Hstorically, the endpoint gold standard for demonstrating clinical benefit in oncology drug trials has been and still is overall survival (OS). From the FDA perspective, OS provides the most objective and comprehensive means of evaluation, but the analysis is not without its drawbacks.

Because evaluation of OS requires large patient numbers and much longer follow-up compared with other endpoints, data receipt is delayed. Trials based primarily on survival are therefore always more expensive and time consuming, and FDA approval takes longer than when based on studies using other accepted primary efficacy endpoints, such as time to progression (TTP) or objective response rate (ORR). Plus, overall survival may be confounded by subsequent therapies, unlike endpoints such as TTP or ORR, which are often used as surrogates for OS in patients with solid tumors.

These tumor assessment endpoints are less expensive and much more time effective than OS, and are, therefore, an attractive alternative for demonstrating efficacy. Because they are based on subjective evaluations (i.e., clinical or radiological observations), however, they are also less objective than overall survival and must be used in accordance with strict quality control measures to ensure validity.

Reducing bias

Efforts to encourage uniform interpretation of endpoint data among assessors and to minimize assessment variability across the study should be part of the study design prior to patient enrollment. A good example of a design intended to reduce variability of endpoint assessments is a double-blind read with adjudication. By this process, radiological images are initially assessed by multiple, often two, independent blinded reviewers. If the reviewers disagree, then a third adjudicating assessment (i.e., a tie-breaker) is made by an additional reviewer who may or may not be blinded to prior assessments. To further minimize variability, all independent reviewers should be uniformly trained and periodically tested to evaluate intra- and inter-reader variability.

Implementation of double-blind read with adjudication in Phase III oncology trials became more widespread in the late '90s based, in large part, on the need for accurate, unbiased radiological assessment in patients with solid tumors. This need arose partly because of the addition of Subpart H to new drug approval regulations promulgated in 1992, which allowed for accelerated drug approval using surrogate endpoints involving radiological assessment of tumor sizing. By the end of the millennium, however, it had become apparent that although endpoints such as ORR can support regulatory approval in select solid tumors, relief of tumor-related symptoms and drug toxicity should also be considered when making approval decisions.

Moving into the 21st century, solid tumor trials started using independent endpoint assessment committees (EACs) that used the data from the independent imaging assessment committee in conjunction with information from sites. Although these endpoint assessment and adjudication committees use a combination of independent blinded imaging data and site data, their goals and processes are the same as the independent imaging assessment committee, and their determination of response is often positioned as the primary efficacy result of a Phase III trial.

This is particularly important because tumor response alone is not always sufficient data for regulatory review. The FDA and other regulatory agencies are looking increasingly to a more objective and encompassing view of the drug's effect, and the use of the EAC offers an evaluation of this efficacy. Data for regulatory submission is not only comprised of bias-free tumor assessment with standardized quantification (e.g., WHO or RECIST criteria), but must also include clinical and pathology data to confirm the radiological analysis.

Not only do EACs reduce physician bias (blinded reviews coupled with off-site specialists serve to eliminate the potential for prior knowledge of the status of the patient) but they also help to overcome the limitations inherent in imaging that may make precise diagnosis difficult. Especially valuable for subjective endpoints, they provide a far more complete method for determining total tumor and drug effect when a complicated definition of efficacy is required.

Still relatively novel, EACs are most prevalent in oncology trials, although the FDA's push toward more robust and definitive datasets for regulatory approval means that they will become important to all therapeutic areas in the future.

Current FDA efforts to draft guidance documents for EACs closely resemble those for cancer endpoints in clinical trials, including survival, tumor response, response duration, relief of symptoms, or composite clinical benefit endpoints. The FDA has intended to develop the series of guidances on oncology endpoints through a process that includes public workshops and discussions before its Oncologic Drugs Advisory Committee, and each subsequent guidance is to focus on endpoints for specific cancer types to support drug approval or labeling claims.

Building the EAC

For oncology trials, the EAC is defined by the oncologist(s), who utilizes data beyond radiology images, including clinical data, pathology results, and other physical measurements (e.g., skin lesions for cutaneous lymphoma, etc.) to more holistically determine the disease status of the patient. Specifically, board-certified oncologists are required, but it is also often of value to have radiology and pathology representation on the committee. All members must be trained under an EAC chair to minimize variability among assessments and to determine the different roles each clinical specialist plays in them.

The EAC chair should be a member of the committee and will be the point of authority for determining information needed for assessment (i.e., adverse events, histology, calibrated photos, magnetic resonance imaging, etc.). They should work with the imaging CRO, sponsor, FDA, and radiologist to design the EAC charter, which is essential to ensuring unbiased assessment of all clinical outcomes. Specifically, this responsibility includes development of a decision algorithm for response assessment and determination of work load and time estimates for all committee members. Once the charter and training and testing programs have been designed, full training commences and the program is put into operation.

Objectivity of the earlier mentioned guidances and workshops is especially critical in fields in which endpoints are determined by clinical or radiological examinations, rather than by predefined laboratory measurements. This is where an independent EAC can be invaluable. Through a charter, EACs can facilitate an unbiased assessment of clinical outcomes in a multicenter Phase III oncology trial to enable a uniform, controlled-assessment process that is independent of the sponsor. Most importantly, the endpoint assessment and adjudication committee charter enables objectivity of endpoint assessment and approximates the critical review that the FDA might impose on clinical endpoint data. The endpoint assessment and adjudication committee charter should be reviewed rigorously in consultation with the FDA, resulting in a charter that may prove to be a useful example in generating independent assessment of pivotal clinical outcomes.^1-7