Modern Dosimetry in Radiation Oncology Clinical Trials

Abstract

Clinical trials in radiation oncology are the best vehicle to optimize our strengths in therapeutic technology, define progress in our field, and improve patient outcome. Trials advance our knowledge in each disease site and provide us information to improve the radiation dose-volume for both tumor control and therapeutic sequelae to normal tissue. An increasing number of systemic and targeted therapies have been developed and are currently in early phase clinical trial design. Ultimately, these new therapies will need to be tested with standard-of-care therapy including radiation oncology. Therefore, during a study, it is essential that radiation therapy is delivered in a uniform and consistent manner for the credibility of the study. If the radiation therapy component of the study does not have a structure or management for maintaining therapeutic compliance, including a real-time data management strategy, it becomes difficult to trust the study outcome and apply the outcome to daily clinical practice. In this chapter, we review the strategy and process involved in the management of dosimetry in radiation oncology clinical trials and how this can impact clinical trial management, primary study endpoints, and the overall success of the study.

1. Introduction

As imaging and therapy technologies have matured over the past 25 years in the field of radiation oncology, the processes of radiation therapy treatment planning and execution have undergone extraordinary change. Accordingly, data acquisition and data management in clinical trials, including radiation therapy, have likewise undergone considerable change commensurate with the technological process improvements in our field. As radiation oncology committees began to develop influence in the National Clinical Trials Network (NCTN), group leadership recognized the need for radiation therapy guidelines to be imbedded in clinical trials and clinical data, including simulation and portal images, to be submitted to a quality assurance (QA) center for review. It was recognized in the early development of radiation oncology clinical trials that a system for credentialing institutions and investigators for participation was an important step in ensuring that radiation therapy treatment could be delivered in a consistent manner. In this phase, the Radiological Physics Center (RPC) developed a plan to quantify thermoluminescence dosimetry (TLD) monitoring of each accelerator used for clinical trial participation. Inventories were kept, confirming both the participating investigators and equipment that would be used. As protocols matured and disease-specific protocols required enhanced QA management strategies, the RPC developed a series of phantoms that could be sent to institutions to verify the radiation therapy dose delivered. The phantoms had TLDs imbedded in selected areas for tumor and normal tissue doses. Institutions would irradiate the phantoms and return to the RPC for evaluation. This process would test tumor treatment targeting and image processing in multiple disease areas. The hepatic phantom would also test motion management during therapy. Credentialing confirmed uniform radiation dose execution across centers participating in clinical trials and likewise generated confidence in each institution that therapy delivered at their site was consistent with treatments delivered at other centers. This was important, especially in the early phase of clinical trial development, as computational and quantitative radiation oncology varied between institutions, and establishing a common platform for computation and radiation treatment planning was essential for consistency of treatment execution.

During this era, the infrastructure of radiation therapy treatment planning began a seismic change. For decades, radiation oncology simulation was based on fluoroscopic simulation with planning performed in two dimensions through the isocenter of the therapy treatment field. The ability to perform computations off isocenter was limited and could not be easily adjusted in the care plan. The advent of volumetric radiation therapy treatment planning introduced a new era in radiation therapy treatment planning and delivery of therapy. Computer tomography treatment planning made radiation oncologists and physics planning teams think in terms of volumes treated in three dimensions. The language of radiation therapy moved from calculations measured by isodose lines to dose defined by volume. In this capacity, the importance of imaging in defining the target for treatment became essential and as important a component of QA in clinical trials as uniform computational metrics. The Quality Assurance Review Center (QARC) managed radiation oncology clinical trials for both the adult and pediatric clinical trial groups as well as imaging for the pediatric group. Because of expertise in the collection of planning objects including imaging, QARC began collecting relevant imaging used to define the target for radiation therapy and outcome imaging to confirm the site of recurrence and site of injury when this occurred. In this capacity, imaging could be repurposed as a vehicle to validate disease status depicted on case report forms. The imaging could be reviewed by study and site investigators to ensure that the tumor target was appropriately defined and that treatments were conducted in a protocol-compliant manner. As trials became more complex and digital transfer tools became functional at an enterprise level, the review of objects could be conducted pre-therapy in a real-time manner from anywhere in the world.

This became important for multiple reasons. In what would be referred to today as intermediate and high-risk Hodgkin lymphoma, patients entered on the Pediatric Oncology Group (POG) protocol 8725 were treated with chemotherapy, and half were randomized to receive radiation therapy to all areas of the original disease. The initial publication did not reveal an advantage to those who underwent radiation therapy. However, a secondary analysis of the data was completed by one of the authors (TJF) and demonstrated a 10% statistically significant survival advantage to patients who were treated on study to radiation therapy volumes that were study compliant. Deviations on study with radiation therapy were uniform due to excluding areas of the original disease from the radiation therapy treatment fields, therefore implying that using radiation therapy as a treatment consolidation tool, all areas of original disease need to be treated as part of the treatment planning strategy. This spawned a new approach in the evaluation of radiation therapy treatment plans in cooperative group trials. The initial response in the next iteration of clinical trials involving Hodgkin lymphoma in POG was to move the retrospective review of radiation therapy treatment plans to pre-treatment review of radiation therapy treatment objects to limit deviations and achieve consistency in treatment plans between participating institutions and investigators and limit the influence of non-uniform radiation therapy delivery on the clinical trial design and outcome. Today, clinical trials are managed with real-time and adaptive approaches, including review of radiation therapy treatment plans, using nimble informatics tools that can connect study and site investigators together as soon as digital data arrive at a QA center. Clinical trials today in Hodgkin lymphoma are using informatics tools to augment therapy to those with limited response to induction therapy and titrating therapy to those with rapid response to induction therapy. Advanced stage Hodgkin lymphoma patients on study are now treated with an emphasis on chemotherapy and radiation therapy delivered to areas of incomplete response to systemic therapy as defined on anatomic and metabolic imaging. To achieve the goals of this study, primary and response imaging on therapy as well as radiation therapy treatment objects need to be reviewed at QA centers to ensure the correct areas are being treated to the protocol-compliant dose and the dosimetry meets the study defined constraints to normal tissue. The perception that collecting the data adds to cost and is a burden to site investigators, and there are many examples where limiting data acquisition negatively affected trial outcome and interpretation of the data, leads to unrecovered financial loss and generates a false narrative to trial conclusions [1, 2, 3, 4, 5, 6, 7, 8].

The HeadSTART protocol was designed as a randomized study to test the efficacy of adding Tirapazamine to chemoradiotherapy for the management of locally advanced squamous cell carcinoma of the head and neck. Preliminary phase 2 data were favorable, and a phase 3 study was designed to affirm the favorable data of the phase 2 study. The study was designed with on-treatment review of radiation therapy treatment objects (first 3 days of management) and not real-time review to promote worldwide participation on study and facilitate trial accrual. The data management was performed by QARC. Unfortunately, nearly 25% of the patients on study were asked to adjust treatment volumes to ensure protocol coverage of what was interpreted as gross tumor, and of those asked to make changes, less than 50% made the requested adjustment. As a result, the deviations on study significantly influenced trial analysis, and the trial did not reach the desired endpoint. Patients with compliant plans de novo had the best outcomes. Those that made the requested changes and those in retrospect who had plans that did not meet dose-volume tumor constraints but thought not clinically meaningful had identical clinical survival outcomes, which were 10% less than those who had compliant plans at the presentation. Therefore, the data make a strong argument for real-time review of objects for study compliance, which can be easily accomplished today with modern informatics tools [8]. RTOG protocol 0617 was designed to test cetuximab with radiation therapy in non-small cell lung carcinoma. Two radiation schemes were imbedded into the study, which included a high-dose (74 Gy) and a lower dose (60 Gy). The study demonstrated no clinical advantage to the higher-dose arm; however, ironically, the local control in the high-dose arm for the first several years on study was 10% worse, and this was statistically significant. Although dosimetry was reviewed and deemed compliant to study metrics, there was no pre- or post-therapy imaging collected on the study. Therefore, the tumor targets and contouring of disease could not be validated, and outcome imaging was not collected to affirm or confirm the site of failure and the relationship of failure to either excluding the disease from treatment or to dose gradient. In this study, cases could be considered compliant to metrics, and the contouring of disease could only be evaluated on the radiation therapy planning systems. Therefore, data acquisition strategies in clinical trials need to include all elements that clinical investigators use to apply therapy to patients in the clinic [9]. The trial must acquire similar information to be confident in the interpretation of the study results. Children’s Cancer Group (CCG)/POG protocol 9961 tested the ability to perform both dose and volume titration to children with standard risk medulloblastoma. Important molecular biomarkers were collected as part of the trial design, and this has proven to be invaluable in re-defining risk factors in this disease. Radiation therapy objects were reviewed in real-time. However, in this study, 10% of patients were deemed ineligible for study on retrospective review mostly due to the presence of persistent disease in the posterior fossa post-surgery or evidence of spinal disease not identified at the time of entry onto the study. These patients who in retrospect had a disease at presentation more advanced than study requirement had a significantly worse outcome, therefore establishing the need for real-time review of imaging pre-therapy to ensure the correct patient is entered onto the correct study and the study population is uniform. Each patient on study is a valued resource and deserves our undivided and full attention to facilitate the outcome they deserve [10, 11]. Outcome imaging is essential as it becomes the primary vehicle to assess the level of responsibility imposed by therapy on the location of treatment failure including injury imposed by therapy as well. QARC (now the Imaging and Radiation Oncology Core (IROC), Rhode Island) houses outcome images on Children’s Oncology Group (COG) trials, which have been essential in distinguishing disease progression from therapy effect in multiple disease groups, including the central nervous system and late effects imposed by therapy on normal tissue [6, 7].

Therefore, modern dosimetry analysis by members of radiation oncology physics teams works in conjunction with radiation oncologists to ensure contours to both tumor and normal tissue are constructed correctly, and the therapy plan is delivered to protocol specifications and clinical standards to ensure the data can be trusted and applied into clinical practice. In each of these examples, limitations in data and data review as part of the trial design directly influenced the interpretation of the study. Therapy is becoming increasingly complex in both treatment design and treatment execution. Clinical trials currently incorporate the complexities of modern therapy including stereotactic programs, compressed fractionation, radiopharmacy, and particles into trial design and execution. In many studies in pediatric oncology, photon and proton treatment delivery reside synergistically in the study with identical volume guidelines in order to not limit study accrual. In more common adult disease areas, modern physics teams and QA centers will be able to successfully apply strategies for the management of dosimetry on clinical trials for these technologies using the clinical experience of investigative teams [12, 13, 14, 15, 16, 17, 18, 19, 20].

2. Dosimetry of modern protocols in radiation therapy

2.1 Qualification

The initial step in the process of evaluation of physics and dosimetry for clinical trials is to visit the IROC Houston website ([e-mail]) to file an inventory list of onsite radiation oncology equipment for treatment planning and delivery. This will include the available equipment for radiation therapy treatment planning, accelerator information, imaging and image validation, and personnel involved in the clinical trial process. This will also identify personnel who will be involved in the clinical trials process including physicians, physicists, and data management staff. IROC will upload the data acquired through the inventory to both the Clinical Trial Support Unit (CTSU) and the Cancer Therapy Evaluation Program (CTEP). As part of this program, all megavoltage photon, proton, and electron beams at each participating site have reference beam output calibration. Brachytherapy audit tools have likewise been developed. If the TLD/optically stimulated luminescence dosimeter (OSLD) measurement is 5% out of alignment, IROC Houston initiates a review process, which will include a review of procedures. If the discrepancy cannot be resolved at this level of interaction, a site visit is performed, which measures calibration, QA procedures, image guidance and multileaf collimator function, and treatment planning system calculations with IROC-measured dosimetric performance [21, 22, 23].

These processes ensure confidence that the institution and involved investigators have the equipment and expertise to perform fundamental functions to participate in clinical trials involving radiation therapy and can design and execute patient care in a protocol-compliant manner and meet dosimetry constraints to both tumor and normal tissue targets.

2.2 Protocol development and clinical trial support

Clinical trials remain the best vehicle to move evidence-based knowledge forward and improve clinical care. For the NCTN, IROC becomes involved with study investigators at the time of the concept sheet development. This interaction is particularly important when emerging and developing technologies are being introduced into national and international studies as a structure can be placed into the study to ensure treatment symmetry between institutions, support the development of a uniform study population, and answer protocol-specific questions. This decreases the likelihood that an asymmetric delivery of radiation therapy between institutions will influence the primary study question even when the primary question does not have a radiation therapy study endpoint. During this phase, IROC will develop protocol-specific metrics including tumor target and normal tissue dosimetry constraints to be applied as well as the definition of what would be considered a volume or a dosimetry deviation on study. IROC will write the imaging and radiation therapy protocol guidelines for pediatric and adult cooperative groups for each planned study to make certain the protocol is written in language compatible with the goals and objectives established by the CTEP. The written protocol will include imaging strategies and acquisition pathways for the definition of target and outcome evaluation, radiation therapy dose prescription including percent dose to target volume, protocol compliance objectives and definition of deviations, radiation therapy planning instructions including immobilization strategies and image guidance, QA procedures and implementation, and the data submission process and timing of submission of information, especially on studies that permit adaptive volume adjustment based on response to chemotherapy. The formatting also permits patient-specific information including dosimetry to be submitted to Rave and the National Cancer Institute (NCI) Cancer Research Data Commons. Examples of dosimetric guidelines with volumetric language written into studies are listed in Table 1. Note that standard nomenclature is required for target and organ names [24].

Target/OAR—Standard name	Description	Metric	Per protocol	Variation acceptable	Deviation unacceptable
PTV1	PTV1	D95%[%]; Dose to 95% of the volume	> 98% of protocol dose	>95% of protocol dose	<95% of protocol dose
GTV_PET	GTV identified on PET scan	D100%[%]; Dose to 100% of the volume	> 98% of protocol dose	>95% of protocol dose	<95% of protocol dose
Lungs	Both lungs minus the GTV	Mean[Gy]	≤ 20Gy	≤ 22Gy	> 22Gy
Heart	Heart	V30Gy[%]	≤ 50%	≤ 55%	> 55%
Spinal cord	Spinal cord	D0.03cc[Gy]; Max dose to 0.03 cc	≤ 50Gy		> 50Gy

Table 1.

Protocol-required target and normal tissue constraints. The table above shows the expectations of radiation therapy dose criteria for specific targets.

2.3 Credentialing

For participation in clinical trials involving radiation therapy through the NCTN and the CTSU, the initial step in the process is to contact the IROC Houston office ([e-mail]). A step beyond qualification, credentialing is the vehicle used to determine that the clinical physician/physics team has the resources and expertise to meet individual and specific requirements of a study. IROC implements credentialing through multiple mechanisms that are specific to the clinical trial. Examples include stereotactic therapy phantoms with image guidance including contouring reviews as well as a liver phantom designed to test treatment of multiple lesions with motion management. The phantoms have TLD imbedded into selected areas for validation. Many of the phantoms are technology specific and can be applied across individual protocols and NCI Groups. For example, once an institution passes the phantom for intensity modulation therapy, the credentialing extends to all protocols using intensity modulation across all the NCI Groups. The phantoms can be repurposed for use with varied technologies. For example, the head and neck phantom has been used to credential for intensity modulation and can also be used to credential institutions for proton head and neck therapy. Phantoms have also been made specific to individual disease sites and protocols. For example, pediatric total body radiation therapy and pediatric spine have specific phantoms for participation in clinical trials. The pediatric total body phantom is designed to measure dose uniformity to the target and lung dose. The pediatric spine phantom is designed to test dose across the target in the spinal cord and evaluate dose titration to vertebral bodies largely with proton-directed therapy. The portfolio of phantoms provided by IROC Houston is robust and addresses an important role in the validation of dosimetry for clinical trials. The phantoms are generalizable, and credentialing can be reapplied to protocols once a site has completed the process and is approved for participation in the specific trial.

Credentialing extends into additional evaluations for knowledge assessment. This is very common in industry-sponsored clinical trials. In this circumstance, the institution will receive a case from the QA center and submit tumor and normal tissue contours with a radiation therapy treatment plan that would meet protocol specifications. This process serves an additional purpose of helping site investigators succeed with the data submission process. The data submission objects are reviewed by the QA center. Pre-treatment review of a similar case from an institution can serve as a benchmark case and be repurposed as a surrogate for a knowledge test. Completion of an end-to-end phantom can test all aspects of case development including assessment of the dosimetry of the planned approach to care including image guidance. Once completed and approved, the specific trial can be assured that the institution and identified providers of care possess the skill and have the clinical technology and infrastructure needed to successfully execute treatment for clinical study patients. As technology matures, credentialing becomes of increasing importance in assessing study outcomes. Credentialing for an individual trial can be repurposed and accepted by study investigators at the discretion of the study team and radiation oncology committee members [21, 22, 23]. A list of phantoms available for clinical trials is presented in Table 2.

Phantom type (anatomical location of disease)	Modality	Specialized technique being tested
Head and neck (H&N)	IMRT
Head and neck (H&N)	Proton
Lung	IMRT	Motion management
Lung	Proton	Motion management
Liver	IMRT	Motion management
Liver	Proton	Motion management
Spine	IMRT/3D
Spine	Proton
Head	Stereotactic	Small field dosimetry
Head	Proton
Prostate	IMRT
Prostate	Proton

Table 2.

Partial list of credentialing phantoms available at IROC Houston ([e-mail]).

2.4 Data acquisition and management

To ensure the quality of the data submitted to QA centers, the IROC process validates that institutions have forwarded accurate and complete information for protocol review. Protocols are complex, and data for individual patients often need to be submitted in sequence for management. For example, intermediate risk COG Hodgkin lymphoma trial AHOD0031 required diagnostic imaging review during chemotherapy to assess response and reassign patients into secondary randomization points based on response to therapy adjudicated by studying anatomic and metabolic imaging. Radiation therapy was a tertiary point of randomization based on the completeness of response on imaging. Outcome imaging was also acquired on the study to validate sites of failure and assess toxicity. Therefore, a substantial volume of data had to be acquired over multiple time points over years of participation by each study patient. Patient accrual on this study was 1733 patients. The data provide a wealth of information concerning assessment of disease and therapy response using anatomic and metabolic imaging and are an extraordinary resource for providing structure and imaging guidelines for the next generation of clinical trials in this disease. The success of this study was driven by the processes imbedded in IROC for data acquisition and data management. IROC works to reduce the burden of data submission by providing nimble tools for data submission with data anonymization with continuous improvements to automate the process. IROC uses protocol-specific scripts to perform real-time review of the data, assess completeness of the record, and standardize nomenclature including extraction header and additional identification information from DICOM files. Once submitted, users and members of the QA staff can evaluate and identify gaps in data. The process provides valuable feedback to institutions, which in turn provides an economy of scale for the data submission process and improves trial efficiency.

2.5 Case review

The process of case review ensures that each patient treated on study is cared for in a protocol-compliant manner. The protocol may require treatment approaches that may not fully align with physician preference or department approaches to care; however, the objective is to treat all patients on study in a uniform manner to ensure the study population is treated in a similar manner with respect to tumor dose/fractionation and normal tissue constraints. Although these are specified in the study, contours may be applied differently between institutions and investigators; therefore, the quality review process works to establish common ground between institutional management and study requirements. Imaging has evolved to be an indispensable component to the QA process. Imaging serves many roles in clinical trials. Imaging ensures that the correct patient is entered on the correct study. As evidenced in the medulloblastoma A9961 clinical trial, more than 10% of study participants were not eligible for the standard risk study due to retrospective review of imaging revealing more extensive disease than reported in the primary disease site in the posterior fossa and spine. Similar to asymmetry in the interpretation of response assessment in clinical trials, this has prompted investigators to acquire, manage, and review images in real-time by QA staff and study investigators to achieve the objective of creating a uniform study population including acquisition of outcome imaging to validate clinical outcome including site of failure. Important to radiation oncology and dosimetry, imaging has become essential to the mission for defining target volumes for radiation therapy, and a case review now requires all relevant anatomic and metabolic imaging to be available as part of the review process.

Historically when images had to be copied and data were transferred in hard copy, case reviews could only be performed in retrospect or at best, during the early phase of therapy. This had value, however, even when site investigators adjusted radiation therapy treatment plans in the first week of care at the request of QA review to meet protocol guidelines; patient survival in the HeadSTART trial was 8% less than if the plan met guidelines at the time of data submission. This prompted NCI group leaders and study investigators to investigate pre-treatment and real-time review of imaging and treatment objects to ensure patients on study had protocol-compliant treatment plans and the response to therapy using imaging objects can be validated and is uniform for study interpretation. Tools for digital data transfer for both imaging and radiation therapy today are readily available, nimble, and cost effective. Hundreds of thousands of dollars are lost when patients are deemed study ineligible post-study entry; therefore, making certain the correct patient is part of the correct study improves the confidence investigators have in the interpretation of study outcomes. Although time and experience are needed for institutions and their information transfer teams to be comfortable with digital data transfer and information de-identification, today’s real-time review has become a requested standard practice and is available on most studies requiring radiation therapy managed by the NCTN QA centers. Real-time review permits site and study investigators to review plans simultaneously to resolve gaps between site investigators’ care plans and study objectives. The process ensures that each patient on study can be treated in a study-compliant manner and that the goals of the study population can be achieved.

Once a study is completed, QA centers continue to collect data on study patients, including completion objects and outcome imaging. This is essential to clinical trial function as outcome imaging is our best vehicle to confirm the status of the patient. Outcome imaging supports our maturation as physicians as we learn to interpret the imprint of therapy on imaging and teaches us how to improve our interpretation of disease progression and therapy effect. The evolving field of artificial intelligence becomes more credible with large datasets to interpret patterns for quantitative analysis. Coupling protocol objects with patient outcome will support the entire oncology community and bring all of us to a broader understanding in applying imaging to outcome analysis.

There is an inherent danger on the horizon with respect to protocol management. There is a perception within studies that too much information and imaging are requested for data transfer to QA centers often hidden under the cloak of “real world” data management. We have sufficient information that too little data collection can lead to study interpretations that are invalid, limiting confidence in the final analysis of the study. The trial data are of vital importance, and limiting data acquisition will contribute to weakening confidence in the results of the trial. It is important to support data collection that is reasonable for the objectives of the study, and if a study has a simple endpoint, a more limited collection of data is reasonable. However, many modern trials ask sophisticated trial questions with secondary and tertiary points of randomization imbedded with response to therapy driving the secondary points of randomization. Trials of this nature require nimble data transfer, with QA centers playing an important role in sharing trial information in real-time between site and study investigators. Although there is a balance between data acquisition and cost, we need to avoid our mistakes of the past in placing a ceiling on data acquisition. The cost is in developing the infrastructure. Once in situ, data transfer is nimble and, in the end, highly cost effective if it improves clinical care.

2.6 Secondary analysis

An important aspect of data collection is to have the informatics infrastructure available to clinical trialists and investigators to ask questions with the data that were not anticipated at the time of trial design. We cannot envision every question before embarking into a clinical trial, and there are moments in clinical trial evaluation, we need to pivot mid-trial as objectives may change based on trial data evaluation and as additional data emerge to help us ask better questions. In this area, data from each of the NCTN Statistical Data Centers coupled with imaging and radiation therapy data housed at IROC are essential to be available both mid- and post-trial for additional analysis and review. In radiation oncology, analysis of acute and late effects from therapy is essential to refining and improving our service to patients. Treatment planning volumetric objects from radiation therapy need to be coupled with outcome imaging to assess dose-volume relationships to outcomes. In this manner, we can assess radiation dose to volume kinetics for toxicity and how the metrics are altered with chemoimmunotherapy. The information gathered on clinical trials is invaluable as the patient population was treated in a uniform format from multiple institutions. Therefore, when analyzed correctly, the data provide information that can be directly applied to clinical care and influence metrics for both tumor control and toxicity in the next iteration of clinical trials. This is an area where QA centers and NCTN data management can support The Cancer Imaging Archive (TCIA) in moving data to the national archive for all investigators to participate and repurpose for research. Data housed in this manner become extraordinary research for projects in artificial intelligence and other endeavors [1, 2, 3, 4, 5, 6, 7, 8].

3. Conclusions

Dosimetry is an essential component to the clinical trial process. Quantitative assessment of tumor coverage and normal tissue dose-volume data is the best vehicle to define metrics for tumor control and toxicity. Accurate and uniform dosimetry assessment is crucial to validating clinical trial outcomes. Real-time review of imaging and radiation oncology treatment objects pre-therapy can improve study compliance and secure protocol-compliant dosimetry for tumor targets and normal tissue contours and dose to volume. This generates a uniform study set and serves to limit deviations on study. In turn, the dataset becomes an important platform to evaluate both primary study objectives and secondary objectives not anticipated at the time of trial design. The data on study can be repurposed for multiple studies. For example, the intermediate risk Hodgkin lymphoma study AHOD0031 has generated multiple secondary papers including but not limited to evaluation of the pattern of failure with chemoradiotherapy and interpretation of response to bone lesions as seen on anatomic and metabolic positron tomography imaging. Data are available in nearly all oncology disease sites. Each study has objects that can be repurposed and provides an opportunity to study patients with unusual and exceptional responses to therapy as well as those patients who have biomarkers for therapeutic resistance and progress on primary therapy. These objects can be uploaded to TCIA with clinical information to be used by worldwide investigators for additional studies moving forward. The data are an extraordinary resource, and it remains our responsibility to provide the informatics infrastructure to move imaging and radiation oncology objects for study to a common platform in a public domain linked to pathology and genetic information as a DICOM object linked to the platform in a manner similar to an image. This has the potential of being a worldwide resource for all to learn [25, 26, 27, 28, 29, 30, 31, 32].

Acknowledgments

Supported in part by Imaging and Radiation Oncology Core (IROC) Grant, CA180803.

Conflict of interest

The authors declare no conflict of interest.