Centralized Clinical Trial Imaging Data Management: Practical Guidance from a Comprehensive Cancer Center’s Experience

Abstract

Medical imaging is an integral part of clinical trial research and it must be managed properly to provide accurate data to the sponsor in a timely manner (Clune in Cancer Inform 4:33–56, 2007; Wang et al. in Proc SPIE Int Soc Opt Eng 7967, 2011). Standardized workflows for site qualification, protocol preparation, data storage, retrieval, de-identification, submission, and query resolution are paramount to achieve quality clinical trial data management such as reducing the number of imaging protocol deviations and avoiding delays in data transfer. Centralization of data management and implementation of relational databases and electronic workflows can help maintain consistency and accuracy of imaging data. This technical note aims at sharing the practical implementation of our centralized clinical trial imaging data management processes to avoid the fragmentation of tasks among various disease centers and research staff, and enable us to provide quality, accurate, and timely imaging data to clinical trial sponsors.

Background

Medical imaging is an integral part of oncology clinical trials that help bring novel cancer drugs from discovery to market. The Dana-Farber Cancer Institute (DFCI) is a National Cancer Institute (NCI) Designated Comprehensive Cancer Center and an academic teaching hospital, where robust and efficient clinical trial management programs are required to maintain quality standards and provide accurate data to sponsors in a timely manner [1, 2]. Data management challenges arise when numerous trials are conducted across multiple collaborators, such as disease centers, investigators, clinical research coordinators (CRCs), sponsors, and contract research organizations (CROs) [3]. Furthermore, medical imaging technology and informatics have seen growing complexity and quality improvements requiring imaging expertise to effectively manage imaging data [3–8]. Previously, CRCs at DFCI did not always know who to contact in imaging regarding clinical trial imaging protocols. Imaging scans were often ordered without technologists being aware of protocol-specific requirements resulting in imaging deviations and subsequent queries. Anonymized medical images were placed on digital media and then, either shipped from the CRCs to the sponsors via courier or uploaded electronically by the CRCs. Delays occurred between the time imaging personnel anonymized images onto CD/DVD media and the time the CRCs shipped or uploaded the images, causing more queries regarding the delay in the sponsor’s request of receipt of the images. Therefore, there were several instances between initial imaging scheduling and final imaging data transfer where protocol deviations or delays could occur. In this technical note, we describe our practical experiences and approaches used through our centralized clinical trial imaging data management workflow to address these challenges in order to improve quality, avoid unnecessary delays, and share our experiences with others.

Methods

We have established a centralized clinical trial imaging team and developed standardized procedures for site qualification, protocol preparation, and imaging data management including storage, retrieval, de-identification, submission, and query resolution. The full life cycle of these procedures is summarized in Fig. 1. Our team members include the lead imaging research technologist and designated research technologists in each imaging modality who serve as key imaging contacts for clinical trial site qualification, protocol preparation, and query resolution, along with the Imaging IT staff who designed and implemented infrastructure and workflows for the imaging data management and the imaging data associates who process and submit data to the study sponsors. Team members’ roles and responsibilities from the initiation of site qualification procedures through the completion of imaging data transfer and query management are described in further detail below.

Fig. 1.

Centralized clinical trial imaging data management workflow

Site Qualification and Protocol Preparation

The lead imaging research technologist works with CRCs, imaging modality supervisors, and designated imaging research technologists for site qualification and imaging protocol preparation by performing the following tasks [9]:

1. Qualifies imaging modality scanners by completing clinical trial site qualification surveys/questionnaires and coordinates phantom scan acquisition and transfers;

2. Reviews clinical trial protocols, imaging manuals, and site operations manuals to determine imaging acquisition, processing, and data transfer requirements. Relays relevant requirements to technologists in each modality and the imaging IT team;

3. Participates in clinical trial Site Qualification Visits and Site Initiation Visits;

4. Coordinates and provides clinical trial training for imaging technologists, radiologists, and nursing staff;

5. Writes Imaging Standard Operating Procedures (SOP) and Protocol Summary Sheets (PSS) outlining the requirements for patient preparation, image acquisition, and reconstruction methods for each participating imaging modality;

6. Coordinates creation of non-Standard of Care (SOC) scanner research protocols with designated imaging modality research technologists to ensure protocol compliance, standardization, and reproducibility across multiple imaging time points for the duration of the clinical trial;

7. Verifies that electronic transfer access is available to all members of the imaging IT team;

8. Maintains imaging trial regulatory binders and participates in site monitor audits;

9. Assists in resolving imaging-related sponsor queries; and,

10. Serves as the communication liaison between groups involved in clinical trials, including imaging radiologists, technologists, supervisors, radiation safety team, nursing staff, imaging IT specialists, clinical trial investigators, sponsors, CROs, schedulers, and CRCs.

Imaging Data Management

The research imaging data transfer workflow is described in more detail below.

Request for Clinical Trial CD/Electronic Data Transfer Form

To facilitate imaging data transfers for clinical trials, we developed an intranet web-based request form using the Livelink collaboration tool. The CRCs are granted access to the secured intranet link by the imaging IT team and request imaging data transfers through the web-based “Request for Clinical Trial CD/Electronic Data Transfer” form (Fig. 2).

Fig. 2.

Intranet secured web-based clinical trial CD/electronic data transfer form

On this form, the CRC enters relevant information for the scan transfer. Three different identifiers (Brigham and Women’s Hospital Medical Record Number [BWH MRN], DFCI MRN, and patient name) are required for each request to verify patient identity and accuracy of submission. Multiple lines are available for multiple imaging studies and timepoints in the research study. Research identifiers (DFCI Institutional Review Board [IRB] protocol number and Subject Identification [ID] number) are used to confirm the protocol’s requirements for imaging data transfer and to redact protected health information (PHI) during de-identification. Once requests are submitted, a confirmation page appears with a link to check the status of all the requests made by that CRC with a list of all previously submitted requests. Imaging personnel assign and process requests by updating and filling out the imaging services section of the request forms (Fig. 3). As a result, the worklist report queue is automatically updated with the status and the assignee.

Fig. 3.

Intranet secured web-based clinical trial CD/electronic data transfer form imaging personnel section

Electronic Transfer Database

The imaging IT team developed an electronic transfer database to reference the various data transfer requirements for each clinical trial (Fig. 4) that:

1. Identifies research collaborators: study sponsor, CRO, principal investigator(s), primary and secondary CRCs, and specific cancer center/department;

2. Lists the required electronic transfer methodology;

3. Lists trial identification—DFCI assigned unique IRB numbers, sponsor trial IDs, sponsor trial name, site ID, CRC contacts, sponsor, and CRO;

4. Establishes archives for all trial-specific imaging guidelines and trial documentation on our private, secured shared drive;

5. Lists pertinent imaging guidelines;

6. Verifies imaging personnel access to transfer images and electronic transfer methodology training; and,

7. Includes additional special requirements information.

Fig. 4.

Clinical trial database: main menu (left) and search trial form (right)

Imaging Data Storage

SOC clinical trial data do not require any special arrangements for their acquisition and storage. They are sent to the hospital Picture Archiving and Communication System (PACS) for permanent storage.

Special data storage arrangement is required for protocols that require non-SOC acquisition of imaging data. Non-SOC imaging data are not stored on our central PACS but on a dedicated research PACS. Since raw data is stored on the scanners for just a short period of time, we established a workflow where a technologist sends data to our dedicated research PACS and confirms the number of images archived with our imaging IT team. Once approved by the imaging IT team, the technologists can delete the data off the scanners. This ensures that important protocol data is confirmed and not lost during routine scanner data deletion.

Imaging Data Retrieval

Images are retrieved from our central hospital PACS (SOC images) or dedicated research PACS (non-SOC images) and processed on either Apple (OsiriX MD) or Windows (Hermes GOLD) workstations. The number of images is verified between PACS and local storage to ensure that all images are obtained [10].

De-identification and Redaction

Once images are retrieved locally and image counts are verified, the images are redacted with research identifiers following the Health Insurance Portability and Accountability Act of 1996 (HIPAA) and clinical trial guidelines. All irrelevant and PHI-containing forms and text documents are deleted from the study. All required images should not contain any PHI in the header or embedded in the pixels of the Digital Imaging and Communication in Medicine (DICOM) image files. Therefore, all DICOM images with embedded PHI such as topograms, secondary captures, or save screens are either deleted or edited with pixel de-identification tools. For this task, we use pixel de-identification modules included with the sponsor-provided electronic transfer tools or freeware Pixelmed Publishing DicomCleaner if those modules are not available [11].

Once these processes are completed, Medical Imaging Resource Center (MIRC) DicomEditor is used to remove DICOM header PHI information. Patient ID and patient name are redacted to a prefix of DFCI-assigned IRB number followed by subject ID number assigned by the research study [12]. All exceptions from the standard anonymization profile are noted in the electronic database for the specific trials.

Image Transfer

De-identified images are transferred using the sponsor-preferred method/application. Often times, the software has built-in de-identification features and requires study parameters to be entered for submission. Some protocols require the completion of Data Technical Forms (DTFs) with important information about scan acquisition. For these cases, the CRCs will send a partially blank DTF (with study ID, subject ID, and time point information completed) to the lead imaging research technologist in advance, so that the DTF can be attached to the patient’s paperwork on the day prior to the exam. The technologist performing the study completes the DTF at the time of acquisition and alerts the lead imaging research technologist when the form is ready for pick up. If retrospective completion of a DTF is required, most information can be retrieved from DICOM headers, electronic medical records, and our clinical imaging modality databases, which include fields for technologists’ input of parameters that could not be found from other resources [13].

Query Resolution

For the imaging-related queries, we verify that all images have been submitted and we resend them if any images did not arrive to the destination. For protocol deviations, we identify the reason of the deviation and take measures to prevent them from reoccurring. An audit trail documenting the use of electronic forms allows for improved accuracy to address queries. We also created a quality control redundant database to prevent unnecessary reprocessing of previously requested scans as follows: processed requests are queried and appended to the table of processed requests, while new requests are compared with the processed requests followed by a query that generates a list of potential duplications. The query searches for a combination of IRB, MRN, and scan dates to find matches. For all duplicate requests, the study CRCs are then asked to determine whether data reprocessing is necessary.

Results

With the number of processed clinical trial imaging studies approaching approximately 7000 per year in our institution, the implementation of standardized processes for clinical trial imaging data management has resulted in workflow improvements with institutional-wide benefits. Clear communications and responsibilities between collaborators have helped with quick adoption and improvement of this workflow. DFCI investigators and research staff know how to contact the proper centralized imaging contacts, and all imaging-related data management are off-loaded from clinics and CRCs onto our centralized imaging team. This has helped standardize image acquisitions and reduced delays in site qualifications, imaging transfers, and query resolutions.

Discussion

At our institution, hundreds of clinical trials involve imaging in the trial design and the centralization of the clinical trial imaging data management within the Department of Imaging has had a positive impact. The previous fragmented model of utilizing CRCs for clinical trial imaging data management was inefficient, as CRCs are not trained in imaging modalities and imaging data workflows. The previous lack of a main point of contact also led to delays in answering imaging-related queries which was resolved by creating a lead imaging research technologist position as the main point of contact for the entire clinical trial imaging management process. Furthermore, the establishment of dedicated research PACS for non-SOC studies, imaging databases and archives were key to the successful transition in consolidating imaging clinical trial tasks. This centralized role has helped standardize our imaging research activities, set up scanning protocols for the successful implementation of imaging in these clinical trials, ensure compliance, address and resolve issues, and improve communication between the various multidisciplinary groups. Queries within our quality control redundancy database has been very helpful to our imaging transfer team in identifying duplicate transfer requests submitted by CRCs, thus avoiding the unnecessary reprocessing and resubmission of imaging data.

Conclusion

The creation of a centralized clinical trial imaging team, standardized imaging acquisition protocols, centralized clinical trial imaging data management, and standardized workflow described above has consolidated multiple tasks from site qualification procedures through the completion of imaging data transfer and query resolution, and has enabled us to provide quality, accurate, and timely imaging data to clinical trial sponsors.

Compliance with Ethical Standards

Conflict of Interest

The authors declare that they have no conflict of interest.