Today's imaging clinical trials are complex, often spanning the course of months to years, involving geographically dispersed sites, utilizing a variety of medical imaging equipment and relying on expert medical and technical personnel. To ensure high quality imaging data from clinical trials, standardization of image acquisition (e.g. patient positioning, anatomical coverage, and instrument settings) and quality control need to be established, monitored and enforced.
With advanced medical imaging a routine part of evaluation of today's therapeutics, sponsors are turning to Imaging Core Labs (ICLs) for the management and independent review of clinical trial data. Imaging core labs can provide end-to-end support for sponsors (study initiation through to data archiving and regulatory submission) as well as validated methodologies and services (e.g. guidelines, training, and site visits) to obtain high quality imaging data with less variance and ensure delivery of precise endpoints.
As part of their commitment to generating high quality data, ICLs have developed and implemented performance metrics as a way to track and evaluate image quality and identify site performance issues early that require immediate remediation. Two important performance metrics, the percent of site queries and the percent of non-evaluable images, are discussed in more detail below.
During the course of a clinical trial, imaging data arrives at the ICL on a regular basis and is evaluated for image quality by radiological or other imaging technologists prior to being sent for independent review. When there is an issue with a piece of data, a query can be generated and sent to the study sites electronically for quick resolution.
If the query rate in a study exceeds a certain threshold, it may indicate insufficient site training or significant deviation from study protocol with regard to image acquisition. In this scenario, ICLs can take special measures to reduce the number of image queries including implementing specialized site training. These types of measures are often sufficient to significantly reduce the number of image queries, maximizing the amount of usable data and contributing to overall trial efficiency.
Percent of Non-Evaluable Images
For multisite studies, varying technical capabilities at hospitals and imaging facilities can result in sub-optimal imaging data. This can be especially problematic when sites are unable to provide digital images in the required format, increasing the incidence of non-evaluable images, which require repeat scans (if possible) or contribute to a pool of missing data for the study.
By tracking and monitoring the incidence of non-evaluable images, an ICL can work together with sponsors and sites to identify locations where subjects can receive adequate scans that comply with the study protocol. Furthermore, the ICL can provide additional training to optimize parameters for digitized images.
The metrics described here provide an example of ways in which the operational capabilities of an imaging core lab can provide improved study outcomes through decreased data variability. These tools contribute to greater statistical confidence in study findings and to increased clinical trial efficiency while providing significant cost savings for sponsors.
To learn more about the use of performance metrics in clinical trials and read about specific case studies in which these metrics were applied, see my recent article “Metrics in Medical Imaging” in The Monitor.