Identifying and reducing sources of variation in quantitative and dynamic molecular imaging

Learning Objectives Help researchers and clinicians become more aware of, understand, and reduce or eliminate drug delivery related sources of variation in their research and clinical practices.

The Need The SNM and its members recognize that research into new procedures, drugs, and equipment advances the practice of medicine. The SNM Clinical Trials Network is striving to standardize imaging protocols across multicenter clinical trial sites. This standardization helps to reduce variation, potentially reducing the number of patients required for a clinical trial and thus saving costs. This also helps avoid the situation where a clinical trial fails because the beneficial effects are swamped by the variation. In clinical oncology practices, the key questions are becoming “Is a tumor responding to the current therapy?” and “Will a tumor respond to a proposed therapy?” This requires assessments over time where the variations in the studies do not overwhelm the patient’s response to the treatment. Sources of variation related to drug delivery and the time course of drug delivery to and through the patient are often underappreciated or overlooked in molecular imaging research. A hand injection was good enough in the past. However, these sources of variation are becoming much more important as the questions being asked become more sophisticated. Awareness is the first step to a solution. Understanding the sources of variation can lead to their mitigation or elimination in the planning stages of a research protocol or clinical practice. The Content And extensive mind map of sources of variation and a table of effects are presented. This information is compiled from many years of SPECT, PET, CT, MR, and therapeutic fluid delivery research and clinical practice observations, as well as reasonably expected patient variation and operator limitations. Several strategies to eliminate or mitigate the identified sources of variation are also presented. Actual improvements achieved by employing selected strategies will be illustrated for SPECT, PET, and other imaging modalities