Clinical Considerations for Modern Dosimetry and Future Directions for Treatment Planning

Abstract

Technology and computational analytics are moving forward at an extraordinary rate with changes in patient care and department workflows. This rapid pace of change often requires initiating and maintaining the educational support at multiple levels to introduce technology to radiation oncology staff members. Modern physics quality assurance and dosimetry treatment planning now require expertise beyond traditional skill based in computational algorithms and image management including quality assurance of the process of image acquisition and fusion of image datasets. Expertise in volumetric anatomy and normal tissue contouring are skills now performed by physics/dosimetry in collaboration with physicians and these skills are required in modern physics dosimetry training programs. In this chapter, challenges of modern radiation planning are reviewed for each disease site. Skills including future applications of image integration into planning objects and the future utility of artificial intelligence in modern radiation therapy treatment planning are reviewed as these issues will need to be added to modern training programs.

1. Introduction

Prior to the advent of volumetric radiation oncology treatment planning, physician, physics, and dosimetry teams would construct and calculate radiation therapy treatment plans at the center of the target referred to as the isocenter. Calculations would be derived based on depth measured at isocenter. Beam shaping devices which shaped dose at the isocenter were applied to the sloping surface of the target at a single level. Plans would be calculated to isodose lines which would unintentionally not define the volume and location of areas of radiation dose asymmetry. In breast cancer patients, the areas of asymmetry would be at the medial and lateral regions of the breast in rib/chest wall and extend for the length of the field which by default would include multiple rib segments. In this era injury, when it occurred, was simply ascribed to radiation therapy with limited attention to dose and volume treated. Radiation is not a drug; however, we did not have volumetric computational tools to be more exact in our review of process and convince our medical and surgical colleagues otherwise. We could not determine a specific dose volume effect relative to injury as we did not have tools to validate this point. Our approach to treatment planning changed with the introduction of computer tomography into simulation and volumetric driven radiation therapy treatment planning. With tools for contouring targets with reconstruction algorithms, radiation oncologists and treatment planning teams could visualize targets in three and four-dimensional volumes and review the juxtaposition of target volumes with normal tissue structures. This provided radiation oncologists opportunities to apply therapy in non-coplanar modulated geometries with beam arrangements that were more specific to each patient’s target volume.

Radiation therapy treatment planning and dose prescription permanently changed with the introduction of advanced technology. Dose was prescribed relative to volume, not isodose lines, and contouring normal tissues provided the infrastructure to develop strategies for conformal avoidance of normal tissue. Altering fluence profiles by moving multi-leaf collimators across radiation therapy treatment fields provide the opportunity to generate voxel-based dosimetry to further improve dose asymmetry to tumor targets and place sharper dose gradients across normal tissue targets decreasing the risk potential for injury. The weakness in target specificity was therapy reproducibility and image validation of the targets before and during therapy. This was addressed with several manufacturing improvements including the integration of cone beam computer tomography into linear accelerators, use of four-dimensional treatment planning to develop therapy targets, and optical tracking tools to validate the lack of motion during treatment. Volume modulated arc treatment delivery provided the opportunity to decrease the time of treatment delivery with simultaneous multi-leaf and gantry motion. By decreasing treatment time delivery both durability and reproducibility of daily positioning could be confirmed. Motion, including deep inspiration breathing, is now validated with optical tracking systems. Decreasing the time of treatment delivery with volume modulated arc therapy provides confidence that patient care will not be influenced by intra-fractional motion.

These improvements have served to secure the success of radiation oncology moving forward as tumor targets are treated accurately with confidence and normal tissue protection is optimized. The improvements have also served to change the work scope and skill set of colleagues in the radiation therapy physics and treatment planning community. With the advent of volume-based planning, image integration into targeting has become the standard of care. With two-dimensional radiation therapy treatment planning, often the information needed to generate a standard treatment plan was fully available to the dosimetry and physics teams at the end of the simulation hour. Today, most of the work in planning and targeting is performed after the simulation process. The simulation hour is used to create devices for immobilization, perform three- and four-dimensional imaging, and establish target coordinates for planning. Physicians contour targets for treatment after images are acquired and processed, often with diagnostic images fused into radiation therapy planning images. The work of the planning team cannot begin in earnest until the targets for therapy are contoured and constraints for normal tissues are defined for the objectives of the planned therapy. If there are delays in physician contouring, unintentional time constraint can be placed onto the planning and treatment validation process. The planning teams need to be well versed in volumetric therapy language as clinical, motion, and planning targets are applied to the intended areas of therapy by planning teams following protocol and/or institutional guidelines. The plan is developed as best as possible within the confines of the normal tissue volume constraints and validated through the quality assurance process. Image guidance and tracking process for quality assurance is initiated and maintained by the planning teams in collaboration with the therapy teams. Planning teams are essential in all services housed within the department of radiation oncology from the time of simulation to treatment validation. Planning teams are involved in brachytherapy and stereotactic therapy with varied imaging and dose computational algorithms required for modern patient care. The skill set for the modern planning team is diverse requiring knowledge of all aspects of modern planning equipment and tools.

Therefore, the role of the planning team in radiation oncology has expanded to image-based volumetric dosimetry and plan validation. Dosimetrists now have an extended role in defining volumetric anatomy and plan validation. In this chapter we will review skills required by dosimetry and planning teams in each disease and discipline area; the role of imaging and dosimetry in both daily work scope and clinical trials; the skill set for the modern planning team; and define what a modern planning group might resemble soon [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11].

2. Central nervous system

Patients requiring radiation therapy to the central nervous system (CNS) require a broad range of department services as these patients comprise both pediatric and adult populations including patients with primary and metastatic disease. Patients can require therapy to the entire CNS as well as stereotactic therapy to small targets with sharp dose gradients across critical normal tissues. The objectives for each patient have similarities with the primary goal to successfully treat the tumor target with conformal avoidance to as much central nervous parenchyma and critical structures as feasible and not compromise dose to target. In both pediatric and adult populations, sparing normal tissue now has near equivalent importance in patient management to tumor control and this has become essential to the treatment planning community. The CNS has multiple critical normal tissue structures with limited self-renewal capacity, therefore conformal avoidance when possible is important for optimal clinical outcome. Imaging plays an essential role in defining targets and accurate contouring is fully dependent on image fusion and quality. Most tumor structures are not well visualized on computer tomography. Fusion software is aligned with bony anatomy and the irregular shape of skull landmarks lends itself to accurate integration of multiple image sets for contouring. Investigators have developed protocols in glioblastoma multiforme using multiple magnetic resonance (MR) sequences involving spectroscopy, fluid-attenuated inversion recovery (FLAIR), and contrast images using dose painting strategies to help limit dose to critical structures. Spectroscopy is helpful when tumor abuts the corpus callosum and can define areas where disease extends to the contralateral hemisphere, FLAIR defines edema which can house disease, and contrast defines regions of blood-brain barrier disruption by disease. Positron tomography imaging with amino acids can define tumor in deoxyribonucleic acid (DNA) synthesis often not well visualized with gadolinium. The datasets help create multiple target volumes which can be treated as a single plan with individual fractionation and total dose to each target. For patients with metastatic disease, treatments are delivered with radiation treatment planning including radiosurgery to subtotal volume CNS targets and hippocampal sparing for improved neurocognition. The growth of MR imaging has facilitated the development of subtotal volume therapy. Pediatric radiation therapy on selected germ cell protocols is delivered to spinal fluid pathways with temporal lobe sparing and standard risk medulloblastoma therapy boosts are now planned to image targets and not the entire posterior fossa in order to spare normal tissue. The plans require creativity with a balance of constraints between multiple normal tissue targets with dose limitations applied to the CNS tissue volume in the treatment field.

Often tumor targets come in close approximation to normal tissue and planning teams need to be fluent in multiple aspects of field geometry including table motion, off-axis fields, and six-degree couch motion and place dose gradients across structures including optic nerves, chiasm, and the cochlea when needed. Artificial intelligence (AI) will have influence in this aspect of care as field design can be optimized to constraints through an iterative process once the contours have been drawn and processed. The plan, once approved, is validated through a quality assurance process and treatment can begin once the patient’s therapy portal images are generated and approved. Figure 1 is the treatment plan of a patient with neurofibromatosis with an astrocytoma in the posterior fossa occupying the fourth ventricle showing high and intermediate risk volumes defined on MR with conformal avoidance of the cochlea.

Figure 1.

Dose painting intensity-modulated radiation therapy (IMRT) plan for a posterior fossa glioma.

Planning for diseases in the CNS is clinically important as normal tissues of the CNS have limited self-renewal potential, therefore conformal avoidance to as many structures as possible with radiosurgery (SRS) and stereotactic radiation therapy is essential for outcome. Few diseases alter the well-being of the patient more than injury to the CNS imposed by disease and treatment. Limiting sequelae of therapy is an essential goal for the dosimetry team [12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23]. Further improvements in small field dosimetry permit multiple lesions to be treated in a segmental manner with a single plan with one isocenter. Figure 2 is an example of a single isocenter plan simultaneously treating multiple lesions in the CNS using the Varian RapidPlan system with volume modulated arcs. The arc permits simultaneous dynamic motion of the treatment gantry and multi-leaf motion to optimize delivery to tumor and limit dose to normal tissue.

Figure 2.

Dose delivery to multiple lesions with a mono-isocenter in the central nervous system using a single plan to treat all lesions. (A) Pre-SRS MRI; and (B) Isodose lines (Rx = 18Gy to 11 tumors in single fraction treatment) on post treatment MRI 8 months later.

3. Head/neck

With the advent and recognition of human papilloma virus (HPV), the incidence of head and neck carcinoma has significantly increased including patients who are lifelong non-smokers. These are challenging patients to plan as there are a multitude of normal tissues to provide both conformal avoidance and titration of dose asymmetry. Our knowledge of dose constraints continues to mature and we are applying strategy to as many normal tissue volumes as the primary target will permit. Spinal cord dose is limited to 50 Gy to 1% when feasible. Dose volume constraints are applied to the mandible, retropharyngeal muscles, carotid vessels, larynx, brachial plexus, and thyroid when possible. When tumors involve the paranasal sinus and skull base, constraints need to be applied to orbital structures, pituitary gland, optic chiasm, cranial nerves and temporal lobe including the hippocampus when possible. These constraints are balanced with tumor location and normal tissue anatomy coupled with knowledge of tissue molecular biomarkers.

In these patients, gross tumor volume (GTV) is often well defined on anatomic and metabolic imaging coupled with findings on physical examination. This permits more expert targeting of disease location as well as clinical target volumes (CTV) thought to be at risk for disease, often considered one lymph node station beyond gross tumor. Planning target volumes (PTV) are designed for daily set up variability, however with the advent of optical tracking tools and improved image guidance, the traditional need for a 5 mm PTV can be re-visited as arbitrary application of expanded targets can often extend target dose into normal tissue structures and outside of the patient if not carefully applied. Vertebral bodies can often be natural barriers especially if volume expansion places the spinal cord at risk for additional radiation dose. Modern image guidance has likewise improved daily reproducibility of radiation therapy treatment. This allows departments to re-visit the concept of a PTV since daily patient treatment is more secure and consistent. Titrating the volume will decrease radiation dose asymmetry and improve clinical outcome.

In applying constraints, it will be important for therapy teams to track outcome through pathways previously less well studied. For example, although it is likely that the dose/volume relationships of retropharyngeal tissue treated influences outcome, recognizing this is driven in large part by the location of primary disease. We need strategies including speech/swallowing colleagues to study this effect and learn where to dose/volume reduce when feasible. Audiologists will help radiation oncologists apply metrics to outcome for alterations in hearing imposed by therapy. Building a portfolio for outcome analysis will support process improvements in radiation therapy planning and support the identification of patients who would potentially benefit from supportive intervention prior to the appearance of visible sequelae of management. Figure 3 represents the plan of a patient with recurrent paranasal sinus esthesioneuroblastoma with conformal avoidance of the optic chiasm.

Figure 3.

IMRT plan for ethmoid sinus esthesioneuroblastoma. Note conformal avoidance of the globes.

This is a growing population of patients who will benefit from the attention to detail required for optimal tumor control and normal tissue function [21, 24, 25, 26, 27, 28, 29, 30].

4. Thorax

Lung cancer has evolved over the past two decades from a disease of habit to a disease of molecular biology. Lesions are now treated with multiple techniques including discontinuous planning such as dose painting/altered fractionation to peripheral nodules and more traditional fractionation to regions of central mediastinal disease where tumor abuts critical central structures. With the advent of immunotherapy, the situation has become more complex as toxicity can occur in both high and low dose regions making planning and dose constraints challenging to meet. Lung tumors can be large and often are in less favorable thoracic locations to meet cardiac and pulmonary constraints. Accordingly, modern lung cancer protocols that include immune-radiotherapy place strict constraints on V20 and V5 with a dose ceiling of 60 Gy. To meet constraints, nearly all modern industry and National Cancer Institute’s (NCI) National Clinical Trials Network (NCTN) clinical trials treat limited to no elective at risk volumes. Only gross tumor as defined on anatomic and metabolic imaging is often contoured with dose constraints placed on cardiopulmonary volumes, soft tissue, chest wall, and spinal cord.

Thoracic planning will remain an enigma with imperfections applying dose to structures. Most thoracic plans are now performed with intensity modulation. If left to its own device, intensity modulation will titrate dose asymmetry (hot spots) but unfortunately push dose to vulnerable pulmonary parenchyma if a strict low dose (V5) and moderate dose (V20) constraint is applied. When these constraints are applied, dose will be driven back to high dose segments which in turn will create hot spots, largely in anterior/posterior soft tissue intentionally lateral to the spinal cord. This has the potential of increasing dose to the chest wall and soft tissue structures. Often this is viewed as acceptable in order to prioritize more limited dose to pulmonary parenchyma. This results in the need for balance of constraints.

As the number of cancer survivors grow [31], the modern cancer patient is asking different questions concerning outcome beyond the question of tumor control. Thoracic malignancies including primary lung, esophagus, and lymphoma can generate dose to vascular structures and therefore cardiac dosimetry is an important element to thoracic therapy. Historical therapy directed to the heart as an unintentional target is associated with coronary artery disease, myocardial dyskinesis, valvular disease, electrical conduction changes, and pericardial disease. It is now also recognized that as blood migrates through chambers during treatment, radiation can decrease white cells and other blood elements. Once thought to be exclusively related to marrow dose, the heart becomes a vehicle for immune suppression through therapy. Therefore, for all epithelial and liquid diseases of the thorax, attention to detail to the heart and cardiac subsegments need to be assigned constraints when feasible. This information becomes invaluable to both the cardiologist and primary care provider teams in evaluating the patient and creating a survivorship plan including cardiac prevention strategies. We cannot evaluate radiation as a “drug” and to optimize survivorship programs, defining dose to subsegments and vessels will provide meaningful information to patients and care providers. Future AI tools will help optimize consistent application of contours to subsegments which will optimize strategies for conformal avoidance by the planning team. The esophagus abuts the left atrium; therefore, radiation oncologists can provide information on dose to the electrical conduction system as a cardiac subsegment as part of survivorship planning as tumor target will abut the posterior wall of the atrium. Improvements such as this instill confidence in providers and patients recognizing that we place value on outcome. The same approach can be adopted to other normal tissue volumes including pulmonary parenchyma. Often thoracic lung cancer patients have compromised baseline function with limited pulmonary capacity, therefore conformal avoidance of parenchyma is important.

Future strategies for application of tools for planning will include functional coefficients for cardiopulmonary volumes. Currently we can only assign anatomical coefficients without recognizing that function may be an important component to future planning paradigms. Lung will remain a balance of constraints including low/intermediate, and high dose parenchymal segments as well as cardiac, esophageal, spinal cord, the chest wall volumes. Integrated databases with images, radiation therapy plans, and outcome with help us further refine planning strategies for thoracic oncology patients (Figure 4) [32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63].

Figure 4.

Esophagus patient with a pre-operative radiation therapy treatment plan demonstrating conformal avoidance of cardiac and pulmonary volumes.

5. Abdomen/pelvis

Radiation therapy plays an important role in the treatment of abdominal and pelvic malignancies which include the need for stereotactic therapy and brachytherapy. This requires an expanded skill set for the planning team as the team must prepare for additional therapy superimposed on teletherapy.

In the past two decades the liver has become an important focus for radiation therapy. Systemic therapies have improved for metastatic disease and primary hepatocellular carcinoma has significantly increased in incidence; therefore, radiation oncologists are applying advanced technology tools to the management of these patients including stereotactic therapy for definitive management and as a bridge to transplant. Planning teams become fluent in fusion of multiple MR and metabolic image sets into planning computer tomography including motion management techniques for successful treatment planning and delivery. Liver targets are often difficult to visualize in early iterative versions of cone beam computer tomography to validate target positioning for therapy, therefore planning teams have used versions of fiducial tracking to validate positioning for treatment set up. Multiple constraints are applied to liver targets including mean dose and partial volume dose. Constraints are applied to gastric/small tissues in close approximation to the liver as well as cardiac, pulmonary, and renal constraints applied in selected areas relative to target volume location. This is an important area of clinical research as radiation therapy is also being supplied via radio pharmacy with Y-90 and other compounds in development. Additional theranostic therapies have hepatic uptake, therefore this dose will need to be calculated as well for dose volume analysis.

Volumetric dosimetry is needed in this area. Although Y-90 can be applied to the region of disease, tumor vascularity may prevent uniform application of dose and the area of intended therapy may be underdosed with migration of therapy away from the intended target. The strategy for volumetric dosimetry with both diffusion kinetics and evaluation of migration is being developed and will likely include serial single positron computer tomography images obtained frequently (daily) and fused into a planning computer tomography to evaluate areas of disease potentially undertreated which can then be augmented with stereotactic therapy. Likewise, areas of dose migration can be identified as regions of conformal avoidance for the radiation planning team through these processes. Modern planning teams need to be nimble in image fusion and registration in order to optimize patient care in this group of patients. Figure 5 represents a stereotactic hepatic radiosurgery treatment.

Figure 5.

Hepatic stereotactic body radiosurgery plan.

Upper abdominal therapy is often directed to biliary, pancreatic, gastric, and lower esophageal targets. In this cohort of patients often bowel, renal, liver, and occasionally cardio/pulmonary constraints need to be applied with thought. In patients treated in a post-operative setting, tolerance to bowel may need to be titrated, especially if it is devascularized and fixed in position. This is commonly seen after pelvic surgery. Likewise, the regions of the gastrointestinal and biliary anastomoses must be identified for conformal avoidance/dose titration as best as possible. Invariably, these areas can reside in high-risk regions, therefore advanced planning techniques need to be applied to optimize patient care and avoid injury.

For treatment to the pelvis, often bowel, bladder, and rectum are considered targets at risk. Mucosal surfaces of these organs are tissues of self-renewal potential and injury is often related to limitations in cell re-growth along the mucosal surface resulting in bleeding/nerve exposure resulting in pain. Long term effects such as perforation are unusual appear to center of areas of previous surgical intervention which are inherently devascularized/fixed in position. This is true for all surgical colleagues including gynecological oncologists, urologists, colorectal surgeons, and surgical oncologists. When possible, these areas need to be identified pre- radiation therapy for dose titration when not intentionally included in high-risk tumor volumes. Although we can treat more tissue than a surgeon can remove including extended nodal volumes, outcomes in post-operative patients require careful exchange between the radiation oncologist and surgical colleagues in designing target volumes of interest to optimize patient care and place dose gradients across tissues considered vulnerable to injury.

There is increasing interest in the use of radiation therapy for abdominal and pelvic malignancies and use of stereotactic techniques when feasible to limit sequelae of management. The modern physics team will increasingly use advanced technology in the care of these patients. As we improve our technology, we must be cognizant of knowledge moving forward. For example, most in the radiation oncology community were unaware of insufficiency fractures in the sacrum associated with radiation therapy. However, with symmetric application of extended targets beyond the gross tumor volume defined by pre-sacral lymph nodes, if one is not careful full dose can be applied through the planning target volume (PTV) if the targets are applied in a symmetric manner. With modern MR sequences, we see the fractures on occasion at radiation doses under 6000 cGy, therefore it is important to place dose gradients across the sacrum and try not to place full dose across the entire structure. Investigators have found the pre-therapy exercise programs designed to maintain flexibility supports treatment reproducibility. This will become an important aspect of survivorship programs [64, 65, 66, 67, 68, 69, 70, 71, 72].

6. Breast

Patients with breast cancer will continue to require significant attention to detail by physics and dosimetry in the development of their care path for radiation therapy treatment planning. Because of irregular topography and multiple sloped surfaces of the breast which are patient specific, breast planning will not easily succumb to exclusive planning through AI. Optimizing radiation therapy treatment plans remains a balance of constraints. With modern treatment planning techniques, most breast geometries can be planned with no more than 5% of the breast target receiving 107% of the prescription dose when treatment is applied to breast tissue. Contouring, however, must be carefully done as breast requires both the radiation oncologist and the planning team to simultaneously think in 2, 3, and 4 dimensions to achieve objectives of management including contouring regional anatomy and establishing goals. If the oncologist contours from midline to the latissimus muscle in a left-sided breast patient, it is difficult to meet normal tissue constraints for the left ventricle, mean heart dose, and left pulmonary parenchyma as defined by most clinical trials. If arcs are applied, often constraints cannot be met for total lung and contralateral breast if tumor target dose homogeneity constraints are too strict. In these circumstances, the planning engine wants to please the planning team, therefore if constraints to normal tissue are not included in the planning strategy, the engines will work to limit radiation dose asymmetry through the target. However, the more homogeneous the plan becomes through the therapy target/breast, the more dose is “pushed” into normal tissue including heart/lung, and contralateral breast. Therefore, the planning teams and the radiation oncologist must provide balance to constraints to meet both tumor coverage without exaggerated dose to target volumes and limit dose to normal tissue. One of the key elements to planning breast radiation is to contour breast tissue without intentionally contouring volumes to midline or posteriorly to the latissimus. When possible, contour breast tissue including the surgical cavity in the target volume. Because of applying underlying cardiac and pulmonary constraints, dose will naturally be “pushed” to the medial and lateral chest wall parenchyma. If one contours to midline, dose can easily be pushed well into the contralateral breast without intent due to the attempt to provide full coverage to this point and well as the latissimus laterally. As the target becomes more inferior, therefore is less breast tissue in both the medial and lateral planes, providing an opportunity to titrate volumes in these locations and limit cardiac dose for left-sided patients. Therefore, contouring becomes a very important aspect to breast cancer radiation therapy. If the volumes are over-drawn or drawn in a casual manner, there is less physics and dosimetry can do to improve the situation. In a similar manner, strategies for regional care need to be carefully designed to balance dose to target and normal tissue. The internal mammary lymph nodes follow the anatomy of the internal mammary artery and there is direct drainage to level 3 and the sub-pectoral region as well as the axilla, therefore strategy must be applied to extent of regional coverage as seen in Figure 6 as normal tissue dose is influenced by the contour and dose assigned to the target. Although anatomic guidelines have been established for clinical protocols, recent publications have suggested that regional failures can occur beyond the target volumes defined from anatomical guidelines, therefore we must identify high risk patients and possibly extend our field edges beyond our traditional boundaries. Patients with triple negative disease often now undergo positron emission tomography prior to initiating therapy. In these selected patients, nodal disease is at times identified beyond traditional field boundaries, therefore volume of regional disease now is driven by imaging.

Figure 6.

MRT plan for breast radiotherapy with intentional (left) and unintentional (R) inclusion of internal mammary structures. Note difference in cardiac dose.

Breast remains a disease requiring thoughtful application of radiation therapy to the breast, surgical cavity, and regional targets. Because outcomes remain excellent, attention to detail including normal tissue constraints is essential to modern planning [73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91].

7. Extremities

For both adults and children, treatment of the extremities requires significant attention to detail. There are multiple tissue compartments that can be affected by therapy and each can have consequence for late effects of management. These can include major blood vessels, bone, joint spaces, nerve, and muscle bundles. Children require additional conformal avoidance to growth plates when possible for optimal outcome for growth and development. This is an area where natural barriers may be helpful in placing dose gradients across structures to limit late effects, especially bone and joint spaces.

Because the extremities resemble cylinders, volume modulated arcs often play an important role in radiation therapy planning. It is important to spare a strip of normal tissue as it extends through the cylinder. Although most oncologists try to generate the strip in a linear manner, is it likewise clinically acceptable to spare tissue in a non-linear serpentine manner if the strip remains non-interrupted. Of equal importance is the concept of immobilization. Historically, radiation oncologists approached this from a perspective that casts and rigid structures provided the optimal security for treatment reproducibility. However, even within a rigid structure, alterations in anatomy can occur either on a pre- or post-operative basis rendering the utility of rigid structures. Modern alignment and optical tracking tools have altered our approach to this situation. With the generation of these tools, we can build a “virtual cast” and maintain alignment without the rigor of rigid tools. Patients are more comfortable, we are more secure in daily treatment reproducibility, and treatment time is considerably decreased further contributing to security in daily set up. The volumes in these cases lend well to advanced technology therapy and conformal avoidance of key structures can be more readily accomplished. Figure 7 represents a patient with an extremity lymphoma with popliteal adenopathy using volume modulated arc treatment to treat the circumferential target with sparing oof posterior soft tissue for lymphatic drainage [92, 93, 94, 95, 96, 97, 98, 99].

Figure 7.

Near circumferential extremity target with popliteal adenopathy. Note sparing of normal tissue in the posterior compartment.

8. Pediatrics

Pediatrics is a special subgroup of patients requiring considerable care and attention to detail. This population requires many departments within the hospital setting including anesthesia and social services to manage and optimize patient care. Often the targets treated in children are both similar and dissimilar to adult counterparts. CNS tumors often require similar normal tissue avoidance strategies with careful attention to the volume of temporal lobe treated including essential structures including but not limited to the chiasm, cochlea, retina, and brain stem. Diseases including germ cell tumors and medulloblastoma require therapy to the CNS fluid pathways, therefore strategy of care is different than adults and includes the craniospinal axis. Intensity modulation and protons have been used to limit exit dose to cardiac, pulmonary, liver, bowel, vertebral body, and renal volumes. Pediatric diseases often requiring whole lung therapy and strategies are now more routine to provide conformal avoidance of the heart using four-dimensional technology and volume modulated arc therapy. Neuroblastoma, sarcoma and other soft/bone tissue primary lesions often require advanced technology therapy application with a growing interest in theranostics for neuroblastoma care. In NCTN clinical trials, more than 25% of pediatric patients are now treated with protons when they receive radiation therapy, and this number is expected to increase moving forward. Figure 8 is a 7-year-old patient requiring whole lung therapy being treated with cardiac avoidance. Note the ability to limit dose to the cardiac structures using volume modulated arc therapy [93, 99, 100, 101, 102, 103, 104, 105, 106].

Figure 8.

Conformal avoidance of the heart in a 7-year-old patient during whole lung radiotherapy.

9. Imaging in radiation oncology

With the advent of volumetric radiation therapy planning systems, imaging has become essential for modern therapy including dosimetry. Prior to the development of computer tomography-based simulation, patients were treated with two-dimensional planning with fluoroscopy simulators. Computer tomography simulation was a paradigm shift in radiation oncology. Today imaging is the infrastructure to all elements of activity in the department. Thoracic and upper abdominal patients are simulated with four-dimensional imaging. Many head and neck and nearly all CNS patients are planned with fusion imaging to accurately define target volumes of interest. Many CNS patients are now planned with multiple MR sequences which when applied are used for dose painting on clinical protocols using FLAIR, spectroscopy, and contrast as areas to target. Cone beam computer tomography has been incorporated into linear accelerators for daily target validation and can be applied for adaptive radiation therapy planning. Portal imagers have a dual role as a dosimeter. Single photon emission computed tomography (SPECT) imaging will play an essential role in computational analytics for theranostics at multiple time points to evaluate dose to volume and migration. Today, all patients are treated with image guidance to ensure stability and reproducibility in daily positioning. Optical tracking tools are used to ensure three-dimensional stability of patient set up during treatment and imaging is used to validate outcome. These changes in daily work and workflow have largely occurred during the past two decades, therefore physics staff and dosimetrists have become nimble in image acquisition and fusion. Modern planners have become expert in radiographic anatomy and often are responsible for contouring normal tissues including subtle structures such as the optic chiasm and the cochlea. These skills are far different than the skill set required two decades ago, and this evolving expertise is an important aspect to modern planning teams [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11].

10. The role of AI in radiation oncology

As treatment planning becomes more complex, tasks that can be performed by AI will become invaluable to department workflow and an essential tool for planning teams. There are evolving standards in radiation oncology which must include the tools of artificial intelligence to optimize patient care moving forward. Departments will need to become as efficient as possible, even in performing tasks with advanced complexity. There are changing reimbursement models for radiation oncology and it is anticipated that compressed fractionation models will continue to mature. Therefore, even though the number of treatments to patient may decrease, there will be no decrease in the number and complexity of plans created for patients which will place planning teams including dosimetrists at the crossroad between devoting time to optimize a patient plan and efficiency of time applied to the task. We will need AI tools to help support and facilitate plan development. There are tools available for auto-contouring normal tissue and AI tools are in development to support planning including radiosurgery including the plan seen in Figure 2. This will save time and effort for planning teams and permit planners to focus on optimizing plans relative to constraints established for each patient [107].

11. Dosimetry in clinical trials

Clinical trials including the NCTN and industry are the primary vehicles used to validate process improvements in patient care. Quality assurance in clinical trials is important as the process ensures consistency in trial execution and we can trust the outcome of the trial. Physics and dosimetry teams play an important role in the development and execution of clinical trials. Members of the Imaging and Radiation Oncology Core (IROC) will write guidelines and data management strategies into each clinical trial for imaging and radiation oncology. IROC will review clinical trial objects in real time to ensure completeness of the data acquisition process and re-compute patient analytics for target and normal tissue dose in a single platform to harmonize dose to volume for each patient on study. It is only through processes such as this that data can be powered to understand toxicity relative to dose/volume and ensure protocol coverage of tumor targets. Moving forward, large databases housed in a uniform format with planning and outcome imaging will be required to optimize our understanding of tumor contouring and normal tissue metrics. Annotation of these cases can be used for the development of AI programs. Clinical trials remain our primary resource to confirm standards of care and the quality assurance physics and dosimetry teams of IROC serve an important role in the development and standardization of standards. The Cancer Imaging Archive (TCIA) will be an important resource to further enhance our knowledge by associating radiation therapy data with pathomics and radiomics [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11].

12. Conclusions

Radiation therapy will play an important role for oncology care moving forward. With increasing evidence of the success treating patients with oligometastasis and refinement of tools for radiosurgery including protons, theranostics, and particle therapies, radiation oncology will remain an important component for patient care for decades to come in nearly all disease areas using sophisticated treatment strategies. An increasing number of patients with cancer are now treated with curative intent and the plans developed for these patients is becoming increasingly complex with varied normal tissue constraints, especially seen in patients requiring re-treatment. Many disease programs including breast and prostate are being treated with compressed fractionation, therefore there will be a decrease in the number of treatments, yet the number of plans will increase, and the complexity of each plan will increase. Therefore, the skill set of the planning teams will not resemble the skill set required for work a decade ago. Tomorrow’s planning team will be fluent in AI for plan optimization and applied imaging/image fusion for target definition. Planning teams will require skill and knowledge to create sophisticated treatment plans in a timeframe required for timely care. Although an exciting new era for radiation oncology, the skill required for treatment planning and treatment execution is increasing and departments will need to provide educational resources to meet this need moving forward. Training programs will need to adapt to meet these evolving standards. The future of radiation oncology is strong; however, we must continue the process of self-improvement to meet the needs of patients moving forward.

Conflict of interest

The authors declare no conflict of interest.