Medical imaging in late phase clinical trials has evolved from paper trail nightmares to concise electronic transmission.

As the medical imaging industry expands beyond current therapeutics and into new therapeutics, it is increasingly important that radiology sites are able to provide the appropriate reproducible and high standards for image quality with reasonable effort throughout imaging trials. Leveraging available technology and systems to support training, communication, and image transfer will aid sites and study teams to improve image submission, completeness, and quality.

Medical imaging studies are heavily dependent on a manual process workflow, available trained resources, and available technology to support the required functions of a clinical trial. In the world of clinical trials, medical imaging may be required for safety evaluations, or may support the primary or secondary endpoints for a specific therapeutic indication. That being said, it is critical to the success or failure of a clinical trial that the image acquisition and image transfer occur in a well-controlled and timely manner at each radiology facility to ensure image data integrity. It is common practice for several clinical studies to be running simultaneously at one investigator’s site where imaging may or may not be a component. The imaging component of a clinical trial is only one component of a much larger picture; therefore, it is vital that the contract research organization (CRO) community continue to evaluate more efficient and reliable means for image capture and transmission from the radiology facilities.

On average, three to five imaging visits per patient are common for a mid-size oncology study. Typically, radiology sites are instructed and trained to ship each imaging visit, via courier, to the central imaging lab within 48 hours after acquisition. This allows time for the core lab to perform adequate review and provide immediate feedback to the radiology facilities. Feedback from the lab is a crucial step in the process, where this is the mechanism to raise the awareness of the radiology site that there was an acquisition deviation for a particular visit. The remediation at this point could be a complete rescan (which is not typical due to per protocol visit window limitations) or performing appropriate scanner corrections for future timepoints and patients. The radiology site may delay sending image visits within that 48-hour window, and may further complicate the issue by batching multiple imaging visits or patients in one courier shipment. In this instance, the chain of custody is broken. The ability to provide rapid, high-quality feedback from the imaging core lab to the radiology site has now been lost. This metric may even be grimmer depending on what portion of the world the images are traveling from. Shipment in batches has proven to be disadvantageous in terms of the final quality of the imaging data assessments. If you consider all the possibilities, delays of 1 to 2 months or greater are possible for the independent review process.

While the importance of timely image acquisition and receipt remains a high priority, the consistent, accurate completion of the associated paperwork that accompanies each imaging visit is also a critical step. Radiology sites should follow an acquisition protocol that includes various parameters, such as the usage of contrast media, the anatomy to be covered, the image resolution, etc. Such information is captured in a study-specific image transmittal form (ITF), which is sent along with the imaging visit. Prior to the imaging visit being shipped, the site completes the ITF and archives a copy. Once the imaging package is received at the central imaging core lab, personnel would then perform data entry of all relevant information into a study-specific imaging database.

An incomplete or inaccurate ITF is yet another area for a gap to occur that could cause significant delays with the independent review process. If one could imagine how long it took for images to arrive at the imaging core lab, how long could it potentially take to have a question resolved? Chasing down missing packages, inquiring about late scans, following up on outstanding queries, tracking down missing ITFs, sorting out the imaging visits and ITFs if sent in a batch; these activities cost a great deal of time, resources, and money and may ultimately affect the quality of the data.

These delays are not acceptable at many levels; however, there may be a more significant impact to one imaging study than another, which could ultimately cause the start-up of a study to fail.

The majority of imaging trials occur over a longer period of time and the independent review process is spread out; the actual analysis of patient data may only occur once patients have come off study (ie, progressed, died, etc). In these cases, there is some wiggle room for delay, but not much. On the other side of the spectrum, there are imaging studies that may require eligibility reviews, in which an imaging core lab plays a specific role regarding whether a patient will be included in a clinical trial. All clinical studies contain a predefined list of criteria that may include or exclude a patient from a study; for example, the patient must be 18 years of age or more, or certain medications or procedures may exclude that patient from participating. The inclusion/exclusion criteria are very project specific and are tailored to the compound/ device being evaluated. The central imaging core lab will have an independent physician review the images and submit their inclusion/exclusion findings directly to the site, a CRO, or the sponsor. The turnaround time for these studies from image acquisition to image shipment, image receipt, image processing, image queries, and independent review must be measured in days, on average 7 to 10 days for the entire process. Therefore, the challenges described above are a recipe for disaster, as each step of the eligibility review process needs to be well orchestrated and for the most part fl awless. The world of image receipt as we know it needs to be changed and the imaging community should consider the available technologies.

Electronic Data Capture (EDC)

EDC has been talked about for many years; the earliest in-house pilot studies date back to the late 1980s. In the fi rst generation of EDCs, companies preconfi gured personal computers with EDC software and protocol defi nition, and shipped them to investigator sites. The data was entered locally and transmitted to a central database using a dial-up connection.

In the late 1990s, the Internet solved the problem of technology distribution, working with only one copy of the data collection software, the study protocol, and the patient data. All the investigator needs today is a standard PC with a Web browser and connection.

EDC in clinical trials offers several advantages to the paper process:

  • Data validity: If study patients must be between 18 and 65 years old for inclusion, an eCRF can quickly identify where an investigator is mistakenly recruiting an under- or over-aged patient and prevent this occurring.
  • Range checks: An eCRF can use simple data range checks to identify data that looks invalid. For example, a patient appears to be only 1.85 feet tall because an investigator ticked the “feet” instead of the meters units box on their paper CRF.
  • Dynamic eCRF: An eCRF can be dynamic, so that, for example, when an investigator indicates that a patient is male, questions concerning pregnancy, childbirth, and lactation can instantly be removed. This prevents the occurrence of impossible data where a human investigator may have accidentally ticked an inappropriate box.

eCRFs in Imaging

Technology also affects the way an independent image assessment of study images is performed. There are three main technology-supported components: image visualization; support of the assessment, eg, to measure an anatomical structure; and capture of the results via an eCRF closely shadowing the implicit workflow of the evaluation criteria of choice. These three components are highly integrated. This said, it is obvious that an independent review assessment can not be done in a clinical EDC system, which is mainly designed for the clinical portion of a trial. However, integrating EDC and imaging works well for investigators when entering data and the back-office integration, including randomization data.

EDC usage is a good prerequisite for a site to receive also the allowance from their local IT to use the Internet (via sFTP or HTTPS) to upload the image data to servers, which can be accessed by the imaging CRO. Besides many advantages, there are a few challenges to be addressed proactively.

The image upload tool will, in the near future, not be integrated in the EDC. Therefore, specific upload accounts/ per site need 1) to be managed by the imaging CRO and 2) the site users need to deal with an additional password.

The performance perception of the users needs to be managed. To pack the CD into its padded envelope and fill out the shipment information will take more overall time than logging in and uploading. But uploading of image timepoints of 20-100MB might take a short while depending on the provided network/Internet connectivity.

The reality in clinical technology infrastructure setups is that, unfortunately, we might need support of three groups of players to make this happen: clinical network experts, clinical data security officers, and the radiology department’s technology specialists. Obviously, if you rely on the collaboration of these groups, which might compete in their daily lives, potential challenges can be expected.

Among the advantages of the electronic transfer are cost savings due to reduced shipment costs; an intelligent workflow to upload, improving accuracy; a trackable chain-of-custody; and avoiding lost or damaged images due to shipping.

Future Vision

In the future, integrated applications for clinical and imaging data will be key to speeding up the process, increasing quality, and decreasing costs. Integration will include:

  • providing centralized access control;
  • using the DICOM header to perform a first QC when uploading images;
  • a check for certified scanners and technical compatibility;
  • follow-up on missing timepoints;
  • linking the completion of a visit in the EDC system with a successful submission of image data.

The usage of an EDC system for clinical data is an opportunity to use the EDC infrastructure for electronic image upload and to change the current paper ITF process to an integrated, digital process. Early case studies show positive results. With increased available bandwidth and the users’ access to online applications, image collection as denoted today will evolve toward an electronic image submission with integrated quality control.

Gunter Bellaire is director of operations, medical imaging, and Kevin Jaynes is program director for Perceptive Informatics, which is a subsidiary of PAREXEL International, a global contract research organization.