Clinical outcomes and efficacy of stereotactic body radiation therapy in children, adolescents, and young adults with metastatic solid tumors

Abstract

Objective:

The objective of this study is to report disease outcomes and toxicity with the use of stereotactic body radiation therapy (SBRT) in the treatment of pediatric metastatic disease.

Methods:

All pediatric and adolescent young adult (AYA) patients’ who received SBRT were included between the years 2000 and 2020. Study endpoints included local control (LC), progression-free survival (PFS), overall survival (OS), cumulative incidence (CI) of death or local failure and toxicity. The end points with respect to survival and LC were calculated using the Kaplan–Meier estimate. The cumulative incidence of local failure was calculated using death as a competing risk.

Results:

16 patients with 36 lesions irradiated met inclusion criteria and formed the study cohort. The median OS and PFS for the entire cohort were 17 months and 15.7 months, respectively. The 1 year OS for the entire cohort was 75%. The 6- and 12 month local control was 85 and 78%, respectively. There were no local failures in irradiated lesions for patients who received a BED₁₀≥100 Gy. Patients who were treated with SBRT who had ≤5 metastatic lesions at first recurrence had a superior 1 year OS of 100 vs 50% for those with >5 lesions. One patient (6.3%) experienced a Grade 3 central nervous system toxicity.

Conclusion:

LC was excellent with SBRT delivered to metastatic disease, particularly for lesions receiving a BED₁₀≥100 Gy. High-grade toxicity was rare in our patient population. Patients with five or fewer metastatic sites have a significantly better OS compared to >5 sites.

Advances in knowledge:

This study demonstrates that SBRT is safe and efficacious in the treatment of pediatric oligometastatic disease.

Introduction

Pediatric cancer remains the most common cause of non-accidental death in the United States and is estimated to account for 16,850 cases in 2020 (age 0–19 years). ^1,2 Technical advances in the delivery of radiotherapy and surgery combined with the development of novel chemotherapeutic regimens have improved the 5 year overall survival (OS) of all pediatric malignancies to nearly 85%. ¹ While outcomes remain excellent for leukemias (5 year OS:87%) and lymphomas (5 year OS:94%), patients with certain solid tumors including malignant bone tumors, soft tissue sarcomas, and central nervous system neoplasms have significantly inferior survival rates. Recurrence, whether distant or locally recurrent, is a major cause of morbidity and mortality in pediatric patients and accounts for 67% of all deaths among 5 year survivors. ³ Strategies to address recurrence vary greatly among different subsites of tumors, but typically includes either second-line chemotherapy, surgery for resectable recurrences, or palliative radiotherapy. Palliative radiotherapy remains very effective in the palliation of symptoms in pediatric solid tumors, but may be associated with subablative doses of radiotherapy which may not provide durable local control.

Recently, hypofractionated stereotactic body radiation therapy (SBRT) has emerged as an alternative strategy which can deliver a high dose per fraction to the tumor, and therefore a higher biologically effective dose (BED) than conventional radiotherapy. The impact of an SBRT approach on local control of recurrent or metastatic lesions in pediatric patients has not been well elucidated in the literature. In this study, we performed a retrospective review of pediatric and adolescent and young adults (AYA) patients treated at our institution (Ohio State University/ Nationwide Children’s Hospital) with SBRT for recurrent or metastatic cancers and its impact on local control of metastatic lesions.

Methods and materials

Patient selection

We performed an IRB approved retrospective review of all pediatric cancer patients (n = 298) treated at our institution between January 2000 to January of 2020. From this initial data set, we identified metastatic pediatric patients who were treated with SBRT. Patients were excluded if they had less than 2 months of follow-up or if they received conventionally fractionated radiotherapy to metastases (>5 fractions). After excluding patients who did not receive SBRT (n = 272) and received conventional fractionation to their metastases (n = 10), 16 patients with 36 distinct lesions irradiated with SBRT met eligibility criteria and comprised the overall cohort.

Immobilization

Stereotactic treatment requires rigid immobilization for reproducibility and will vary depending on treatment location while also being mindful of comfort for our pediatric patients. Solitary targets within the brain were set up and treated utilizing a Qfix Encompass SRS thermoplastic mask. Targets within the head and neck, cervical spine, and upper thorax regions utilized a full head and neck thermoplastic mask with a rigid or custom head rest for support with arms at sides using indexed hand pegs.

Lower spine targets and inferior targets within the chest and lung were set up using a CIVCO SBRT body pro-lok setup system including an indexable carbon fiber table platform and arm shuttle with a full body vac-lok bag. Body pro-lok bridges with clam-lok cushions were also utilized to help aid in immobilization as well as an indexed knee sponge for comfort and reproducibility. A respiratory belt or plate was included in patient setups with targeted locations near the diaphragm to assist in restricting respiratory movement. SBRT targeted locations within the pelvis required an indexed lower vac-lok bag immobilization with arms on the patient’s chest holding a ring for comfort. Vac-lok bags were used for immobilizing extremities with careful consideration in regards to machine clearance. Lower extremities were set up feet first while upper extremities were set up akimbo with the patient offset on the table.

Radiation planning

The gross tumor volume (GTV) was defined as the gross disease seen on clinical exam or radiographic imaging (CT/MRI or PET/CT). For metastatic sites in the lung, liver, abdomen or pelvis an internal target volume (ITV) was generated using the 4DCTs to encompass tumor motion. The use of a clinical target volume (CTV) varied by site with no CTV used for the majority of sites, but based on clinical judgment a margin of 3–5 mm was allowed. A 0–5 mm planning target volume (PTV) volumetric expansion was used to account for setup error. For spine metastases, no PTV margin was used, and the GTV and vertebral bodies were contoured based on consensus spine radiosurgery guidelines. ⁴ One target volume was typically used for contouring and planning. Treatment was delivered daily with a cone beam CT supervised by the treating Radiation Oncologist and no breaks were given between fractions.

Treatment planning

Pediatric CT protocols were selected to reduce patient imaging dose with 0.13 cm slice thickness. 4DCTs were acquired for motion management in patients with tumor location in the chest or abdomen. Beam energy was dependent on treatment site utilizing 6 MV, 6FFF, 10 MV, or 10FFF with flattening filter free selected when possible to decrease treatment times with increased dose rate. Most patients were treated with volumetric arc therapy (VMAT) with full arcs for centrally located targets while partial arcs were chosen for unilateral target locations. All patients were treated with photons. Three-dimensional planning was a chosen technique for an extremity utilizing parallel opposed static beams with a dynamic conformal arc. Treatment planning utilized a 0.1 cm grid size for dose calculation.

VMAT goals with inverse optimization allowed for 95% of the PTV to receive prescription dose. A steep dose gradient was achieved sparing surrounding organs at risk. Acceptable hot spots ranged from 110 to 150% of the prescription dose based on the disease site and surrounding normal tissue structures. If multiple targets are present in the same area, a single isocenter multitarget (SIMT) approach was used with care taken to choose appropriate collimator and gantry angles. Forward planning goals included 95% of the PTV to receive prescription dose and dose escalation to the GTV when achievable. The median prescription dose was 25 Gy delivered in a median of 5 fractions. Dose constraints were used from available Children’s Oncology Goup Trials (ARST 1431) and American Association of Physicists in Medicine (AAPM) Task Group 101. Typical normal tissue constraints for central nervous system structures were a maximum point dose of 27 Gy for the spinal cord, 30 Gy for the thecal sac, and 28.8 Gy for the sacral plexus and cauda equina. For lung structures, we aimed to keep 1500 cc of the right and left lung below 11.2 Gy, and the trachea and ipsilateral bronchus to 34.2 Gy maximum point dose. For bone structures, we constrained the femoral heads to a volume of <10 cc receiving 27 Gy.

Clinical data

Variables analyzed include sex, tumor histology, previous RT, age at first cancer diagnosis and SBRT, number of metastatic lesions (>5 or≤5), surgical resection at time of SBRT, and SBRT dose. The follow-up duration for each patient was calculated from the end of SBRT radiation treatment to the date of last follow-up or death from any cause. We also calculated the biologically effective dose (BED) of each SBRT regimen assuming an α/β of 10 for tumor. Toxicity was scored using the Common Terminology Criteria for Adverse Events (CTCAE) v. 5.0.

Statistical analysis

Patient characteristics were summarized using the median (interquartile range [IQR]) for continuous data and frequency (percent) for categorical data. OS was defined as the date of initial SBRT treatment to the date of death and censored at the date of last follow-up for those still alive. Progression-free survival (PFS) was defined as the date of initial SBRT treatment to the date of initial local progression, metastatic disease, death of any cause, or censored at date of last follow-up for those without progression The end points with respect to survival and PFS were calculated using the Kaplan–Meier estimate. Cumulative incidence of local failure was calculated using death without local failure as a competing risk. Survival end points were defined using each individual patient, however, analysis of local failure end points treated each tumor location as independent. 95% confidence intervals are provided for Kaplan–Meier and cumulative incidence estimates. The log-rank test was used to compare survival curves and Gray’s test was used to evaluate difference in cumulative incidence functions. Analyses were performed using R, v. 3.6.3 (R Core Team, R Foundation for Statistical Computing) with the survival and cmprsk packages.

Results

Patient and treatment characteristics

Table 1 summarizes the patient baseline demographic and tumor characteristics of our cohort. The median follow-up for the entire cohort was 12 months. The median age at SBRT was 16.2 years and 75% of the patients were female. The three most common tumor types were Ewing sarcoma (19%), Osteosarcoma (19%), and Rhabdomyosarcoma (12%). The most common tumor locations for SBRT were the spine (31%), lung (22%), and pelvis (19%). Of the 16 patients, 9 (56%) received a course of radiotherapy (RT) prior to SBRT to a median dose of 50.4 Gy in a median of 28 fractions. The median time from the first course of RT to the start of initial SBRT was 32.5 months (IQR: 10, 93.3) for 9 patients who received prior RT. Of the entire cohort, only 6 (37.5%) of patients received systemic therapy concurrently with SBRT. Systemic regimens included Temsirolimus/Cytoxan/Vinrolebine, temozolomide/vincristine, pazopanib, and irinotecan/vincristine. Patient’s with lung metastases received a median BED₁₀ of 100 Gy (IQR: 45–100 Gy) compared to 48 Gy (IQR: 38–56) for those treated for bone tumors.

Table 1.

Patient characteristics

Patient characteristics	N = 161
Age at Initial RT (years)	12.0 (8.0, 15.0)
N without initial RT	7
Age at SBRT (years)	16.2 (12.6, 18.1)
Difference in age (SBRT – Initial RT) (years)	3.5 (1.5, 7.8)
N without initial RT	7
Sex
Female	12 (75%)
Male	4 (25%)
Tumor histology
Ewing sarcoma	3 (19%)
Osteosarcoma	3 (19%)
Rhabdomyosarcoma	2 (12%)
Ganglioglioma	1 (6.2%)
Pleomorphic xanthroastrocytoma	1 (6.2%)
Synovial sarcoma	1 (6.2%)
Renal cell carcinoma	1 (6.2%)
Adenoid cystic carcinoma	1 (6.2%)
Wilm’s tumor	1 (6.2%)
Neuroblastoma	1 (6.2%)
Melanoma	1 (6.2%)
Initial treatment dose per fraction (Gy)	1.8 (1.8, 1.8)
N without initial RT	7
Initial treatment number of fractions	28 (10, 31)
N without initial RT	7
Previous Irradiation Prior to SBRT
Yes	9 (56%)
No	7 (44%)
Pre-irradiation surgery
Yes	15 (94%)
No	1 (6.2%)
Systemic therapy during SBRT
Yes	6 (38%)
No	10 (62%)
Number of metastatic lesions
≤5	8 (50%)
>5	8 (50%)
Median follow-up in months	12 (5, 16)
Lesion characteristics	N = 361
SBRT location
Spine	11 (31%)
Lung	8 (22%)
Pelvis/Hip/Femur	7 (19%)
Brain	5 (14%)
Upper extremity	3 (8.3%)
Chest wall	1 (2.8%)
Liver	1 (2.8%)
SBRT total dose (Gy)	28 (25, 40)
SBRT dose per fraction (Gy)	6 (5, 8)
SBRT number of fractions	5 (5, 5)
BED	48 (38, 72)
BED
≤48	25 (69%)
>48	11 (31%)
BED
<100	30 (83%)
≥100	6 (17%)
1Median (IQR); n (%)

IQR, interquartile range; SBRT, stereotactic body radiation therapy.

Treatment outcomes

The median OS and PFS for the entire cohort were 17 months (CI: 15.7, NA) and 5.6 months (CI: 2.9, NA) respectively. The 1 year OS for the entire cohort was 75% (CI: 54%, 100%) (Figure 1A). At a median follow-up of 5.6 months (IQR:3.4, 10.9), there were 5 local failures out of the 36 lesions which were irradiated (13.9%). The 6 months and 12 month PFS for the entire cohort was 37% (CI:19%, 74%) and 22% (8%,60%), respectively (Figure 1B). The 1 year cumulative incidence of death and local failure was 33% (CI:18%, 49%) and 14% (CI:5%, 27%), respectively (Figure 1C).

Figure 1.

Clinical outcomes of SBRT for recurrent pediatric tumors stratified by (A) OS (B) PFS and (C) cumulative incidence of local failure and death without local failure. BED, biologically effective dose; LC, local control; OS, overall survival; PFS, progression-free survival; SBRT, stereotactic body radiation therapy.

There was no statistical difference in incidence of death or local failure when stratifying by the median BED (p = 0.13 and p = 0.66, respectively; Figure 2A). However, for patients who received a BED₁₀ ≥100 Gy, which correlates to a dose of 10 Gyx5, there were no observed local failures to the irradiated lesions. Patients who were treated with SBRT who had ≤5 metastatic lesions at first recurrence had a superior 1 year OS of 100 vs50% for those with >5 lesions (Figure 2B). A model-based summary of estimated BED values and 95% CI for each histology is shown in Figure 3. The three histologies associated with the highest BED were Adenoid Cystic Carcinoma, Synovial Sarcoma, and Osteosarcoma (Figure 3). The tumor locations associated with the highest BED delivered were liver and lung (Figure 3).

Figure 2.

Clinical outcomes of SBRT for metastatic pediatric tumors as described by (A) lesion-level outcomes for cumulative incidence of local failure and death without local failure stratified by Biologically Effective Dose (BED) <48Gy₁₀ versus ≥48Gy₁₀ and (B) patient-level outcomes for overall survival stratified by ≤5 or >5 metastatic lesions.

Figure 3.

Lesion-level summary for the distribution of Biologically Effective Dose (BED) values by histology and location of lesion.

Toxicity outcomes

No patients experienced a Grade 4 or 5 toxicity in our cohort. One patient (6.3%) experienced a CTCAE Grade 2 acute dermatitis that resolved. One patient (6.3%) experienced a CTCAE Grade 3 toxicity. The patient with Grade 3 toxicity developed a seizure 3 months after fractionated stereotactic radiosurgery (fSRS) for a left parietal lesion to a total dose of 25 Gy in 5 fractions. The patient’s seizures resolved with antiepileptic medications. No patients who received systemic therapy concurrent with SBRT experienced a CTCAE Grade 3 toxicity or higher. Toxicity data can be found in Table 2.

Table 2.

Acute and chronic toxicity outcomes for pediatric patients treated with SBRT

Toxicity type	CTCAE Grade 2	CTCAE Grade 3
Total	1 (6.3%)	1 (6.3%)
Acute pneumonitis	0 (0%)	0 (0%)
Chronic pneumonitis	0 (0%)	0 (0%)
Acute dermatitis	1 (6.3%)	0 (0%)
Radiation necrosis	0 (0%)	0 (0%)
Seizure	0 (0%)	1 (6.3%)

CTCAE, Common Terminology Criteria for Adverse Events; SBRT, stereotactic body radiation therapy.

Discussion

This is one of the growing number of studies to report clinical disease outcomes for SBRT in recurrent and metastatic pediatric solid tumor malignancies. We show that SBRT is efficacious and well tolerated in pediatric patients. In addition, our study sheds additional critical insight on the potential impact of the number of metastatic lesions and BED on local control, which should inform future clinical trials.

Clinical disease outcomes in our study are comparable to those previously published in the literature. The Memorial Sloan Kettering Group recently published their experience of hypofractionated palliative SBRT in 62 pediatric patients with 104 lesions and demonstrated a 1 year LC of 74%, comparable to our 1 year LC of 78%. ² Recently, investigators at the Royal Marsden Hospital in the UK reported on the use of SBRT in 14 patients with 18 lesions and reported a 1 year local control of 78.6%, similar to the local control in our series. ⁵ Casey et al, previously reported on the use of hypofractionated RT in Ewing sarcoma and Rhabdomyosarcoma in patients with 49 bone metastases to a median BED of 42.4 Gy and demonstrated favorable local control with a cumulative incidence failure rate of 6.6% at 1 year. ⁶ Tinkle et al, recently published their experience of 55 patients treated to with SBRT to 107 lesions at a median dose of 35 Gy in 5 fractions and found a similar 1 year estimated cumulative local failure rate of 25.2%. ⁷ In addition, the largest Phase II trial to date using SBRT for pediatric sarcoma bone metastasis demonstrated in 14 patients a 6 month lesion specific local control of 95% and 2 CTCAE Grade 3 toxicities (bone fracture and dysphagia). ^8,9

A direct randomized comparison between SBRT and conventionally fractionated radiotherapy regimens has not been performed and our conclusions are limited to small retrospective studies. Previous clinical outcomes of palliative fractionated radiotherapy in persistent/progressive pediatric malignancies have demonstrated an improvement in pain in 90% of the radiation courses. ¹⁰ In addition, a small retrospective study of palliative conventionally fractionated radiotherapy in 21 patients using 30 Gy in 10 fractions noted an overall response rate of 90% with a median OS of 1 year. ¹¹ Interestingly, the authors noted that nearly 41 days were spent in treatment. In a large combined study from multiple national and international centers, 88 pediatric patients treated to 131 lesions with a mostly conventionally fractionated palliation (92%) demonstrated favorable outcomes with 83% of patients experiencing either a complete response (CR) or partial response (PR). ¹² Furthermore, 73% of distant metastatic sites treated with RT were controlled in a small retrospective study of Stage IV rhabdomyosarcoma demonstrating the impact of local therapy on metastatic sites. ¹³ In summary, clinical outcomes with SBRT compares favorably with conventionally fractionated treatments, with a shorter number of treatments, which significantly improves patient quality of life.

The better than expected OS for patients with ≤5 metastatic sites who were treated with SBRT warrants further investigation. However, these findings are consistent with a survival benefit seen in adult patients with a limited number of metastases (oligometastasis). The SABR-COMET trial in adult patients with oligometastatic cancers (Defined as ≤5 metastatic lesions) demonstrated a median OS of 41 months for patients treated with SABR vs 28 months for those treated with standard of care treatments alone. ^14,15 In a recently reported Phase II randomized trial of SBRT with maintenance chemotherapy in metastatic non-small cell lung cancer (Also defined as ≤5 metastatic lesions) vs chemotherapy alone demonstrated an improvement in PFS in of 9.7vs 3.5 months in favor of the SBRT arm. ¹⁶ In another multicenter trial, Gomez et al reported on 74 patients who were randomized to either local consolidative therapy vs maintenance treatment (≤3 lesions) with Stage IV NSCLC and demonstrated an improvement in median PFS of 11.9 vs 3.9 months. ¹⁷ In addition, the impact of an optimal dose fractionation and BED has not yet been clearly defined. In our study, no significant benefit in LC was observed when stratified by BED₁₀. These findings are consistent with the results reported by Lazarev et al, which stratified BED with a 30 Gy cut point and found no difference in LC. ² However, one intriguing finding in our study was the lack of local failures in patients who received a BED₁₀ ≥100 Gy. In our study, our patients with six tumor locations received this dose and there were no local failures (three lung and one liver). While the small sample size limits our ability to interpret this result, it could be plausible that very high biologically effective doses are needed for durable local control.

The use of SBRT in recurrent or pediatric cancers must be judiciously weighed against the risk of side-effects and its impact on QOL, given the poor OS of these patients due to metastatic progression outside of the irradiated field. Several studies have demonstrated that the majority of these patients have a 2 year OS of 20–30%. ^2,13,18 Therefore, careful consideration of patient symptomology, number of metastatic lesions, and potential toxicity should be considered during consultation. Treatment related toxicity in our study was very low, with only one Grade 3 event, consistent with the toxicity events observed in other SBRT studies. ⁸

Our study has several limitations that must be acknowledged. First, given the inherent retrospective nature of the study, selection bias could impact the survival and cumulative incidence. Next, given the limited patient numbers and limited number of patient events, we were unable to perform a multivariable analysis and elucidate prognostic factors associated with survival in these patients. Our study also included a wide range of dose fractionation schedules and histologies which limits our ability to determine the most optimal radiation dose and fraction size. In addition, the superior 1 year OS for patients with ≤5 metastatic sites is only hypothesis generating and should be cautiously interpreted, given the wide confidence intervals and small patient numbers. The majority of patients in this study did not receive SBRT to all metastatic sites, and therefore an analysis of total consolidation vs partial consolidation was unable to be performed. Finally, quality of life (QOL) was not measured in this study, making it difficult to determine the impact of an SBRT fractionation regimen vs conventional fractionation on QOL measures.

Prospective clinical data will prove to be crucial to the widespread use of SBRT in pediatric patients. COG AEWS 1221 is evaluating the use of an SBRT regimen of 30–40 Gy in 5 fractions for metastatic Ewing sarcoma patients who have bone metastases. In addition, a French SBRT trial will be evaluating the use of SBRT in paraspinal metastases of 27 Gy in 3 fractions or 35 Gy in 5 fractions. ¹⁹ As we await the results of these trials, a collaborative multi-institutional pooled analysis of retrospective data will be critical to more accurately delineate the pediatric patient population that would most benefit from this approach.

Conclusion

This is one of a growing number of studies of an SBRT palliative approach to the treatment of recurrent or metastatic pediatric solid tumors. We have demonstrated that SBRT is both safe and effective alternative to conventionally fractionated palliative radiotherapy. Future prospective trials with multi-institutional collaboration will be necessary to evaluate appropriate patient selection and the optimal radiation dose regimen.