Physicists should perform reference planning for CBCT guided online adaptive radiotherapy

Abstract

Radiotherapy (RT) is at crossroads, as the technological improvements are pivotal in improving its treatment efficacy. This presents a paradoxical question: are the people driving the technology, or is it the technology driving the people in this field? In the recent years, one of the advancements that draw the most attention is the online adaptive radiotherapy (oART), which is known for its time consuming and resource intensive nature. As literature continues to pour in regarding knowledge sharing on how to best utilize this technology in terms of workflow, personnel training, and resources allocation, it is evident that a lot of critical steps in this process require expert attention. Even though commercial platforms largely streamline the process and build in automation for efficiency, there are still multiple steps that need to be performed manually or require further exploration. Amongst these steps, the reference planning specifically designed for oART is a critical step to ensure high quality and efficient treatment for each adaptive fraction. A contentious debate arises who should shoulder the responsibility in the reference planning for each oART session? Herein, we have invited a panel of medical physicists from four major institutions who are experts in oART to discuss their experience in building the oART workflow. The proposition for this parallel-opposed debate is “Physicists should perform reference planning for cone beam CT (CBCT) guided online adaptive radiotherapy.” Arguing in favor of the proposition, we have Dr. Mu-Han Lin and Dr. James Kavanaugh. On the opposing side, arguing against the proposition, we have Dr. Minsun Kim and Dr. Carlos E. Cardenas. Dr. Mu-Han Lin is an associate professor and the Director of Treatment Planning at the Department of Radiation Oncology, University of Texas Southwestern Medical Center. Dr. Lin also serves as the lead physicist of Ethos and Unity adaptive therapy service. Her research focus is translating and implementing artificial intelligence models and automations to improve clinical workflow. She and the team at UTSW have successfully implemented the AI malignancy prediction, AI synthetic CT, AI segmentation, and AI dose prediction modes into routine clinical use. She is also actively involved in the society such as the AAPM working group on treatment planning (WGTP) and the Task Group No. 395. Dr. James Kavanaugh is an Assistant Professor of Medical Physics at Mayo Clinic in Rochester, Minnesota. His clinical focus is leading the implementation and growth of the online adaptive radiotherapy program utilizing two CBCT-based Varian Ethos platforms. Research interests focus on the development of novel treatment delivery techniques, including spatial fractionation, knowledge-based planning, adaptive radiotherapy, and multi-criteria optimization, and the translation of these techniques to clinical practice. He is a member of the AAPM's Working Group on Treatment Planning and serves on AAPM Task Group No. 308. Dr. Minsun Kim received her Ph.D. degree in Applied Mathematics from the University of Washington (UW). Currently, she is an Associate Professor at UW and serves as the Clinical Physics Lead at the UW Medical Center. Her research focuses on the spatiotemporally optimal design of radiation treatment planning for better clinical outcomes and her clinical interests include adaptive radiotherapy and automation to improve patient safety and workflow efficiency. She is involved in the professional activities of the American Board of Radiology, American Association of Physicists in Medicine, The Radiosurgery Society, and NRG Oncology. She also serves as Physics Section Editor in Practical Radiation Oncology. Dr. Carlos E. Cardenas received his PhD degree in Medical Physics from The University of Texas MD Anderson Graduate School of Biomedical Sciences. Currently, he is an Assistant Professor and Director of Automated Treatment Planning in the Department of Radiation Oncology at the University of Alabama at Birmingham (UAB). His research laboratory focuses on leveraging artificial intelligence and automation to improve the quality of radiation oncology care. His clinical interests include the translation and evaluation of novel clinical automated solutions such as those used for adaptive radiotherapy. He provides clinical coverage for UAB's adaptive radiotherapy service, which utilizes the Varian Ethos treatment unit for online adaptive radiotherapy treatment delivery. He is a member of the AAPM's Working Group on Treatment Planning and serves on AAPM Task Groups No. 384 and No. 395 which are tasked with developing recommendations for online adaptive radiotherapy. Recent advancements in online imaging, AI-generated contouring, and auto-planning in RT have enabled the creation and successful clinical adoption of an efficient CBCT based oART platform.1-4 This oART process requires the creation of a reference plan generated on the traditional simulation CT and baseline anatomical contours. During this reference planning stage, a treatment planner is responsible for setting the formulas for margin expansion, deriving planning structures and optimization parameters. Subsequently, the oART process will use these predefined parameters from the reference plan to generate contours, automatically derive planning structure, and re-optimize the plan during each oART treatment fraction. Medical physicists’ responsibilities for a traditional linear static RT workflow include quality control checks (plan check, patient specific QA, weekly chart checks) throughout the planning and treatment process.5 Once a plan is approved, the treatment delivery accuracy falls on the attendings and radiation therapists through the on-board image matching and patient positioning process. This compartmentalized approach, bifurcating the treatment planning and treatment delivery, is not feasible within the CBCT-guided oART process, as the parameters used to create a reference plan has an intended direct impact on the quality of the available adaptive plans for each individual fraction. Therefore, it is crucial for the oART team to have a comprehensive understanding on both the treatment optimization and the adaptive plan creation when they need to make decisions on selecting and executing high quality treatment plans. Within the oART team, the physicist is the lone member with the technical knowledge, clinical treatment planning skillset, and innovative problem-solving expertise needed to ensure the quality of the adaptive reference plan. As such, we hold the position that the radiation oncology physicist should be responsible for generating the reference plan for CBCT-guided oART at the current stage. Medical physicists have the technical and planning experience required for the creation of quality baseline reference plans within the Ethos treatment planning environment. Initially, the intelligent optimization engine (IOE) in the Ethos TPS has a relatively steep learning curve for planners. Most of the optimization tools available in traditional treatment planning systems are hidden within the IOE and a planner only has indirect control of the process by modifying the ranking of desired dosimetric goals.6, 7 Good understanding on how the IOE translates ranked goals into a desired treatment plan requires clinical experience and practice using this new technology. Physicists have the best technical knowledge of the IOE and its performance on the resulted oART plans. Generating an oART reference plan requires additional considerations from planners, medical physicists are well-equipped with the clinical and physics considerations for online adaptive plan robustness. The reference plan for the CBCT-based oART should be robust against the potential changes in anatomy to ensure high quality adapted plan for each fraction. Planners should consider the potential uncertainties and take their impact into account when defining beam angles, derived optimization structures, and non-conflicting optimization objectives. This forward projected planning strategy is then translated into a clear, patient specific directive for the online adaptive planning. With the utilization of auto-contouring, there is no more “pushing” structures or volumes that planners can add on the fly since they cannot be easily reproduced during oART. Instead, the planning structures need to be pre-determined and, ideally, derivable from anatomical structures or targets, allowing the structure to be adjusted automatically in response to anatomy or target changes during the course of treatment. For plan generation, the IOE is driving the optimization based on the ranking of planning goals setup by the planners. These goals need to robust against future potential conflicts, taking into account both planning priorities and the dynamic relationships of organs/targets throughout the course of treatment. The knowledge on how anatomy can change and its associated impact on adaptive plan quality is gained through experience covering online adaptive cases. The online process is intense and time sensitive, requiring quick informed decisions in response to changing anatomy. Minimizing the variation in contouring/planning can help the team member covering oART to stay focused and avoid the stress of comprehending an unfamiliar planning strategy. Without sufficient experience of interacting with the IOE or detailed knowledge on contouring/optimization objectives used to generate the reference plan, it is difficult to make specific suggestions while at the machine that will result in a quality adaptive plan. Having the physicist creating the reference plan will give them the opportunity for hands on experience to identify the key aspects impacting plan quality and collaborate with physicians to develop the contour, planning, online imaging strategies for each individual treatment site. This kinetic can also stimulate more development ideas to improve the process. Indeed, the continued development of the CBCT-based adaptive treatment planning process within the clinic is necessary given the nascent state of the technology and its growing utilization. The technology is constantly being refined, with inclusion of more AI-powered auto-contouring, modifications to the optimization engines, improvements in dose calculation efficiency, and changes to automated systems/workflows. Additionally, direct feedback from physicians as their understanding of the tools and outcomes evolve will influence the goals and approach to ART planning. Accounting for rapid changes in both the technological tools and clinical objectives will influence how reference plans are constructed for future patients. Having the physicist as the creator of the reference plan will enable efficient changes to the planning process while simultaneously eliminating the possibility of the coverage team being unaware to these modifications, thus minimizing potential errors that could impact the quality of the plan delivered to the patient. Finally, the involvement of physicists in planning will help the physicists to better develop the checks to catch potential errors prior to impacting patient's treatments while simultaneously improving overall efficiency. This involvement should include the (1) planning phase: to effectively flag the deficiency in the reference plans and get it corrected in time; (2) physics chart check phase: to ensure the reference plan strikes the key aspects to enable a robust adaptive treatment; and (3) online treatment phase: to efficiently check patient setup, imaging, and contouring are all within the specifications to produce a high quality plan. A physicist with detailed involvement in all three phases will allow the development of a comprehensive series of quality controls that minimize the potential propagation of errors that impact the treatment delivered to the patient. In summary, the involvement of physicists in all aspects of the process, including reference planning, will ensure the creation of high quality plans in a rapidly changing technical and clinical environment. Additionally, physicists will gain experience through reference plan creation, and then provide feedback to further improve existing planning approaches, overcome challenges, and further advance the field by generating consensus and standardization guidelines in the adaptive planning process. The CBCT-guided oART is still in its infancy. Once the process is streamlined, other team members, such as dosimetrists, could also take on the reference planning seamlessly when the process reaches its maturity. This will speed up the widespread application of the online adaptive therapy to really change the profile of patient care. Similar successful examples, such as the development of SRS and SAbR, had physicists’ involvement in the planning significantly contribute to the standardization and generalization and inspired the developments in planning tools to make these advanced technologies widely implemented in all practices. Having physicist participating in adaptive strategy design is also in line with the future of medical physicist's role—more front facing to investigate and advocate for improvement to patient care.5 RT relies on the collaborative efforts of a multidisciplinary team, composed of highly skilled professionals trained in various disciplines, who work together to provide optimal patient care. Within this team, medical physicists in Radiation Oncology play a crucial role in ensuring the comprehensive safety of patients undergoing radiation therapy. Their responsibilities encompass diverse aspects, including quality assurance of treatment equipment, ensuring the high quality of patient-specific treatment plans, and assuring the safe delivery of treatments. As technology advances and becomes more complex, the role of medical physicists is evolving to adapt to these advancements while upholding patient safety. Medical physicists serve a key role in the clinical implementation of novel technologies, having the overarching goal of incorporating and overseeing new clinical processes introduced with new technologies while ensuring patient safety. Online ART is no exception. While their expertise and knowledge are essential for the successful implementation of oART, it is important to note that their vital role is in training and supporting other professionals to ensure safe and effective utilization of oART rather than taking over their jobs. When it comes to treatment planning, medical physicists are expected to provide comprehensive training and support to dosimetrists. This involves educating them about the principles and operation of the new technology, its potential applications, and the best practices for its use. Training sessions can include hands-on workshops, lectures, and case studies that enable dosimetrists to gain a thorough understanding of the new technology. These sessions also focus on addressing any potential challenges or pitfalls that may arise during implementation of an oART program. In addition, physicists provide ongoing support and guidance to dosimetrists as they navigate the complexities of the new technology. Physicists serve as a resource for troubleshooting technical issues, providing expert advice, and ensuring that treatment plans are accurately and precisely executed. By equipping dosimetrists with the necessary knowledge and skills, physicists empower them to acquire proficiency in optimizing patient plans with the new technology. Upon conducting proper initial training and establishing continuous communication with medical physicists, we believe that dosimetrists are able to develop the necessary expertise to create high-quality adaptive radiotherapy treatment plans. Their unique skill set and dedicated focus on treatment planning make them well-suited for the task, ensuring optimal patient outcomes and treatment efficacy. Dosimetrists are highly trained professionals with in-depth knowledge of treatment planning techniques and software. Certified medical dosimetrists are required to complete the necessary didactics and clinical experience before passing the board exam and take 50 continuing education credits for each 5-year cycle to maintain their knowledge and stay updated with the latest treatment planning techniques. They closely communicate with radiation oncologists to understand the clinical needs of each patient and this work environment allows them to gather insights from different perspectives and integrate them into adaptive treatment plans. Their capacity to communicate and coordinate with the team ensures that treatment adaptations are seamlessly implemented. While physicists have diverse responsibilities ranging from quality assurance of treatment equipment to safe delivery of patient plans, dosimetrists are dedicated to treatment planning and therefore have time allocated solely to the intricate task of developing optimal treatment plans. This exclusive focus allows them to devote the necessary time and attention to details, ensuring that each treatment plan is meticulously tailored to the individual patient's needs. Their dedicated time allocation enables them to thoroughly analyze a patient's clinical background, perform quality assurance checks, and collaborate with other clinical team members to refine and optimize treatment plans. By having this dedicated time, dosimetrists are also responsible for staying updated on any last-minute changes in the patient's treatment plan, as failure to incorporate these changes in a timely manner can potentially harm the patient. We believe that the collaborative relationship between medical physicists and dosimetrists allows them to further enhance their expertise and develop a deeper understanding of the technology's intricacies. By focusing on training dosimetrists instead of taking over their responsibility, medical physicists create a synergistic environment where multiple disciplines work together to leverage the benefits of new technologies. This approach not only ensures the efficient and effective integration of the technology into clinical practice but also promotes professional growth and development among the technical staff. Physicists can empower dosimetrists to take an active role in the treatment planning process and foster collaboration between the two disciplines, especially when it is a challenging process like online adaptive radiotherapy. This collaborative approach enhances the overall quality of patient care and ensures that the treatment plans are optimized for each individual patient. We believe that the most effective approach to ensuring patient safety and achieving high quality treatment plans lies in the synergistic effort among medical physicists, dosimetrists, and physicians. By leveraging the expertise and contributions of each team member, we can optimize patient care. This collaborative effort, where each professional performs at their best, leads to enhanced patient safety and improved clinical outcomes. First, we would like to start with the points from our opponents that we agree with. Medical physicists play a crucial role in the clinical implementation of novel procedures. Lectures, hands-on workshops, and case studies on CBCT-based oART are good educational opportunities for dosimetrists and other personnel involved. Additionally, we are in an agreement that a close partnership between physics and dosimetry during the planning process will leverage the expertise of both parties in developing customized patient specific treatment plans. However, we argue for the unique responsibilities and vital roles of medical physicists throughout the planning and delivery of oART. Specifically, the critical knowledge and practical understanding on creating adaptive plans from reference plans in daily adaptive session requires medical physicists to develop and maintain hands-on experience with the auto-planning features in the CBCT based oART system. Unlike conventional planning that is well studied with a wealth of knowledge in literature, the CBCT-based oART planning, as an emerging technology, still has many clinical aspects being investigated and may present significant challenges to staff members who are not as experienced as well-trained medical physicists. Even for medical physicists with extensive experience in the conventional planning, the auto-planning features in the CBCT-based oART TPS are unique and require specialized training. Hence, the hands-on planning experience with this system is necessary for a medical physicist to better oversee the on-treatment procedure and navigate the auto-planning, for which the planner cannot see the hyper-parameters for optimization. The goal of medical physicist performing reference planning at the current phase is NOT to take over dosimetrists’ job. Instead, the goal is to develop a stronger collaboration between medical physicists and dosimetrists during the ongoing implementation and advancement CBCT-based oART. As our opponents mentioned, dosimetrists are highly trained professionals and with proper guidance, they will be able to perform the reference planning very well. However, CBCT-based oART introduces brand new auto-planning concepts and tools that not allow an adoption with the existing and expertise in understanding of the reference plan and adaptive when to on a simulation CT and potential changes present on daily CBCT is critical for the physicist covering the adaptive especially when plan is necessary to approach involves extensive planning and development before implementation of technology. resources would be required for oART development this development approach, since one would need to the oART process to the of an direct involvement in the reference planning is a more efficient and effective to build robust adaptive Training dosimetrists on would be the opponents have that one important role of the is to closely work with the radiation to their clinical goals into a treatment plan. They can direct feedback from the and make allowing for a tailored plan that each patient's needs. Within the traditional workflow, the role is to these However, within the oART workflow the close collaboration between and planner while the adaptive from the on how the changing anatomy and impact on the adaptive plan a improvement for future planning. In most it is the not who is present at each adaptive and can this feedback from the into changes in the planning process. agree that physicist's involvement in adaptive therapy planning is opponents that physicists can get involved in the adaptive contouring and deriving structure setup before planning and planning strategies before the plan These are However, this will add to the time from simulation to treatment With their process, how many institutions can to add these steps during planning prior to practical knowledge about the how effective can a physicist be to identify the pitfalls of planning guidance and multiple of planning the between dosimetrists and a constructed reference plan it to the on-treatment adaptive it will of a quality plan being generated and the covering team would need to the the quality plan, or change at the treatment machine to improve the plan. the physicists are not with the of the how can physicists be able to oversee or direct the oART process and in For in the in which the of oART not generate an plan, should a be changes can be to make the In we on our proposition that physicists should perform reference planning for CBCT guided oART at the current A partnership with dosimetry while generating reference plans will allow for the physicist to maintain their knowledge and associated with the plan creation process, incorporate dosimetry training, and insights into the development of oART treatment As the oART technology and workflow a of planning from physics to dosimetry can for and that have well Additionally, medical physicists can use the in-depth knowledge and practical experience gained from simulation to treatment of oART to clinical operation in their and further on to other institutions who are oART. We agree that it is crucial for the oART team to have a comprehensive understanding of the treatment planning and adaptive plan creation We also agree that the being a member with technical knowledge and a pivotal role in ensuring the quality of the reference plan. However, this not the physicist taking over the role of creating the reference plan, it that the physicist's plan in Unlike adaptive plans for treatment sessions with time the reference plan is with time and sufficient resources available to the In such an diverse perspectives and from both disciplines, dosimetrists and physicists, can improve the quality of the plan and patient safety. This approach in to one that both physicists and dosimetrists contribute perspectives and to patient care. The proposition that physics of planning ensures proper of contours, and planning we that of predefined patient-specific planning parameters could be performed prior to planning to any could be implemented in the planning workflow to ensure prior to plan optimization within the We agree that the planning techniques and optimization tools within the IOE significantly from those in traditional However, our that the to knowledge and expertise to dosimetrists lies with the medical rather than the role of planners. While physicists may a deeper of the optimization and hidden within the Ethos it is essential to that dosimetrists a of planning diverse patient and clinical medical physicists can contribute by developing planning that dosimetry can use as points in the planning process. have and their experience developing Ethos TPS planning which to further optimize the planning process and improve plan these planning and for evaluation and could further enhance medical physicists’ role in adaptive treatment leading to a implementation of planning by safety is also to a through rather than solely on the of an medical physicists responsibility is treatment planning, this could be a However, in the of medical physicists have diverse responsibilities various aspects of patient care. They may not be as as dosimetrists in the which can potentially patient safety not We are of the that there are to physicists treatment planning in terms of resource allocation and patient the physicist is not dedicated to the planning process. We that the planning strategy required for a reference plan to robust against future treatment adaptations is important in a reference plan. The of an online adaptive planning with its during the reference that this is novel to both physicists and dosimetrists, it the for collaborative efforts between these two disciplines, each their unique This the of prior training before implementing new technology in clinical with institutions that are of novel technologies and that knowledge to and staff a crucial of physics We with the of Lin and Kavanaugh regarding the of incorporating knowledge from the reference plan when adaptive plans time is the knowledge that physicists gain during the adaptive planning process at the which to reference plan This the of collaborative for the of the online adaptive radiotherapy program. The of a robust feedback between physicists and dosimetrists as the in the implementation of oART for enhanced patient care. This feedback the working among the crucial team When it a more robust environment for the field and, improving patient care. The have to The no of