| In adopting a medical imaging strategy, Gunter Bellaire of Perceptive Informatics assesses the potential variability of imaging data collected in multicentre trials, and the variability introduced when interpreting that data
Imaging modalities such as computer tomography (CT) and magnetic resonance imaging (MRI) can provide insight into drug efficacy and safety in a non-invasive way that often provides a faster way to reach traditional clinical endpoints (1).
In the past, even for pivotal studies, only site assessments of imaging data (local image review) were taken into account for study result analysis. However, in recent years, the value of a centralised imaging review has been acknowledged and continues to grow in popularity, making it now the method of choice, particularly for pivotal studies (2). In particular, the standardisation of image acquisition, and independence from clinical patient information, supports the objectivity of image analysis.
The independent centralised review (ICR) process aims for the reproducible outcome of expert image review (3,4). There are ways to assess reproducibility, such as comparing the variability among multiple expert reviewers, or looking at the variability of a single reviewer over time. The features of the central review that contribute to reproducibility are: standardised image acquisition; site and reviewer training; and tools to support the review process. This articles focuses on these aspects and discusses the potential effects attributed to each one.
STANDARDISED IMAGE ACQUISITION
Imaging modalities nowadays are highly complicated devices with various degrees of functioning/flexibility. The overall imaging modalities of choice are usually defined in the study protocol. In a multicentre clinical trial, the local imaging standards and the variety of available modality generations and individual software packages (5), which are defined for local usage, require a considered decision to provide guidance for image acquisition that meets the needs of such a trial. |
