The Value of Standardized Imaging Protocols in Oncology Clinical Trials

Joseph Pierro MD and David Raunig PhD | |

Our past blogs have highlighted the implications of conducting clinical trials in a highly regulated industry and how poorly designed or executed clinical trials may lead to underpowered studies, often resulting in the drug development program failing to achieve regulatory requirements.

In oncology trials, which are particularly time- and cost-intensive, robust trial design is even more important. Not only do sponsors have more at stake; patients’ lives often depend on receiving the right therapies as quickly as possible. A lack of imaging protocol standardization can negatively impact image contrast, lesion conspicuity, anatomic area coverage, and more. Standard of care imaging protocols vary by institution and equipment manufacturer which may lead to non-uniform image collection (i.e. lack of standardization)  leading to non-compliance not only with the study design requirements but potentially with relevant regulatory guidelines and statutes as well.

This blog will focus on elements of a protocol that a pharmaceutical developer can identify and standardize, and the importance of establishing and communicating clear, thorough imaging protocol standards.

The FDA accepts overall survival (OS) as the gold standard endpoint in oncology clinical trials. However, these trials are lengthy and more costly to conduct than clinical trials relying on endpoints which may predict treatment response or progression earlier (e.g. lesion measurements assessed with RECIST 1.1 to determine progression free survival (PFS), or Date of Progression (DOP)).

Over the last few decades sponsors have used these endpoints successfully and the issues of independent reader variability has already been discussed in prior blogs. This blog will discuss the topic of imaging protocol standardization, the lack of which has been shown to have an impact on image contrast, lesion conspicuity, anatomic area coverage, etc.

Over the years, many professional groups such as QIBA (Quantitative Imaging Biomarker Alliance), ACR (American College of Radiology), and FDA have focused on the goals of imaging data standardization and harmonization. Their intent is to reduce technical or quantitative variability that may be introduced by different site-based imaging equipment, such as:

  • MRI field strength
  • Manufacturer-specific imaging acquisition sequences
  • Reconstruction algorithms
  • Gradient strength/designs
  • Calibration protocols, etc.

Standardization of image data “will assure that imaging results are comparable from site-to-site, time-to-time, and patient-to-patient” 1 and provide image consistency, better data quality and potentially better study outcomes. Additionally, uniform datasets provided as a result of standardization (big data approach) would enable interrogation via machine learning algorithms to help radiologists detect imaging differences earlier and more consistently, improving diagnostic performance.

Elements of standardization include [list is illustrative and not all inclusive]:

  • Defining which imaging equipment will be used in the trial and assessing scanner performance. This may include the use of imaging phantoms (loosely defined as a device with a defined set of physical properties or characteristics) that are periodically scanned throughout the study to evaluate and monitor scanner performance, resolution, accuracy, reconstruction/correction algorithms, etc.
  • Analyzing the scanners and site-based software used in quantitative image analysis and standardizing image processing and analysis
  • Ensuring the same patient preparation, image acquisition, imaging pharmaceuticals (contrast agent, radiopharmaceutical, dose, timing, etc.) and processing parameters are consistent across all clinical sites with dedicated site training and coordination with their radiology/nuclear medicine departments. These efforts have additional importance in molecular imaging (PET) in standardizing the collection and quantification of standard uptake values
  • Ensuring site personnel understand the protocol and perform all of the imaging required following the protocol without making changes that could result in nonstandard data collection, and that personnel understand the potential regulatory implications for protocol noncompliance
  • Providing sites with imaging manuals including detailed description of the imaging protocol, measurements, assessments etc. when protocols only include general descriptions that do not ensure standardization
  • Establishing methods and standards, for image data transfer/encryption and ensuring sites comply with appropriate regulations on data privacy rules

Our experience has shown that differences in how imaging is performed across various clinical sites may result in potentially significant differences in the appearance of lesions, and may limit imaging assessments or even result in non-evaluable data points.

Robust and quality data improves a pharmaceutical developer’s ability to make decisions with confidence. Harmonizing imaging protocols and standardizing analysis in oncology trials will reduce site-induced effects of variation (i.e. scanner and imaging protocol related differences), produce higher quality efficacy data, and potentially accelerate the therapy’s study lifecycle.

 References

  1. https://www.whitehouse.gov/wp-content/uploads/2017/12/Roadmap-for-Medical-Imaging-Research-and-Development-2017.pdf

Joseph Pierro, MD, is the Medical Director of Imaging at ERT and David Raunig, PhD, is the Senior Principal Imaging Statistician at ERT.

 

 

Share