Survival rates for children with cancer have increased dramatically over several decades, revealing a host of late effects associated with treatment. Proton therapy promises to reduce late effects because of its ability to better localize dose, but the question is by how much? Pediatric health outcomes must be prospectively studied to determine what the margin of benefit is. To facilitate such research, a consortium consisting of pediatric investigators from the US was formed. The goals of the Pediatric Proton Consortium Registry (PPCR) study are to create a comprehensive database of pediatric patients treated with proton radiation therapy in the United States to be used and accessed by participating institutions, to describe the patterns of follow-up at proton facilities, and to describe the acute and late effects in the children treated with proton therapy.
REDCap is the platform for the ∼600-field database which was designed to capture baseline, treatment, and follow-up information of children and families who consent to participate. This body is governed by the steering committee and coordinated through Massachusetts General Hospital. Initial funding is provided by the National Cancer Institute/Massachusetts General Hospital federal share funds.
Currently, there are 4 active sites that have collectively enrolled 232 children in ∼17 months and another 7 that are poised to open soon.
This study serves to introduce the PPCR, to share our structure, process, and preliminary enrollment data, and to openly invite every new proton center in the U.S. to join in hopes of accelerating the pace of outcomes research for children treated with proton therapy. Our primary vision for the PPCR is that it will provide a platform for communication and research among proton therapy providers to accelerate the pace at which we learn from our patients.
Today, nearly 80% of childhood cancer patients are long-term survivors. There is a dual focus of pediatric cooperative groups continuing to cure children at high rates while diminishing the late side effects and morbidities associated with the treatment. The Childhood Cancer Survivor Study (CCSS) has successfully catalogued the late effects of treatment and published seminal articles on how radiation therapy and chemotherapy treatments have adverse effects on childhood cancer survivors . Radiation therapy to pediatric cancer patients exacts a late-effects toll that is inversely proportional to age at radiation therapy yet directly proportional to volume of tissue irradiated and radiation dose. The current focus in radiation oncology is to improve dose localization to the target, which allows for dose escalation, if required, and lower toxicity profiles. There have been tremendous advances in radiation therapy delivery that have had a remarkable impact on its delivery, including better immobilization for reproducible treatment, better imaging for target delineation, and improved delivery techniques with intensity modulation. Particle-beam radiation therapy is a newly recognized modality of radiation that is being harnessed to further improve the dose localization to the target and to better spare normal tissues. The most widely available form of particle therapy is proton therapy, and its dose distribution in most cases is superior to that which can be achieved with even the most sophisticated photon techniques [2, 3]. Unlike photon radiation, which entails both an entrance and exit dose to normal tissues in the process of treating a tumor in a given beam line, proton radiation entails a somewhat lower entrance dose and eliminates exit dose to normal tissues. Therefore, proton therapy dose distributions are usually superior to those of photon therapy by a factor of 2 or greater allowing the potential to further improve clinical outcomes by specifically decreasing toxicity associated with treatment [3, 4].
There are currently 13 operating proton centers in the United States and many more are opening imminently . It is widely accepted within the radiation oncology community that protons have great potential to reduce the late effects of treatment in the pediatric population [6, 7]. Many of these proton therapy centers conduct single-institution research and participate in Children's Oncology Group trials. Nevertheless, because of the relative rarity of pediatric cancers, accumulating the appropriate data is an expensive and prolonged process that impedes the timely and necessary comparison of proton outcomes to existing photon data on late effects. Furthermore, the United States lacks a comprehensive profile of pediatric patients being referred for and ultimately receiving protons. For this reason, the Pediatric Proton Consortium (PPC) has been created to unify and share the experiences of pediatric proton therapy centers across the United States. The current members include Massachusetts General Hospital (MGH; Boston, MA, USA), MD Anderson Cancer Center (Houston, TX, USA), University of Pennsylvania (Philadelphia, PA, USA), University of Florida Proton Therapy Institute (Jacksonville, FL, USA), CDH Proton Center (Warrenville, IL, USA), Washington University (St. Louis, MO, USA), Indiana University Health Proton Therapy Center (Bloomington, IN, USA), Mayo Clinic - Rochester (Rochester, MN, USA), Hampton University (Hampton, VA, USA), ProCure Proton Therapy Center - Oklahoma City (Oklahoma City, OK, USA), Mayo Clinic - Arizona (Scottsdale, AZ, USA), ProCure Proton Therapy Center - New Jersey (Somerset, NJ, USA), Loma Linda University (Loma Linda, CA, USA), Rutgers Cancer Institute of New Jersey (New Brunswick, NJ, USA), and the Provision Center for Proton Therapy (Knoxville, TN, USA). An invitation to subsequently-opening proton centers will be extended to the pediatric provider, director, or chair as long as sufficient funding for this project is available.
Spearheaded by proton investigators at MGH, with funding from the National Cancer Institute and MGH, the first initiative of this group has been the development of the Pediatric Proton Consortium Registry (PPCR). The objective of the PPCR is to create a comprehensive description of the pediatric patients being treated with proton radiotherapy around the country. Based on a preliminary survey and communications with the current open proton centers conducted by the Pediatric Proton Foundation, approximately 700 pediatric patients receive proton radiotherapy a year  (Figure 1).
The potential benefits of the PPCR are varied and great, including the ability to identify referral patterns, evaluate the efficacy of protons for disease control, and better evaluate late effects outcomes. Currently, proton therapy is a very constrained resource for pediatric patients, and rationing of available treatment slots is necessary at some institutions . For unresectable tumors, proton therapy may be a better option than photon therapy because of the enhanced ability to escalate the dose , but it is currently underutilized at some centers either because of poor clinical prognoses or because of limited availability of treatment slots . PPCR data may potentially be used to determine the current underlying barriers to proton therapy treatment, better understand referral patterns for proton treatment and follow-up care, provide estimates on the subsets of tumor types being treated with protons, correlate treatment patterns with socioeconomic and insurance status, and track the changing treatment indications as the availability of treatment slots increases with the emergence of new treatment facilities. The PPCR data will provide metrics for more precise delineation of the overall number and geographic distribution of proton centers needed to serve the growing volume of pediatric patients clinically indicated for proton radiotherapy. Ideally, this comprehensive and systematized integration of clinical and practice data provides a regular forum for communication among pediatric proton therapy providers that will lead to improved practices from which the whole community can learn.
Patients and Methods
The PPCR was conceived of in May 2010 when the primary collaborators convened to discuss the schematics and feasibility of an innovative, prospective, multi-institutional pediatric registry. Following a period of collaborative design, the study first opened to enrollment at MGH in the summer of 2012. At the start of March 2014, there were 5 enrolling sites, with several additional sites targeted to open to enrollment soon. This initial wave of PPCR sites will encompass the largest currently operating pediatric centers, where over 50% of the pediatric patients were treated from 2010 to 2012. A second wave of consortium sites is scheduled to initiate later in 2014.
Identification, Eligibility, Consent, and Registration of Patients
Patients eligible for inclusion in this study must receive proton radiation at a consortium site and be age < 22 years at the start of radiation treatment. Patients may be receiving concurrent therapy, they may have any type (benign or malignant) or extent (local or metastatic) of disease, and they may be participating in other concurrent clinical trials. Once all currently open consortium sites are activated, annual enrollment of approximately 700 patients is expected, although that number will grow as more proton centers open in upcoming years.
Each participating institution will invite all eligible patients to participate in the PPCR in order to minimize potential selection bias. Patients and/or their legal guardian are approached at any time point from the radiation oncology consultation through the last day of radiation treatment. Informed written consent and assent, when appropriate, is obtained after the details of the PPCR are explained, all questions have been answered, and the patients/legal guardians have read the consent form. This is purely an observational study requiring no additional tests, visits, or changes in the treatment plan. The radiation plan, dose, and modality are not affected by participation in the PPCR. The consent allows for the investigators to periodically check in with the patients and their local medical team for the purposes of updating the patient's health status in the database.
When English is not the patient or legal guardian's primary language, an interpreter is used. Prior research has shown that minorities and individuals who do not speak the native language are less likely to participate in research [11, 12]. For patients who decline participation for any reason, non-identifying demographic data, cancer diagnosis, and reason for not participating will be recorded in order to capture in full the cohort of pediatric proton therapy patients, and to assess if there is a difference between those who consent and those who do not. This study is approved by the host institutional review board (IRB) at the Dana-Farber Cancer Institute at Harvard University (Boston, MA, USA). Each institution opening the registry also obtains local IRB approval.
Following Dana-Farber Cancer Institute policy, all patients are registered centrally by the MGH Coordinating Center with the Dana-Farber Cancer Institute Quality Assurance Office for Clinical Trials. Once the participating sites submit the signed informed consent form and documentation of eligibility, the MGH Coordinating Center registers the patient with the Quality Assurance Office for Clinical Trials. Once registered, the site receives confirmation of enrollment and data collection commences.
Establishment of an expansive and robust data set such as the PPCR requires careful consideration and application of data sharing, use, and publication guidelines. Concurrent to the inception of the PPC, the PPCR Steering Committee was devised in an effort to create and employ consistent and impartial practice for requesting the use of PPCR data, reviewing data use requests, and ensuring timely and appropriate publication and dissemination of PPCR outcomes.
The PPCR Steering Committee comprises no more than 12 members and will constitute one appointed co-investigator from each participating site. Should the number of consortium sites be greater than 12, members will be chosen by an election held among all co-investigators. Members will serve for a term of 2 years, and may be re-elected as their site representative if no other candidates are available.
The Steering Committee is charged with the responsibility of reviewing data use requests, approving and prioritizing data use proposals that align with the goals and expectations of the PPCR, and fostering and mentoring ancillary projects that leverage and promote the mission of the registry. Steering committee functions are supported by appointed Data Use and Data Quality Control Sub-committees. The founding Steering Committee includes the site investigators from the first wave of sites initiated for enrollment. Under the governance of this initial Committee, a Steering Committee Charter and Data Use and Publication Guidelines were adopted. During these early stages of the PPCR, the Steering Committee meets to review project milestones and data completion, and to strategize topics for initial PPCR publications. As the PPCR extends to more sites and the data set grows, the role of the Steering Committee will become vital for promoting collaborations and fostering communication among investigators, providing support and guidance on ancillary projects, balancing conflicting priorities and resources within the consortium, and ensuring timely project performance and reporting (Figure 2).
Training, Site Communication, and Data Oversight
Shortly following the opening of the study at MGH, participating first-wave sites were asked to attend an investigator's meeting for site initiation training and to review and provide input on database structure and design. This meeting was held in June 2013 in Boston, MA, USA. Because it was anticipated that there would be some challenges resulting from different practice patterns and capabilities, it was decided to conduct a slow roll-out of the PPCR across sites to help target and solve any issues that might arise before opening at all sites. Minor changes to the database to accommodate practice patterns are expected and will be easier to incorporate and amend in the protocols at the various institutional IRBs if the initial number of sites is smaller. At this meeting, investigators and research staff conducted a detailed review of the protocol, multi-center regulatory processes, and the PPCR database structure. Additionally, a teaching session was held that allowed each site to enter a patient into the test site for additional feedback. The MGH Coordinating Center staff was on hand to answer questions. Additional site initiation meetings will be planned as new sites open to enrollment.
PPCR members participate in a monthly conference call to update all members of current PPCR status, review enrollment at the sites, address data-entry questions or issues, and review and maintain currency on the regulatory requirements. Minutes are disseminated after each call.
Database and Data Collection
Study data are collected and managed using REDCap (Research Electronic Data Capture) electronic data capture tools hosted at the MGH. REDCap is a secure, web-based application designed to support data capture for research studies which can provide the following: (1) an intuitive interface for validated data entry; (2) audit trails for tracking data manipulation and export procedures; (3) automated export procedures for seamless data downloads to common statistical packages; and (4) procedures for importing data from external sources . It is continuously maintained and updated through federal funds from the National Institutes of Health who is committed to having a robust Internet-based database that researchers can use without paying a proprietary fee. REDCap is compliant with Part 11 of the Code of Federal Regulations Title 21.
Using the REDCap application described above, the PPCR has almost 600 customized fields organized into the following sections: (1) demographics; (2) primary diagnosis; (3) baseline health inventory; (4) tumor-related surgery; (5) imaging; (6) radiation; (7) chemotherapy and protocols; (8) follow-up, and (9) off-study. Some of the forms are completed only once but others, such as the imaging and follow-up forms, are longitudinal and therefore used repeatedly as new time points arise, as illustrated in Figure 3. Branching logic is used throughout the database to only reveal relevant questions and answers subsequent to certain questions. For example, one of the major branching points for the database is based on the diagnosis page, which asks if the disease is central nervous system (CNS) or non-CNS. Based on the answers only histologies appropriate to CNS disease are presented when the user indicates the tumor is a CNS tumor. Similarly, other questions are tailored to the issues that patients with CNS primaries are likely to encounter. The follow-up time points are numbered but are not linked to a specific time point. The goal is to collect follow-up data at approximately 3 months after the completion of radiation in order to capture any acute effects of treatment and then annually to monitor for disease control and treatment related late effects. However, follow-up data can be entered at any point and will be logged by date (Figure 3).
The PPCR database was created and is managed by the MGH Coordinating Center and the MGH Biostatistics Center. Prior to data entry of enrolled patients, the database was piloted in test mode by members of the MGH Coordinating Center, an MGH research coordinator not affiliated with the PPC, and by members of the PPC. During this process, patient identifiers were redacted and patient privacy was persevered at all times. This piloting process led to adding, deleting, and clarification of data fields. A procedure is in place to identify and implement future modifications to the database.
Data for the PPCR is gathered by clinical research staff trained both in Good Clinical Practice (CITI Training) and REDCap. Data is extracted from source documentation located in the electronic medical record, MOSAIQ or comparable radiation oncology management system, notes and records reviewed from outside institutions, and from direct patient interactions. Internal audits and those conducted by the MGH Coordinating Center and the Quality Assurance Office for Clinical Trials are performed on a regular basis to ensure data integrity.
Quality Assurance Review Center
Quality Assurance Review Center (QARC) is a program affiliated with the University of Massachusetts Medical Center that is used by institutions and groups (e.g. Children's Oncology Group) as a diagnostic imaging and radiation therapy data management program. QARC's data management standard operating procedures are compliant with ICH, Good Clinical Practice, and Food and Drug Administartion regulations. The QARC database for storage of digital treatment plans is compliant with Part 11 of the Code of Federal Regulations Title 21.
Final treatment plans including computed tomography structures, radiation treatment plans, and dose files are sent to QARC electronically, in DICOM format. Baseline and follow-up imaging, when available, such as computed tomography, magnetic resonance imaging, and positron emission tomogrpahy scans are also sent to QARC for future projects. The plans and imaging are stored and will be reviewed for both quality assurance and future hypothesis generation. Participating investigators will have access to data, de-identified as needed, via remote access or on site at QARC.
As of March 3, 2014, there were 232 patients enrolled across the 5 active sites. At MGH, a total of 162 patients were screened and 141 are currently active. Of the remaining 21 patients, 6 are deceased, and 15 were not enrolled. The reasons for non-enrollment are recorded as follows: not interested (n=6), unable to consent patient prior to completion of radiation (n=4), language barrier (n=1), no reason given/unknown (n=3), and other (n=1). At the CDH Proton Center, 29 patients were screened, 28 enrolled, and all are active. The one patient not enrolled falls into the “other” category. At the University of Florida, 58 patients have been screened and 57 are enrolled and active. The one patient not enrolled falls into the “unable to consent prior to completion of radiation” category. The two additional active sites are University of Washington and University of Pennsylvania. They have not yet enrolled any patients.
The PPCR is a collaborative effort designed to bring investigators together to create a comprehensive database that will encourage faster and more collaborative proton-based research in the pediatric setting. We anticipate it will initially serve to address questions related to patterns of care and acute toxicity in children receiving proton radiation therapy. In the future, it may be comprehensive enough to illustrate patterns of follow-up, late toxicity, and disease control. It is our hope and intention that the PPCR will provide readily available data for additional dose-effect analyses that will answer questions surrounding the treatment of children with both proton- and photon-based radiation therapies. It will be a readily available resource for investigators from participating institutions and may help foster multi-institutional collaborations that extend outside of the registry as the site participants will be regularly meeting and sharing academic ideas. The Steering Committee is in place to facilitate and oversee this process. As an evolving project, it is expected that additional IRB guidance will be needed for questions beyond the scope of initial objectives outlined above.
Once the registry is firmly established at participating centers, the Steering Committee may choose to expand the registry in various directions. REDCap has the capacity to incorporate patient-reported outcomes through patient surveys that can be entered directly into the database by the patient. This may help with our follow-up of patients, which can be a challenge because many patients travel to proton centers for treatment and financial and logistic barriers prohibit follow-up at those centers.
The registry may also open to photon centers to help facilitate comparative effectiveness research. The existing PPCR infrastructure platform may allow the participating sites to open the same or similar protocols for standardized follow-up of pediatric patients. It is the hope that sufficient academic output will be generated to successfully apply for grant funding once the MGH Federal Share/National Cancer Institute funds run out in 2021 (or sooner depending on reimbursement dynamics).
In the approximately 12 months since the PPCR enrolled its first patient, the Steering Committee has been encouraged by the high enrollment rate and low refusal rate. Ideally, the PPCR would include every child treated with proton radiation in order to have a truly comprehensive registry. Since this is not possible with the requirement of consent, we have obtained IRB permission to maintain a screening log with basic data that will enable us to meet the secondary goals of the registry related to treatment and referral patterns, and to determine if the patients who did not consent differed in some basic ways to those patients who did consent. It is too early to make any observations of those patients who were not enrolled in the PPCR.
Challenges have included the realities of our decentralized regulatory process, which have prevented more sites from opening to date. Each IRB has its own requirements and we are actively working with each site to meet these requirements to get them activated and enrolling as soon as possible.
The database is large with almost 600 fields. It is intentionally designed to be a robust archive that can serve the purposes of each home institution for their own research goals. In fact, its very nature has made it attractive to many centers. However, only a proportion of the fields are “required,” which means they are linked to monitoring reports. “Required” fields are those thought to be critical to the fundamental querying and ability to meet the primary endpoints of the study. It is recognized that there are challenges to completing all of the PPCR REDCap fields which can include the following: unknown or missing clinical information especially when the proton center is not the patient's primary hospital, copious research follow-up and data entry that may overwhelm the research staff allocated by the PPCR funding model, lack of staff resources to spend the time to collect and enter all of the data, and patients being lost to follow-up over time. There will undoubtedly be other unanticipated challenges in a project of this scope. We have considered the statistical compromises to having a database that is “half full” and ultimately decided that the opportunity to add as much information as possible was a preferable goal. Our Quality Assurance program will regularly review the data in the registry of all sites. We anticipate a steep learning curve as to what enhances and hinders the quality of the data early in the project. We will make appropriate adjustments as needed.
Short-term future goals include getting all of the PPC sites open to enrollment. We have an active Coordinating Center that is currently working diligently on this goal. Once the sites are open and trained, the Coordinating Center will continue to provide support to promote a high rate of enrollment. This support will take the form of monthly conference calls, frequent email communication, and site initiation/support visits as well as annual in-person meetings either separately or in conjunction with national meetings of the American Society for Radiation Oncology, Children's Oncology Group, or Pediatric Radiation Oncology Society where most of the investigators will already be present.
In addition to publishing patterns-of-care studies and proton cohort studies of pediatric patients receiving proton therapy in the United States, we hope to collaborate with other groups to answer other important research questions. As the PPCR matures there are a number of different avenues and possible collaborations that can be forged, and we hope to pursue these relationships once the PPCR becomes more firmly established. For example, the CCSS is cataloguing late effects in childhood cancer survivors. While the timing of the cohorts for the CCSS and the PPCR are different, the missions are not dissimilar. The PPCR could be altered to collect data through the REDCap survey functions in a similar manner as the CCSS does for the purposes of comparison. An important hurdle to overcome, however, would be the different eras during which the patients were treated. The CCSS cohort includes patients who were treated between the years of 1970 and 1999 (including the expansion cohort added later) [1, 14]. The PPCR will only be enrolling patients prospectively, but conceivably similar methods of data collection could be used in the future. Or, collaborations with Children's Oncology Group studies may be possible that can help foster the Comparative Effectiveness Research goals of the National Institutes of Health and the National Cancer Institute in assessing new technologies, of which proton therapy is a perfect example.
The PPCR is an exciting and unprecedented collaboration within the pediatric radiation oncology community intended to create a comprehensive database encompassing rare cancers and unique technologies. Through data sharing, this model affords a critical number of patients necessary to optimize timely comparative effectiveness research. Ideally, the PPCR will provide a platform for both a broad understanding of the current state of pediatric proton therapy and generate future hypotheses that inform the therapeutic ratio of radiation in children.
ADDITIONAL INFORMATION AND DECLARATIONS
Conflicts of Interest Disclosure: Dr. William Hartsell has a minority ownership interest in CPTI, LLC. Dr. Andrew L. Chang is a consultant for the Proton Center Development Corporation. Dr. Daniel J. Indelicato is part of the Volunteer Board of Directors at the Pediatric Proton Foundation. Dr. Stephanie M. Perkins received travel funding from Mevion Medical Systems. All other authors have no conflicts of interest to disclose.