home > ebr > autumn 2010 > reviewing for eligibility
European Biopharmaceutical Review

Reviewing for Eligibility

Gregory V Goldmacher at ICON Medical Imaging compares local readers to centralised services when choosing eligible patients for clinical trials

Determining subject eligibility for trials through a centralised reading service can offer trial sponsors significant advantages in cost savings and data quality. Local readers may vary in their understanding of trial protocols, have competing commitments, and be biased towards over-inclusiveness. Centralised services must combine accurate determination of eligibility with a quick turnaround time, which requires expert readers as well as a robust and flexible technology that can optimise data management, quality control and reader access.


Over the last two decades, diagnostic imaging has played an increasingly important role in clinical trials. One of the applications of imaging within trials is the determination of subject eligibility. Early phase trials that are focused on safety rarely rely on imaging criteria for inclusion, although the exclusion of patients with conditions that make them inappropriate for a trial occasionally requires imaging. However, trials in Phase II and later frequently use imaging for eligibility. In oncology trials, endpoints such as overall response rate and progression-free survival require an imaging determination of response or progression, and thus the presence of baseline disease measurable by some set of imaging criteria, such as RECIST, is included in the eligibility screening. For trials focusing on cardiac function, a determination of ejection fraction, whether by echocardiography, MUGA or cardiac MRI, is often required. In coronary artery disease trials, the degree of arterial stenosis may be determined by angiography. In stroke thrombolysis trials, computed tomography (CT) may be used to exclude haemorrhage and CT angiography to confirm vessel occlusion. In a peripheral vascular disease trial, angiography may be used to decide whether patients are eligible for other interventions. Other examples of trials which use imaging in eligibility determination include various neurological, musculoskeletal, wound healing and gynaecological indications.


One of the decisions that a trial sponsor has to make in designing the protocol for the clinical trial is whether image interpretations (or ‘reads’) deciding eligibility should be done at each site by the local investigators or their clinical radiology colleagues, or whether they should be performed by a central imaging core laboratory.

There are significant problems associated with relying on sites to perform eligibility reads. Criteria based on imaging can vary in important ways from trial to trial. For example, RECIST 1.0 and RECIST 1.1 differ significantly in the metrics used to determine tumour burden, particularly in their evaluation of lymphadenopathy. Moreover, various modifications of the RECIST criteria have been developed to reflect the unique needs of various types of tumour (such as mesothelioma or prostate cancer), and may be further modified to suit the requirements of individual trials. It is unreasonable to expect that readers at every site will know and be able to apply each trial’s imaging eligibility criteria consistently. At smaller sites, each reader may only encounter a few patients who are potentially eligible for a trial. At larger and more active sites, while there may be more eligible patients, there are often multiple trials in progress, with a different imaging criteria for each. Different readers at a particular site may also have inconsistent interpretations of the criteria and their application.

There are additional factors that add potential inaccuracy to local eligibility reads. Typical contracts between trial sponsors and sites specify that sites are paid for every randomised patient, even if they later turn out to have been ineligible. Thus, there is an incentive for local readers to be overly inclusive, and this can introduce bias into their interpretations. In addition, for a busy clinical radiologist the decision of whether a patient meets the criteria for inclusion in a trial is a minor consideration that competes for his or her attention with the large volume of other clinical work that awaits them.

There are, of course, some challenges that centralised eligibility reads have to contend with. It is an oft-repeated truism in radiology that the most valuable film is old film. In other words, when a finding on an image is ambiguous, one of the most helpful tools in narrowing the diagnosis is to search for prior studies to use for comparison. In one case I recently reviewed – a clinical trial scan performed to assess eligibility in an ovarian cancer trial – a cystic lesion with calcification was identified in the spleen. No other abnormalities were present on the scan. The appearance was consistent with an old haemorrhagic cyst, infection or metastasis. At the local site I might have checked whether the patient had any prior imaging. If a scan from the previous year, for example, had shown an identical lesion, it would have been possible to confidently declare it a benign incidental finding. As it was, the decision was much more difficult.

What are the factors that a core lab must focus on to ensure success in central eligibility reads? Clearly, it is necessary to have expert readers who have been trained to follow the specific eligibility criteria of the trial in question. However, another key factor is rapid turnaround. In order to add value to a trial, a central reading facility or core laboratory must be able to produce a result within a timeframe that allows decisions to be made about patient enrolment. Generally, the actual image interpretation portion of an eligibility read does not require a great deal of time. In contrast to a clinical radiology read or a full independent review within a trial, in which every portion of every image must be scrutinised to avoid missing potentially relevant findings, the read for eligibility is highly targeted. Usually, the radiologist is looking for a single feature or a small set of features. For example, if eligibility requires measurable disease by RECIST 1.0, any malignant lesion greater than 10mm in diameter (with slice thickness of 5mm or less) will satisfy the criteria. It is important to be able to distinguish malignant lesions from benign incidental findings, but most frequently, especially in late stage diseases, it is not difficult to find a qualifying lesion.

Most of the time required for the eligibility assessment, therefore, is the process of receiving the data, performing technical quality assessment, reconciling image and clinical data, and preparation of the data for reading. To perform these tasks, a core laboratory must have a robust infrastructure and process for importing and performing quality checks on image data. There must be a well-designed and smoothly running process for sites to deliver data to the core laboratory. The core laboratory has to have rapid quality control procedures, highly trained technical personnel, and a system for delivering feedback to the sites in case images have to be repeated. Once the images have been checked, there must be a system in place for presenting the images to a pool of qualified readers in ‘real time’, without having to wait for a batch of cases to accumulate. This is most easily accomplished if readers are alerted to the arrival of cases into an eligibility read queue, and can then log in remotely to read the images. The interface for such reads should be familiar, resembling as much as possible the user interface that the readers are familiar with in their clinical work.


Despite the challenges of executing a timely imaging eligibility review, it can markedly increase the fidelity of the data, permitting a more robust statistical outcome for the trial and saving the sponsor the costs of randomising ineligible patients. These benefits make centralised eligibility review an attractive option for those involved in trial planning.

Read full article from PDF >>

Rate this article You must be a member of the site to make a vote.  
Average rating:

There are no comments in regards to this article.

Gregory V Goldmacher is currently Assistant Director of Medical Affairs at ICON Medical Imaging in Warrington, PA, US. He received his undergraduate education from the University of Chicago, and his MD and PhD (in neuroscience) from the University of Texas Southwestern Medical Center in Dallas. He did his residency training in diagnostic radiology at Baystate Medical Center (Springfield, US) and Hahnemann University Hospital (Philadelphia, US), and research fellowships at Massachusetts General Hospital (Boston, US) and Thomas Jefferson University (Philadelphia, US).
Gregory V Goldmacher
Print this page
Send to a friend
Privacy statement
News and Press Releases

Tower to bring the temperature of pharmaceutical logistics to CPhI

Temperature-controlled containers that combine robust, reliable, and reusable benefits for the pharmaceutical supply chain will be showcased on the Tower Cold Chain stand, 91H66, zone ICSE, at this year’s CPhI worldwide, in Frankfurt.
More info >>

White Papers

The Power of Execution in Clinical Research


Medpace, long known for its excellence in conducting Phase I-IV studies on a global platform, is particularly noted for its therapeutic excellence. When that expertise is combined with operational precision, the results are nothing short of stellar. The Medpace full-service outsourcing approach – combined with therapeutic expertise, regulatory know how, and global presence – delivers clinical studies with an industry recognized track record.
More info >>




©2000-2011 Samedan Ltd.
Add to favourites

Print this page

Send to a friend
Privacy statement