Numerous elements of PET imaging are subject to variability, such as scan protocol, data processing and analysis methods. However, reducing variability as much as possible is essential for ensuring that the data acquired for the trial is accurate and robust.
In this blog post, we will focus on the scanner itself and the strategies used to reduce variability among scanners in a multi-center trial.
Quantitative imaging parameters, such as standardized uptake values (SUVs) and metabolic active tumor volume (MTV) are frequently used as imaging biomarker endpoints in a multi-center study. Implementing stringent quality control (QC) for all aspects of the trial can therefore provide confidence that such parameters are comparable among patients and sites, especially for longitudinal studies, irrespective of the PET system used.
For example, a study assessed SUV errors over time for six dose calibrators from two major manufacturers and for six PET (plus CT) scanners from three major manufacturers. The authors recorded errors in SUV measurements ranging from −20% to 47%. Moreover, bias in SUV measurements varied over time and between scanner sites.
As that study shows, cross-calibration of PET scanners across sites as a QC step is mandatory, as PET scanners are likely to have different manufacturers, technical specifications and usage according to local practices. QC of PET scanners can validate that the system is operating correctly, minimize the occurrence of artifacts, identify and resolve potential problems early and eliminate unnecessary repeat scans.
Several guidelines are available to help implement QC processes. For example, the Quantitative Imaging Biomarkers Alliance (QIBA) and the European Association of Nuclear Medicine (EANM) have both published guidance for quantitative FDG-PET imaging in oncology.
In the QIBA guideline, QC includes selecting and qualifying a PET imaging facility, imaging personnel and PET scanners and ancillary equipment. In the EANM guideline, QC and inter-institution PET system performance harmonization includes ensuring a standardized inter-institution calibration procedure to facilitate the exchangeability of SUVs between institutions.
Meanwhile, a recently published article provides practical actions, including a detailed checklist with parameters referencing the DICOM tag information, for this step.
It might seem that the radiologist is the only one who would need to know about calibrating PET scanners. However, this process involves several stakeholders, including:
All these players need to understand the importance of ensuring that the scanner is acquiring accurate data, as well as what to look out for in the data to ensure that QC procedures have been followed. For example, to ensure comparability of SUVs, the minimum set of QC procedures recommended by the EANM include:
The QIBA guidance provides more detailed procedures, including a “Scanner Checklist” that can be used to ascertain a PET scanner’s qualification for quantitative imaging according to its guidelines.
Accreditation schemes, such as the one offered by EANM Research Ltd (EARL), or the Clinical Trials Network (CTN) Scanner Validation Program from SNMMI, are available to provide further confidence that a site is suitable for a clinical trial using PET-based imaging biomarkers as an endpoint.
Recent developments in PET technology, such as time-of-flight and resolution modeling during reconstruction, point spread function reconstruction, digital PET detectors and integrated PET/MRI systems, are leading to efforts to ensure that standards across sites are harmonized.
Care must be taken to ensure that the quantification of the imaging biomarker (be it the SUV or MTV) is comparable across systems. Some might question whether it is possible to balance optimization (i.e., technological advancements) and harmonization (i.e., cross-calibration of different PET systems across multiple sites). A recent editorial suggests that it can be possible.
Navigating the maze of imaging in clinical trials can seem overwhelming. Our operational experts can help in all aspects, including whether the site is equipped with the appropriate hardware and processes to manage images as per regulatory recommendations.