Brain tumor clinical trials imaging: a (well-standardized) picture is worth a thousand words

FDA approval of oncology therapies requires demonstration of direct clinical benefit measured by improvement in how patients “function, feel or survive.”¹ Prolonged overall survival has long been the preferred benchmark for establishing patient benefit; however, outcomes such as progression-free survival and tumor response have been recognized as acceptable surrogates for clinical benefit in the appropriate settings. Historically, FDA has supported the use of imaging as a tool to expedite the drug and device development processes²; in fact, most accelerated approvals of cancer drugs have been granted on the basis of durable objective responses in a refractory setting, including approval of therapies for patients with high-grade gliomas. Assessment of response and progression in neuro-oncology has been particularly challenging given the unique biological, physiological, and anatomic characteristics of the brain. Moreover, determining disease status requires appraisal of not only the tumor, but its vasculature and neighboring reactive cells. The advent of MRI provided exceptional visualization of neuroanatomy and pathological processes such as ischemia; paradoxically, it also increased uncertainty in the ability to interpret radiographic changes.

The utility and limitations of imaging as an endpoint in neuro-oncology clinical trials were discussed at the recent Brain Tumor Clinical Trials Endpoints Workshop co-sponsored by FDA and the Jumpstarting Brain Tumor Drug Development Coalition.³ Variations in imaging acquisition and processing were identified as key impediments to accurately determining disease response, as multiple factors influence MRI quality, including magnetic field strength, echo time, repetition time, voxel dimensions, signal averaging, acquisition orientation, dose of the administered contrast agent, and time delay prior to scanning, as well as the selection and implementation of any postprocessing algorithms. Consequently, standardization of imaging for multicenter neuro-oncology trials emerged as a leading action item aimed at reducing inconsistencies due to heterogeneity of MR scanners and image acquisition parameters. This initiative is in keeping with principles endorsed in FDA's Draft Guidance for Industry Standards for Clinical Trial Imaging Endpoints, first issued in 2011 and revised this year.⁴ This guidance differentiates imaging used in clinical trials intended to demonstrate clinical benefit from imaging used in medical practice and suggests that standardization elevates the bar for studies intended to support the approval of drugs or biologicals: “Imaging process standards help sponsors ensure that imaging data are obtained in a manner that complies with a trial's protocol, that the quality of imaging data is maintained within and among clinical sites, and that there is a verifiable record of the imaging process. Minimization of imaging process variability may importantly enhance a clinical trial's ability to detect drug treatment effects.”⁴ The guidance does not provide specific instruction on how to set imaging standards, as this is beyond the purview of the FDA. Decisions about imaging protocols, comparability, quality, cost, and imaging technologies are best left to disease experts and trial sponsors to develop in the context of specific disease settings.

Accordingly, the Brain Tumor Imaging Standardization Committee was formed with the purpose of drafting a standard clinical trial imaging protocol. Membership included international representatives from academics, industry, clinical institutions, and government. In this issue of Neuro-Oncology, Ellingson and colleagues present the results of this cooperative effort to craft consensus recommendations for a standardized Brain Tumor Imaging Protocol (BTIP). Building on work from the European Organization of Research and Treatment of Cancer (EORTC) brain tumor imaging standardization efforts, and borrowing from the experiences of the Alzheimer's Disease Neuroimaging Initiative (ADNI), the authors have developed guidelines for pragmatic image acquisition recommendations, including minimum sequences to be obtained, acquisition parameter selection, and contrast agent dose and timing. Mindful of the competing priorities of multiple stakeholders, the authors strived to balance quality with practicality in designing a blueprint suitable for use in large clinical trials that would also remain feasible for general practice. The result is an efficient protocol distilled to the essential parameters required to ensure adequate standardization that also approximates what is used in current clinical practice. FDA congratulates the authors for this accomplishment, achieved in the spirit of collaboration and completed in a remarkably short period of time.

Efforts are now under way to explore standardization of imaging protocols for techniques such as diffusion and perfusion MRI. Interest in these MRI technologies has fueled rapid advancements in this dynamic and maturing field. Certainly data acquired from more advanced techniques could provide valuable information and strengthen confidence in imaging endpoints; however, as one ventures further afield from basic imaging techniques, caution must be applied in making premature assumptions and one must recognize that presently data from these studies are largely for exploratory purposes rather than confirmation of clinical benefit. As standardization of imaging acquisition and processing is realized, what remains to be seen is whether this progress will translate to a meaningful improvement in the ability to assess progression and response in high-grade glioma. The theory of constraints tells us to first identify the most important limiting factor in demonstrating the effectiveness of therapies, and then work systematically to improve that constraint. Although lack of standardization of imaging has been cited as a critical factor undermining our confidence in interpreting radiographic data, it is not evident that it supersedes other obstacles such as variability in qualitative image interpretation and lack of unambiguously effective therapies. Therefore, from a regulatory standpoint, standardization efforts may approach a point of maximal value and diminishing returns when considered in the context of all other existing limitations. Finally, the neuro-oncology community should not lose sight of the fact that evidentiary standards for drug approval in the US are based on the ability to demonstrate a direct clinical benefit to patients; therefore, all proposed endpoints, imaging or otherwise, must be judged on a continuum from direct evidence to diminishing evidence of clinical benefit.

Conflict of interest statement. Dr Sul and Dr Krainak served as ex-officio members of the Brain Tumor Imaging Standardization Committee.