Practice Patterns and Recommendations for Pediatric Image-Guided Radiotherapy: A Children’s Oncology Group Report

Abstract

This report by the Radiation Oncology Discipline of Children’s Oncology Group (COG) describes the practice patterns of pediatric image-guided radiotherapy (IGRT) based on a member survey and provides practice recommendations accordingly. The survey comprised of 11 vignettes asking clinicians about their recommended treatment modalities, IGRT preferences, and frequency of in-room verification. Technical questions asked physicists about imaging protocols, dose reduction, setup correction, and adaptive therapy. In this report, the COG Radiation Oncology Discipline provides an IGRT modality/frequency decision tree and the expert guidelines for the practice of ionizing image guidance in pediatric radiotherapy patients.

Introduction

The delivery of advanced radiotherapeutic techniques is directed by tumor localization and patient alignment confirmation. Intensity-modulated radiation therapy (IMRT), volumetric modulated arc therapy (VMAT), and proton therapy (PT), allwith daily image guidance provide highly conformal treatments with the prospect of treatment margin reduction.^1–3Setup errors may result in undertreatment of target volumes and increased doses to critical organs, thereby compromising disease outcomes and increasingthe risk of complications in normal tissue.^4–6 To mitigate these uncertainties, image-guided radiation therapy (IGRT) verification methods have been implemented.

Asin adults, IGRT is commonlyusedto treat multiple malignanciesin pediatric patients, facilitating target volume localization and normal structure avoidance.^7–12Image guidance with both planar/volumetric imaging and internal/surface markers isused to increase treatment accuracy.⁸The use of bony landmarks and fiducials may be necessary to confirm the alignment within two radiographic planes.^7,13Three-dimensional volumetric image guidance with cone-beam computed tomography (CBCT) or computed tomography (CT)-on-rails can provide position reproducibility for body and internal anatomy to within a millimeter if 6-degrees-of-freedom (6DOF) treatment couch positioning is implemented. Also, with improved soft tissue visualization,interval changesin tissues can be observed, expediting treatment re-planning.^7,13,14

While it is imperative to optimize IGRT techniquesto improve pediatric disease outcomes, it is also important to minimize the cumulative radiationexposure to normal tissues from IGRT, thus reducing late effects.^15–17 Volumetric and frequent imaging prolongthe treatment time, which is unsatisfactoryfor older,unsedatedchildren,who may find it difficult to maintain their position,orfor children who require extended imaging volumes for craniospinal, whole-abdomen, or whole-lung irradiation.Prolongingthe anesthesia times for younger children also entails risks that should be minimized. The additional costs and secondary cancer risk due to the added radiation exposurehave been reported for CBCT and portal imaging.^14,18The Image Gently Alliance for pediatric patients,¹⁹endorsed by the American College of Radiology (ACR), the American Society for Radiologic Technologists (ASRT), and the American Association of Physicists in Medicine (AAPM),recommends using lower radiation doses when imaging children.

The Children’s Oncology Group (COG) has reported on the portal imaging practice patterns of member institutionsand provided recommendations to minimize unnecessary radiation exposure without compromising verification accuracy.¹⁰A consensus on IGRT among pediatric providers remains complex because the evolving technology has resulted in practice preferences, imaging frequency, and verification protocols varying greatly among institutions. The adult IGRT protocols are suboptimal forimaging the distinct pediatric malignancies with their diverse prognostic implications, disease-specific treatments, and risk of unique late toxicities.The COG Radiation Oncology Discipline designed twosurveysto understand clinical and technical practice patterns of pediatric IGRT amongmember radiation oncologistsand affiliated medical physicists. This report summarizes the survey results and providesexpert recommendations on IGRT practice in children.

Materials and Methods

Survey participants

A total of 347members of the COG Radiation Oncology Discipline at national and international institutions were invited to participate in the survey of practice patterns of IGRT in patients aged21 years or younger (entire questionnaire in Supporting File I).Participants wereeligibleif they had a valid email address, were working at the institution,anddid not have a suspended membership.

Survey protocol

The survey was conducted between October 27, 2017 and December 17, 2017by using the COG SurveyMonkey website. After the survey responses were received, the COG radiation oncology discipline created a task force to summarize current data on radiation exposure during IGRT and develop pediatric IGRT guidelines for modern use. These guidelines were then reviewed by the ten disease site committees and the COG before a consensus recommendation was reached.

Clinical survey questions for radiation oncologists

The physicians completed a 39-item clinically oriented survey that included a series of demographic questions and clinical scenarios. Because of the diverse tumor histologies, disease sites, and management strategies, we created11disease-specific vignettes (Supporting File II) with both closed- and open-ended options to elicit the participants’ description of clinical practice most effectively. For each vignette, physicians were asked to identify their preferredtreatment technique [photons-parallel opposed, three-dimensional (3D)conformal radiation therapy (CRT),IMRT/VMAT, TomoTherapy®, PT, orother],preferred method of image guidance [none, megavoltage planar imaging (MVi), kilovoltage planar imaging (kVi), kilovoltage (kV) stereotactic imaging, megavoltage(MV) CBCT, kV CBCT, in-room CT, or other], and frequency of in-room verification [daily, weekly, first week, daily two-dimensional (2D) and weekly 3D imaging, frequently in the beginning and weekly thereafter, orother]. kVi, MVi, and kV stereotactic imaging represent 2D planar imaging modalities which help verify patient position predominately based on bony anatomy and static treatment volumes. kV and MV CBCT, in-room CT, and MV CT are 3D volumetric imaging modalities, offering better visualization of soft tissues and body surface.

Technical survey questions for medical physicists

The physicists received a 23-item technical survey. The technical questions concerned institutional capabilities for treating pediatric malignancies [IMRT, brachytherapy, high-dose-rate (HDR) brachytherapy, stereotactic radiosurgery (SRS), stereotactic body radiotherapy (SBRT), and four-dimensional simulation], the pediatric-specific IGRT image-acquisition protocol, and technical practice patterns. Additional specific questions encompassed PTpractice, setup corrections, measures to facilitate dose reduction, and the use of adaptive therapy.IGRT credentialing has been a requirement for several COG sarcoma trials. Based on data obtained from the Imaging and Radiation Oncology Core (IROC Houston and Rhode Island), among 212 COG centers, 68 sites are credentialed for IGRT using bony landmarks and 115 COG sites that are credentialed for IGRT in soft tissue. For PT, 31 proton centers are approved to enroll patients on National Cancer Institute funded clinical trials. All 31 centers have 2D image guidance capability and 16 centers can perform 3Dimage guidance.

Perspectives on IGRT

Based on Likert rating scales, physicians and physicists were asked to provide their perspectives on the importance of various IGRT-related topics. Concepts included, but were not limited to,the risk of secondary malignancies, setup margins, workflow efficiency, imagingdosereduction strategies, and immobilization techniques. Each respondent rated their agreement with the various IGRT priorities on a scale from 1to 5, with 5 representing the greatest agreement and1 representing the least agreement (Table 3).

Table 3.

Prioritization and understanding of IGRT-related topics by medical doctors(MDs) and medical physicists(MPs)

(1= strongly disagree; 5= strongly agree)

Agreement	1	2	3 Responses N (%)	4	5	Total responses	Average
1. Image guidance improves treatment outcomes.
MPs	2(3)	0(0)	4(7)	12(19)	45(71)	63	4.6
MDs	1 (1)	3(3)	17(16)	27(26)	57(54)	105	4.3
2. Radiation exposure from image guidance poses a non-negligible risk of secondary cancer and,therefore,should be lowered.
MPs	3(5)	12(19)	23(37)	16(25)	9(14)	63	3.3
MDs	9(9)	17(16)	40(38)	28(27)	11(10)	105	3.1
3. It is a high priority to improve image quality (DRR, portal images, CT, CBCT) and/or soft tissue contrast.
MPs	1(2)	1(2)	11(17)	22(35)	28(44)	63	4.2
MDs	0(0)	6(6)	17(16)	40(38)	42(40)	105	4.1
4. It is a high priority to determine appropriate setup margins with IGRT.
MPs	0(0)	3(5)	2(3)	24(38)	34(54)	63	4.4
MDs	1(1)	2(2)	8(8)	36(35)	57(54)	104	4.4
5. Better tools or workflow are needed to reduce the time spent on image review and approval.
MPs	2(3)	8(13)	16(25)	16(25)	21(34)	63	3.7
MDs	5(5)	14(13)	22(21)	39(37)	25(24)	105	3.6
6. It is important to develop non-ionizing image guidance techniques.
MPs	3(5)	6(9)	18(29)	21(33)	15(24)	63	3.6
MDs	2(2)	6(6)	30(29)	34(32)	33(31)	105	3.9
7. It is necessary to estimate age- and organ-specific doses from radiological image guidance procedures and implement imaging dose reduction strategies.
MPs	1(2)	11(17)	18(29)	23(36)	10(16)	63	3.5
MDs	0(0)	6(6)	27(26)	37(36)	33(32)	103	3.9
8. It is important to establish practice guidelines such as optimal imaging frequency, recommended age-specific planar X-ray techniques, and CBCT scan protocols.
MPs	1(2)	1(2)	10(16)	24(38)	27(42)	63	4.2
MDs	3(3)	4(4)	15(14)	45(43)	37(36)	104	4.1
9. It is important with IGRT to improve immobilization and reproducibility.
MPs	1(2)	0(0)	1(2)	14(22)	47(74)	63	4.7
MDs	3(3)	2(2)	9(9)	31(29)	60(57)	105	4.4
10. It is important to facilitate adaptive replanning, e.g.,to improve the CT number accuracy of CBCT for dose escalation and efficient replanning workflow.
MPs	3(5)	1(1)	12(19)	30(48)	17(27)	63	3.9
MDs	0(0)	8(8)	28(27)	34(32)	35(33)	105	3.9
11. There are no pressing needs for pediatric IGRT. Efforts should focus on other tasks.
MPs	26(41)	22(35)	7(11)	6(10)	2(3)	63	2.0
MDs	45(43)	33(32)	16(15)	9(9)	1(1)	104	1.9

Abbreviations: CBCT = cone-beam computed tomography; CT = computed tomography;DRR = digital reconstructed radiograph.

Results

Participant demographics and institutional capabilities

Details are presented in Tables 1 and 2.Of the 347 individuals to whom surveys were sent, 105 physicians and 63 physicists responded, resulting in 168 evaluable responses (a 48% response rate). Most institutions (47%) treated more than 30 children annually. Conventional linacphotons (95%) andelectrons (92%) were the most commonly availabletreatment modalities for managing pediatric malignancies. Only 20% of physicians reported having direct access to PT.

Table 1.

Demographics and institutional capabilities reported by physicians in the Children’s Oncology Group pediatric IGRT survey

Survey item	Responses (%)
Number of physician responses	N = 105
Institution location
United States	86 (81)
Canada	9 (9)
Australia	9 (9)
Middle East	1 (1)
Number of pediatric patientstreated per year
<11	11 (15)
11-20	23 (22)
21-30	19 (18)
>30	47 (45)
What modalities are available to treat pediatric patients?
Conventional linacphotons	100 (95)
Conventional linacelectrons	97 (92)
Stereotactic radiosurgery	93 (89)
Stereotactic body radiotherapy	88 (84)
Brachytherapy	71 (68)
CyberKnife	23 (22)
TomoTherapy	21 (20)
Protons	21 (20)
Combined MR-cobalt or MR-linac	2 (2)
Carbon ions	0 (0)
Do you incorporate fiducials into your practice?
Yes, regularly	11 (11)
Occasionally, if surgical implants such as hardware or clips are not already in place	54 (52)
No	39 (37)

Table 2.

Institutional capabilities reported by physicists inthe Children’s Oncology Group pediatric IGRT survey

Survey item	Responses (%)
Number of medical physicist responses	N = 63
Which institutional modalities are available to treat pediatric patients?
Intensity-modulated radiation therapy	62(98)
Stereotactic body radiotherapy	61 (97)
Four-dimensional simulation	60 (95)
Stereotactic radiosurgery for central nervoussystem tumors	58 (92)
HDR brachytherapy*	55 (92)
Brachytherap^	13 (22)
Image registrationmethods for body tumors
We combine multiple methods.	25 (40)
We rely on bony anatomy.	22 (35)
We rely on soft tissues.	11 (15)
We rely on fiducial markers.	5 (8)
Do physicians discuss with physicists pros and cons of IGRT technique and frequency for specific patients?
Yes	56 (89)
No	7 (11)
Current institutional practice if your institution offers proton therapy††
We rely on 2D IGRT and other techniques and will not/cannot incorporate 3D volumetric imaging.	1 (2)
We rely on 2D IGRT and other techniques but plan to incorporate 3D volumetric imaging in future.	0 (0)
We currently utilize 3D volumetric image guidance for selected pediatric proton therapy patients.	4 (6)
We currently utilize 3D volumetric image guidance for all pediatric proton therapy patients.	1 (2)
Our institution currently does not have proton capabilities.	57 (90)

Abbreviations: IGRT = image guided radiation therapy; HDR =high dose rate.

^*Only 60 responses;

^{^}Only 62 responses.

^††As of8 April 2020, the record of proton facility questionnaire at Imaging and Radiation Oncology Core (IROC) Houston indicates all 31 proton centers which were approved to enroll patients in NCI-funded cooperative group trials have 2D image guidance capability while only 16 of them have 3D on-board imaging.

Preferences and perspectives regardingpediatric IGRT

Both physicians and physicists were queried about areas for future practice improvement. Specific to IGRT, most physicians (54%) and physicists (71%) strongly agreed that image guidance improved treatment outcomes. On the possibility of radiation exposure from image guidance posing a non-negligible risk of secondary cancer,37% physicians and39% physicistseither agreed or strongly agreed. A similar percentage of respondents expressed a neutral position. There was strong agreement among both physicians (55%) and physicists (54%) on the importance of identifying setup margins with IGRT. The priorities determined by respondents to additional topicsrelating to future practice improvement are presented in Table 3.

Clinicalscenario–based questions for radiation oncologists

A collective summary of the responses throughout all disease sites regarding the recommended treatment technique, imageguidance, and in-room verification is presented in Figures 1A–1C.As the survey did not ask what individuals would do IF they had all imaging or treatment modalities, the stated preferencemay be biased by the availability of IGRT technologies and also insurance approval.

Figure 1.

Clinical scenario-based survey outcomes for (A) treatment modality/technique, (B)image guidance modality, and (C) image guidance frequency. The number on top of eachcolumn represents the number of responses. Abbreviations: 3D = three-dimensional; ALL = acute lymphocytic leukemia; AP/PA = anteroposterior/posteroanterior; CBCT = cone-beam computed tomography; CRT = conformal radiotherapy; CT = computed tomography; IMRT/VMAT = intensity-modulated radiotherapy/volumetric modulated arc therapy; kVi = kilovoltage imaging; MVi = megavoltage imaging; NTSTS =non-rhabdomyosarcoma soft tissue sarcoma.

Ependymoma

As shown in Figure 1, the most commonlyreported IGRT method for ependymoma was IMRT/VMAT(52%) or PT (35%, treat locally or referral) guided by daily (69%) in-room verification using kVi (28%) or CBCT(26%) or a combination thereof (23%).

Craniopharyngioma

For craniopharyngioma treatment, many respondentssupported the use of IMRT/VMAT (51%) or PT (38%) guided with daily (61%) kV CBCT (25%) or combined kV CBCT/kVi (32%).Regarding tumorchanges duringRT, 86% respondents agreed with the importance of re-assessment.Magnetic resonance imaging (MRI) was performed weekly by 15% of the respondents, every other week by 3%, first week only by 2%, and midway through courseby 1%.

Germinoma

The most commonly reported focal treatment strategy for central nervous system (CNS) germinoma was IMRT/VMAT (59%) or PT (29%) with daily (68%) kV CBCT (30%), kVi (25%), or combined kV CBCT/kVi (23%).

Medulloblastoma

The most frequentlyreported techniquefor managing pediatric medulloblastoma with craniospinal irradiation was PT (52%) or, less commonly, IMRT/VMAT (20%) guided with daily (74%) kVi (33%) or combined kV CBCT/kVi (24%).

Rhabdomyosarcoma

IMRT/VMAT (52%) and PT (37%) were the most reportedmanagement modalities for the rhabdomyosarcoma vignette, with 71% of respondents suggesting daily imaging using kVCBCT (35%) or combination (31%).

Non-rhabdomyosarcoma soft tissue sarcoma

The favored approaches to managingnon-rhabdomyosarcoma soft tissue sarcoma (NRSTS) were IMRT/VMAT (45%), 3D CRT (24%), and PT (21%) guided by daily (57%) kV CBCT (34%), combined CBCT/kVi (28%), and kVi (24%).

Ewing sarcoma

For treating Ewing sarcoma, many providers recommended IMRT/VMAT (57%) or PT (24%) guided with daily (66%) kV CBCT (34%), combined CBCT/kVi (34%), or kVi (22%).

Acute lymphoblastic leukemia

For patients with CNS-3 leukemia receiving whole brain RT, most respondents reported the use of photon therapy with parallel-opposed fields (61%) or 3DCRT (29%) guided with daily (45%) or weekly (32%) kVi (50%) or MVi (27%). No image guidance (i.e., using skin/mask markings) was reported by 7%.

Hodgkin lymphoma

Clinicianpreferences for the Hodgkin lymphoma case included IMRT/VMAT (32%), 3DCRT (28%), and PT (22%) guided by daily (58%) kVi (33%), CBCT (24%), or combined CBCT/kVi (19%).

Wilms tumor

Preferred methods for treating flank or whole abdomen in Wilms tumor patients involved Anteroposterior/ posteroanterior photons (66%) guided by daily (45%) or weekly (35%) kVi (55%).

Neuroblastoma

The favored management for high-risk neuroblastoma included IMRT/VMAT (61%) guided by daily (65%) kV CBCT (30%) or combined kV CBCT/kVi (30%).

Technical survey for medical physicists

Institutional capabilities and image registration

The participating physicistsreported institutional capabilities specifically for treating pediatric malignancies. As detailed in Table 2, 98% respondents reported IMRT capability.Forimage registration for body tumors, 35% relied on bony anatomy and 40% combined multiple methodsor were case, disease-site, modality, or physiciandependent.

Setup corrections

For intracranial tumors in a non-SRS setting, 53% of responding physicists reported correcting patient setup regardless of how small the calculated shifts were, whereas others used tolerances of 1mm (19%), 2mm (16%), 3mm (2%), or>3mm (3%). Re-imaging to confirm patient setupafter correction was completed only for shifts that exceeded a certain amount by 34% of respondents: 19% did not re-image at all, 18% re-imaged only after SRS/SBRT or suspected patient movement, 15% always re-imaged, and 15% had otherresponses. Theuse of 6DOF couches and both translational and rotational setup corrections withpediatric patients variedgreatly between institutions. The percentage of physicistswith 6DOF couches that they used for none, 0-50%, 50%-75%, and 75%-100% of theirpediatric patients was 34%, 32%, 12%, and 23%, respectively.

Pediatric-specific IGRT image-acquisition protocols and dose reduction

Most (85%) agreed that using IGRT changedtheir setup margin or the clinical target volume (CTV)robustness parameter setting. Regardingmodifications of the manufacturer’s default adult IGRT protocols for pediatric patients, most physicists (54%) elected to reduce the kilovoltage peak (kVp)/milliampere second (mAs)setting. However, those who did not reduce the setting stated that there was a lack of guidelines (39%), did not consider there was a need to reduce the setting (3%), orwere not allowed to modify vendor protocols (3%).Mostphysicists (60%)recommended incorporating site-specific pediatric imaging protocols into routine practice.Most institutions neither routinely documentedthe imaging dose (89%) nor subtractedthe image guidance dose from the prescribed dose (95%), with 64% consideringthe imaging dose to beinsignificant when compared to the treatment dose,with no evidence that itposeda high risk to patients. Of those physicists who did not subtract the imaging dose from the prescribed dose or document the dose, 62% answered that it was not possible to incorporate the dose accurately, whereas 25% feltthat it was possible and 13% responded “N/A” (i.e., accounting for the imaging dose was not necessary). As listed in Table 4, the most common method to measure image guidance doses was using an ion chamber with a CT dose index phantom (67%). The most common institutional efforts to reduce image guidance dose were lowering the mAs/kVp or using low-dose protocols from vendors (65%) and using kV imaging in preference to MV techniques (63%) (Table 4).

Table 4.

Institutional measures to calculate/estimate and reduce the image guidance dose to pediatric patients

Survey item	Responses (%)
Methods/tools to calculate or measure image guidance dose
Ion chamber with CTDI phantom	38(67)
Ion chamber/TLD/MOSFET/OSLD with anthropomorphic phantom	6 (10)
TLD/MOSFET on patient	4 (7)
Treatment planning modeling of imaging beam	0 (0)
Monte Carlo simulation	0 (0)
Not calculated or measured	9 (16)
Methods to reduce image guidance dose
Lower mAs/kVp or use low-dose protocols from vendors	41 (65)
Use of kV instead of MV	40 (63)
Use of collimation to reduce scan range	30 (48)
Imaging less frequently	24 (38)
Utilize age- and size-specific protocols	19 (30)
Upgrade image guidance software and hardware	13 (21)
Supplement with non-ionizing techniques	9 (14)
Add filtration or shielding	1 (2)
None	5 (8)

Abbreviations: CTDI = computed tomography dose index; kVp = kilovoltage peak; mAs = milliampere second; MOSFET = metal-oxide semiconductor field-effect transistor; MV = megavoltage; OSLD = optically stimulated luminescence detector; TLD =thermoluminescent dosimeter.

Adaptive planning

Most physicists reported a routine practice of adaptive planning during the treatment course to address tumor or anatomic changes (64%). For nine physicists in institutions offering PT, six respondents (67%) routinely performed adaptive PT, whereas three (33%) did not. Of the 62 of 63 physicists who responded to this question, 79% represented institutions that used setup verification images from image guidance procedures to trigger or make decisions regarding adaptive planning.

Discussion

Key observations

It is well recognized that image-guided techniques increase accuracy of patient positioning which enables normal tissue sparing by reducing the safety margin around tumor. As a result, this increases the confidence and utilization of treatment technologies that produce sharp dose gradients, including IMRT/VMAT and PT. The COG survey results indicate that pediatric radiation oncologists have embraced frequent image guidance by incorporating it into their clinical practice. However, treatment techniques and image guidance modality varied greatly among clinicians. Pediatric-specific IGRT image acquisition protocols and dose reduction methods remain underdeveloped. Physicists cited the lack of guidance as the main reason why adult IGRT protocols continue to be applied to children.

Frequency of image guidance

The survey found that daily imaging guided by kVi, CBCT, or combined kVi/CBCT in-room verification was the predominant strategy for patient-specific diseasesfor approximately 70% of respondents except for whole brain and whole abdomen radiotherapy.Alternatively, parallel-opposed photon beams for leukemiaandanterior/posteriorphoton beams for Wilms tumor were favored. Both daily and weekly imaging guided by MVi or kVi were more commonly preferred for these two diseases.

The survey showsthat 70% of physiciansconducted daily imaging with either planar (e.g., kVi) or volumetric (e.g., CBCT) imaging for children with brain tumorsor rhabdomyosarcomasto ensure appropriate setup. The prevalence was slightly lower (57%-66%) for NRSTS, Ewing sarcoma, neuroblastoma, and Hodgkin lymphoma.The International Pediatric Research Consortium reported the use of daily image guidance in approximately 60% of institutions for children with CNS, abdomen/pelvis, or head and neck cancers.⁷ Seventy-four percent of participants in theInternational Paediatric Radiation Oncology Society survey stated that the sameimaging frequency was used for both adults and childrenwith CNS tumors.¹²The national survey of American Society for Radiation Oncology (ASTRO) members reported daily imageguidance rates of 62%-96%.²⁰ However, daily image guidance for brain tumors in adultswas much lower at 18%. Differences in the imaging techniques reported in other series may be secondary tothe patient population, treatment modalities, and available resources.

2D versus 3D image guidance

The COG survey showed that up to 38% and 42% of the clinicians would use solely 3D image guidance for focal radiotherapy at brain and body sites, respectively. In contrast, the 2016 ASTRO survey²⁰found that 60% and 66%-77% of respondents used 3D IGRT for adult brain and body tumor sites, respectively. Differences between adult and pediatric practices might reflect the concernsregardingthe imaging dose and increased anesthesia/treatment time.⁷Although bony landmarks visible on kVi are generally believed to serve as reliable surrogatesfor localizing intracranial targets, CBCT enables planning marginreduction.^7,13,21For craniopharyngioma with both solid and cystic components, MRI can detect the dynamic cystic change during the radiotherapy course. Periodic,²² bi-weekly,²³or weekly²⁴MRI to assess the necessity of adaptive therapyhas been recommended in craniopharyngioma treatment.

Organ doses from IGRT procedures in pediatric patients

The IGRT literature describesthe use of a wide range ofdoses depending on the imaging modality, beam quality, imaging technique, and patient size. Organ dose estimates for CBCT have been tabulated by Alaei and Spezi,²⁵ and similar tables for other imaging modalities were included in the AAPM TG-75and TG-180 reports.^18,26However, data arescarce on organ doses resulting from IGRT procedures in pediatric patients.²⁷Table 5provides sample organ doses for pediatric patients imaged with various CBCT protocols using Elekta kV X-ray volume imaging (XVI). The organ doses werecalculated using a treatment planning system with an imaging beam model.^28,29

Table 5.

Examples ofpediatric organ doses calculated from Elekta kV XVI CBCT scans^{28,29^}

Age Group (years)	Organ Dose Per Scan (mGy)
Low-dose/head and neck protocol—S20 Cassette, 100 kVp, 0.1 mAs/frame, 205 degrees rotation, 366 frames
	Bladder	Rectum	Bowel	R kidney	Lkidney	Liver	Stomach	Spleen	Heart	R lung	Llung	Esophagus	Gonads
2-5	1.1	0.8	1.2	0.7	1.1	0.9	1.3	1.3	1.2	0.9	1.1	0.9	0.9
6-10	1.0	0.7	0.9	0.6	0.8	0.8	1.1	1.1	1.1	0.8	1.1	*	1.0
11-15	0.7	0.5	*	0.5	0.9	0.7	1.1	1.1	1.0	0.7	0.9	0.8	*
	Brain	Brainstem	Chiasm	Roptic nerve	LOptic Nerve	Rcochlea	Lcochlea	R Eye	L Eye	R Lens	L Lens	Pituitary	Thyroid
2-5	0.9	0.9	1.05	1.1	1.3	0.8	1.3	1.25	1.6	1.35	1.7	1.1	1.2
6-10	1.0	1.0	1.1	1.0	1.3	0.8	1.3	1.2	1.6	1.6	1.6	1.0	1.0
11-15	*	*	*	*	*	*	*	*	*	*	*	*	*
Medium dose/thorax protocol—L20 Cassette, 120 kVp, 0.25 mAs/frame, 360 degrees rotation, 660 frames
	Bladder	Rectum	Bowel	Rkidney	Lkidney	Liver	Stomach	Spleen	Heart	Rlung	Llung	Esophagus	Thyroid	Gonads
2-5	5.5	4.6	5.2	5.2	4.9	5.0	5.0	5.1	5.5	4.8	4.7	5.1	5.8	4.9
6-10	5.1	4.5	4.3	4.1	4.4	4.7	4.3	4.6	4.8	4.6	4.4	*	5.3	4.4
11-15	3.3	3.4	*	4.2	4.1	4.1	4.6	4.3	4.5	4.0	4.0	5.0	4.3
High dose/pelvis protocol—M20 Cassette, 120 kVp, 1.0 mAs/frame, 360 degrees rotation, 660 frames
	Bladder	Rectum	Bowel	Rkidney	Lkidney	Liver	Stomach	Spleen	Heart	Rlung	Llung	Esophagus	Thyroid	Gonads
2-5	29.8	25.9	31.9	30.1	30.5	29.9	29.2	32.9	31.2	29.3	29.4	25.9	34.2	27.2
6-10	26.8	24.5	24.7	23.4	24.2	25.9	25.3	24.6	27.8	26.1	25.7	33.1	33.4	23.6
11-15	17.4	16.8	*	22.6	22.0	22.1	25.3	23.9	23.3	23.0	21.6	24.8	20.4

Conditions: CBCT scans were performed with the isocenter placed approximately at the patient midline. All organs listed were fully within the imaged volume. Effects of thetreatment couch were ignored. Doses reported are average organ doses computed using a treatment planning system. The dose to bone was underestimated by this method; hence, bony structure doses were not included. Number of patients analyzed for calculating average values: 7 in the age 2-5 group, 5 in the 6-10 group, and 4 in the 11-15 group.

Abbreviations:CBCT = cone-beam computed tomography; L = left; R = right.

^{^}Organ doses were calculated based on the methods described in these publications.

^*No data available.

Often presumed insignificant, portal imaging can also contribute an excessive radiation dose to pediatric patients, especially with double-exposure techniques, ranging up to 0.75Gy to 2.5Gy over a treatment course forMV port films.^30,31 More recently, Deng et al.³² provided Monte Carlo-calculated organ doses from kVi/CBCT and concluded that critical structures in pediatric patients receive imaging doses two to three times greater than those in adults. For a given kV CBCT protocol, organ dosesincrease with decreasing patient size and body massindex, making it important to considersize-specific protocols.^29,33 As a result, using adult imaging protocols for pediatric imaging will ultimately increase the radiation dose burden for children.^32,34,35

Risks of secondarymalignanciesfrom IGRT procedures

Accurately predicting the adverse effects and risk of death in individuals exposed to radiation doses of less than100 mSv remains challenging. The Biological Effects of Ionizing Radiation (BEIR) VII Committee reported on the risks from medical imaging radiation, acknowledging the limitations and uncertainties of risk estimates.³⁶Zhouand colleaguesrecently reportedthe cumulative imaging doses from IGRT and the associated secondary cancer risk of 4832 cancer patients.³⁷Based on BEIR VII models, the associated average lifetime attributable risks of cancer incidence per 100,000 persons wereestimated to be 78, 271, and 510 for brain cancer, lung cancer, and leukemiapatients, respectively. However, we advise caution when calculatingthe secondary cancer risk from low-dose exposures in patients with cancer receiving radiotherapy by using models derived from atomic-bomb survivors. The study by Littledetermined that, in many organs, the risk per Gy was substantially lower for therapeutic irradiation than for non-therapeutic exposures.³⁸In any case, it is prudent to limit the imaging dose to what is minimally needed to provide the information required.

Despite the risks associated with imageguidance, itspotential advantages must be acknowledged.^9,39The significantlyreducedsetup margin will decrease the dose not only to adjacent healthy tissuesnear the target that are exposed to higher doses of radiation but also to those tissues distal from the target that are exposed to lower doses, thereby diminishing the risk of secondary cancers. As a result of the use of smaller margins and better positioning with IGRT, higher therapeutic doses are more frequently delivered with modern advanced radiotherapy techniques such as SRS, SBRT, and VMAT. The benefits of being able to make informed decisions about margins, adapt the target volumeduring treatment, and ensureaccurate treatment delivery outweigh the risk of secondary cancer that results from diagnostic imaging or other low-dose exposures.

Strategies for reducing the image guidance dose

Methods to reduce the imaging dose to patients include using separate techniques for smaller patients;reducing the imaging field size by either closing the blades or using cassettes instead of open fields;using an appropriate kVp for the imaged site;using kVin preference to MV;reducing thetotal mAs;using a decreased angular range or projection number for CBCT acquisition; and avoiding repeated imaging withan improved setup device/procedure. These strategies can be employed if they do not compromise image interpretation.For example,administering an adult head-and-neckrather than abdomen/pelvis imaging protocol to imagea pediatric abdomen/pelvis could decrease the imaging dose by a factor of 18.⁴⁰When only bony anatomy visualization is needed for CBCT, reducing the kV from 100 to 80 and the mAs by a factor of 3will reduce the dose by a factor of approximately 8 fromthat with the standard linacimaging protocol, yet the resulting images will be no less accurate in 3D matching with the planning CT images.⁴¹Iterative CBCT reconstruction is now commercially available, providing the opportunity to further reduce the imaging dose without sacrificing image quality.⁴²

Surface imaging does not employ ionizing radiation.Thismethod may help reduce the need for repeated imaging by guiding the initial setup for IGRT and monitoring patient motion during treatment.MR-linac or MR-cobalt systems are recent inventions that localize and track tumors without relying on ionizing radiation.⁴³Although our surveyed respondents reported very limiteduse of surface imaging and MRguidance for positioning, we encourage publications on the accuracy and practical implementation of this technology forpediatric patients.

Weaknesses and limitations of the survey

One main limitation of this study is that not all members responded to the survey.Because the data were anonymized, interpretationsof the datawere limitedby an inability to relate the responses to large versus small centers and the available resources.It was also not possible to determine why non-respondents did not reply, and this could lead to biased estimates of IGRT applicationFurthermore, a bias may exist for those individuals more likely to respond to the survey, depending on their prior experience with IGRT or their current institutional capabilities and limitations guided by insurance authorization.As the survey did not ask what individuals would do IF they had all imaging modalities, the stated preference for imaging modality (2D or 3D) and treatment modality (IMRT or proton therapy) may be biased by the availability ofIGRT technologies and also insurance approval. Finally, the use of internal, national, and international guidelines and the demands of trials can also bias practice and this was not fully explored.

Recommendations

Based on the survey resultsand expert consensus, the COG Radiation Oncology Discipline recommends using the IGRT modality/frequency decision tree (Figure 2) and the Choose Wisely recommendations (Table 6) for usingionizing image guidance in pediatricradiotherapy patients. These recommendations are meant to optimize the benefits of IGRT to accurately treat the tumor while minimizing the long-term risks of normal tissue radiation exposure from image guidance modalities.

Figure 2.

Recommended image guidance decision tree for pediatric IGRT. Abbreviations: 3D = three-dimensional; CRT = conformal radiotherapy; IMRT/VMAT = intensity-modulated radiotherapy/volumetric modulated arc therapy; kVi = kilovoltage imaging; MVi = megavoltage imaging; PT = proton therapy; PTV = planning target volume.

Table 6.

Choose Wisely recommendations for pediatric IGRT practice

The following recommendations for wise selection of pediatric IGRT are based on the community practice revealed by the COG survey results, existing evidence, and COG member consensus.
Image guidance modality
• Guiding 2D treatments with 2D kV imaging is generally sufficient without 3D imaging, and normally give a lower imaging dose.These treatments may include whole-brain irradiation for acute lymphocytic/lymphoblastic leukemia, nodal irradiation fields for lymphoma, or flank/whole-abdomen RT for Wilms tumor.
• 3D imaging is recommended when bony landmarks are not reliable surrogates for tumor positions, when margins are small, or when rotational corrections are needed without the guidance of implanted fiducials.
• Consider 3D imaging to reducemargins before prioritizing 2D imaging to reduceimaging dose.
• Do not use MV imaging for more than verifying the field shape on the first fraction unless the low-dose setting is adopted.Consider an alternative method of using the light field projection on field shape diagram in advance.
• Be cautious about electron therapy and light field verification without image guidance for superficial tumors such as chestwall sarcoma.The majority of pediatric radiation oncologists favor conformal treatment with image guidance.
Imaging frequency
• Donot rely solely on weekly imaging at the start of 3D CRT, including CSI beam placement.Such practice is uncommon. Consider reducing imaging frequency to weekly only after daily imaging has confirmed stable anatomy.
• Do not reduce the imaging frequency solely in an effort to reduce the imaging dose. The benefits of accurate tumor targeting with reduced margins may outweigh the risk fromthe imaging dose.
• Minimize repeated imaging in a session to adjust the patient position. Improve patient setup procedures and immobilization devices to minimize multiple exposures.
Imaging dose reduction
• When both MV and kV imaging are available on the same treatment delivery system, choose kV to reduce imaging dose to patients.
• Use field-limiting devices (e.g., blades, collimators, cassettes) to block radiation-sensitive organs (e.g., lens, thyroid, gonads) if target verification is not compromised.
• When volumetric image guidance is preferred in situationswhere only bony anatomy is used for registration (e.g., for rotational correction), utilize institutional 3D low-dose image acquisition techniques.Superior guidancecan still be provided without exposing patients to a significantly higher dose than that with 2D X-rays.
• Do not directly apply imaging guidance techniques designed for adults to young children without modifications.If it is not possible to modify technique parameters such as mAs, consider using the vendor’s low-dose techniques.
• Consider usingnon-ionizing position verification methods (e.g., surface imaging or MRI guidance) to replace or supplement ionizing radiation methods whenever possible.

Abbreviations: 2D = two-dimensional; 3D = three-dimensional; CRT = conformal radiotherapy; CBCT = cone-beam computed tomography; CNS = central nervous system; COG = Children’s Oncology Group; CSI = craniospinal irradiation; IGRT =image-guided radiotherapy; kV = kilovoltage; mAs = milliampere second; MRI = magnetic resonance imaging; MV = megavoltage; RO =radiation oncologist; RT = radiotherapy.

Conclusions

The COG survey shows thatdaily image guidance was used approximately 60%-70% of the time for most disease sites. Although disease specific, kVi was most commonly used for simple treatments, CBCT was more frequently used for complex treatments. We present recommendations to optimize the benefits of IGRT while minimizing the long-term risks of normal tissue radiation exposure. Further research is required to establish the risks from imaging doses, to provide guidance on pediatric imaging techniques, to develop non-ionizing image guidance approaches, to reduce setup margins, to optimize positioning of pediatric patients, andto conduct cost-benefit analyses.

ABBREVIATIONS KEY

2D

two-dimensional

3D

three-dimensional

AAPM

American Association of Physicists in Medicine

ACR

American College of Radiology

ALL

acute lymphocytic leukemia

AP/PA

anteroposterior/posteroanterior

ASRT

American Society for Radiologic Technologists

ASTRO

American Society for Radiation Oncology

BEIR

Biological Effects of Ionizing Radiation

CBCT

cone-beam computed tomography

CNS

central nervous system

COG

Children’s Oncology Group

CRT

conformal radiation therapy

CT

computed tomography

CTDI

computed tomography dose index

CTV

clinical target volume

DOF

degrees of freedom

IGRT

image guided radiation therapy

IMRT

intensity-modulated radiation therapy

IROC

Imaging and Radiation Oncology Core

HDR

high-dose-rate

kV

kilovoltage

kVi

kilovoltage planar imaging

kVp

kilovoltage peak

mAs

milliampere second

MOSFET

metal-oxide semiconductor field-effect transistor

MRI

magnetic resonance imaging

MV

megavoltage

MVCT

megavoltage computed tomography

MVi

megavoltage planar imaging

NRSTS

non-rhabdomyosarcoma soft tissue sarcoma

OSLD

optically stimulated luminescence detector

PT

proton therapy

SBRT

stereotactic body radiotherapy

SRS

stereotactic radiosurgery

TLD

thermoluminescent dosimeter

VMAT

volumetric modulated arc therapy

XVI

X-ray volume imaging