, and what can be done to enhance the role of imaging in clinical drug development.

Future Improvements in the Use of DCE-MRI As described previously, published recommendations will help to encourage consistent DCE-MRI measurements. These should enable pharmaceutical companies and academic groups to focus on the challenges of drug development over methodological optimization. However, the current recommendations that exist remain appropriately broad since they require applicability to scanners from different manufacturers. Greater harmonization of methods between scanner manufacturers would considerably benefit the DCE-MRI experiment, lowering the barrier to implementation and enabling broader use in drug development.

Increased standards of measurement would deliver more studies that could be subjected to metaanalysis: for example, to study DCE-MRI response versus outcome. With such relationships understood, the role of DCE-MRI in early drug development can potentially move from an interesting secondary objective to delivering primary study goals.

Clinical trials are also likely to benefit increasingly from the multiparameter assessments possible within a single MR scan session. In a study of the pan-VEGF receptor tyrosine kinase inhibitor AZD2171 in glioblastoma patients, Batchelor et al. [148] demonstrated that an extensive MRI evaluation provides significant benefit over DCE-MRI alone. The study incorporated lesion volume and apparent diffusion coefficient measurements, together with DCE-MRI - derived indices, including Ktrans and relative vessel size estimations. Such approaches are likely to be necessary to fully interpret the complex changes occurring following antiangiogenic therapy and to promote understanding of drug scheduling and how optimally to combine different therapies. Improved standardization will both enhance the ability to incorporate DCE-MRI into drug development and promote understanding of how different DCE- MRI response levels are interpreted.

Project Management Made Easy

Project Management Made Easy

What you need to know about… Project Management Made Easy! Project management consists of more than just a large building project and can encompass small projects as well. No matter what the size of your project, you need to have some sort of project management. How you manage your project has everything to do with its outcome.

Get My Free Ebook


Post a comment