Comparison of outcomes of stereotactic body radiation therapy delivered with three different technologies to the lung

Abstract

Purpose/objective(s)

Since the inception of stereotactic body radiation therapy (SBRT), treatment delivery has been performed with volumetric modulated arc therapy (VMAT), helical tomotherapy (HT) and noncoplanar static fields (SF). The purpose of this study is to compare SBRT delivery among these treatment modalities to the lung.

Materials/methods

A retrospective review of SBRT treatments of 30 to 60 Gy in 1 to 5 fractions from 2007 to 2015 was performed. Dosimetric parameters included V5, V20, D2cm, gross tumor volume (GTV) and planning target volume (PTV) size and coverage, rib/esophageal minimum/maximum doses, R30Gy, R50%, and the conformality index (CI). Clinical outcomes evaluated included local control, pneumonitis and other toxicities. ANOVA, Student’s t-test and Kruskal-Wallis test were used to compare the parameters among modalities. Kaplan-Meier estimates of time-to-local failure were produced.

Results

176 Treatments included 106 SF, 36 VMAT and 34 HT. HT had better PTV coverage (p=0.0166) but higher lung V5 and esophageal doses (p<0.001 and p=0.0032). R30Gy, R50%, and CI were significantly better with VMAT SBRT (p<0.001). Clinically, Grade 2+ pneumonitis was associated with larger median GTV’s of 21.39 cc versus 7.65 cc (p=0.0016), larger median PTV’s of 65.62 cc versus 31.75 cc (p=0.0030), and higher V20 6.62% versus 4.08% (p=0.0408). For patients surviving >1 year, overall local failure rate was 9.4%. Actuarial control rates trended toward statistical significance with time to local failure with VMAT being the most favorable group on the Kaplan-Meier curve (p=0.0733).

Conclusion

VMAT showed superior conformality compared to the other modalities. Among the modalities examined, HT had higher values for parameters associated with toxicity such as V5 and maximum esophageal dose, but all were within acceptable limits. There was a trend to better local control with VMAT.

Introduction

Approximately one fourth of all cases of non-small cell lung cancer (NSCLC) diagnosed are stages IA-IIA [1]. According to the National Comprehensive Cancer Network guidelines, extensive surgical procedures are still considered the standard of care in early stage NSCLC for medically healthy patients despite the growing literature on alternative treatments such as SBRT[2]. Many patients with NSCLC have multiple comorbidities that preclude surgery. In this population stereotactic body radiation therapy (SBRT) has become standard of care. By definition, SBRT is hypofractionated (>600cGy/fraction, 1-5 fractions), highly conformal, targeted radiation therapy which is very demanding in terms of quality control from patient simulation to treatment delivery to gain the optimal results. Over the last decade, the introduction of new technologies allow more rapid delivery of dose along with high precision of targeting of the tumor volumes.

The power of SBRT is the ability to deliver high doses of radiation to the localized tumor mass in a highly conformal manner that minimizes damage to normal tissues. Strategies for optimal tumor control must consider microscopic disease extension around the visible mass and tumor movement, primarily due to respiration. Dosimetrically, quality control for SBRT plans includes several indices that help evaluate the conformality of the radiation delivery. Together, they represent how well the radiation conforms to the target and falls off spatially, sparing lung and other normal tissues. R50% is the ratio of the 50% isodose volume to the PTV volume and R30 Gy is the ratio of the 30 Gy isodose volume to the PTV volume. The ratio of prescription isodose volume to the planning target volume (PTV) is known as the conformality index (CI), and the maximum dose at any point ≥2cm from the PTV in any direction is the D2cm, both useful parameters to further refine quality control. The quality control becomes important in terms of dosimetric plan properties and clinical endpoints. There are a few key dosimetric parameters, such as the volume of lung treated to 20 Gy or higher (V20) and the volume of lung treated to a dose of 5 Gy or higher (V5), that have been shown to be clinically relevant for the prediction of Grade 2+ radiation pneumonitis[3,4].

Known adverse effects of SBRT include fatigue, chest pain, pneumonitis and rib fractures. While these are relatively rare toxicities and are well-documented, it is uncertain what effect the type of delivery modality can have on such clinical endpoints. Over the past several years, new treatment platforms have been introduced into SBRT delivery. Each linear accelerator platform, e.g. Helical Tomotherapy (HT) (Accuray, Sunnyvale, CA), volumetric modulated arc therapy (VMAT) or static fields (SF), expresses differing dosimetric characteristics arising from their underlying physics and treatment geometry. It is due to these differing characteristics that each modality should be evaluated separately and compared since outcomes in terms of local control and toxicity could vary. This study compared these three treatment platforms with respect to measures of conformality as part of the quality assurance process and long-term outcomes, especially local control and toxicity.

Materials/Methods

Patients treated from March 2007 to November 2015 were retrospectively reviewed on this study which was approved by the local Institutional Review Board. Staging according to the 7^th edition American Joint Committee on Cancer was done. Criteria that needed to be met included patients with confirmed NSCLC histology or no biopsy, T1a-T2a tumors, N0 tumors, M0 tumors, treatment with SBRT for the tumor and no other history of malignancy in the known past or future.

Treatment details

All patients were simulated supine in the Stereotactic Body Frame (Elekta, Stockholm, Sweden) or CIVCO immobilization devices (CIVCO Medical Solutions, Kalona, IA). All patients had custom vac-lock bags molded to their bodies in order to improve the reproducibility of the treatment positions. Most patients were treated using abdominal compression with the goal of reducing diaphragmatic motion to less than or equal to 1.0 cm. Computed tomography (CT) simulation was done with 2.5 mm slices from the inferior edge of the cricoid cartilage to the inferior edge of the liver. The GTV was contoured on lung windows and included only solid mass without ground-glass changes. A PTV of 1.0 cm superior and inferior and 0.5 cm radially was applied. In patients who could not tolerate abdominal compression due to prior surgery or other comorbidities, a four-dimensional CT simulation was performed to create an iterative tumor volume (ITV) from a maximum intensity projection. For these patients, the PTV was created by placing a 0.5 cm expansion around the ITV per RTOG 0813 [5].

All dosimetry was performed with the goal of covering at least 95% of the PTV with the 80% isodose line ensuring that all hot spots were within the GTV. Dose constraints for dose limiting organs were taken from multiple published trials including RTOG 0618, 0915, 0813 and ACOSOG Z4099/RTOG 1021.

All SF and HT treatments were calculated using a superposition/convolution algorithm on the CMS XiO (Elekta, Stockholm, Swedem) treatment planning system (TPS) and Tomotherapy TPS (Accuray, Sunnyvale, CA), respectively. SF plans calculated in XiO were universally non-coplanar, utilizing from eight fields to 26 fields (including in-fields). HT deliveries were, by definition, delivered using helical arcs. VMAT doses were calculated using the Boltzmann transport-based Acuros algorithm. (Eclipse Planning System, Varian, Palo Alto, CA). VMAT deliveries contain a variety of cases both coplanar and non-coplanar arcs, depending on the location of the treatment volume, and were accomplished using two to seven arcs. In all cases, heterogeneity corrections were used.

The earliest patients were all treated via SF, as it was the only option available from 2007-2009. For plans delivered from 2009-2014, the only arc delivery method available was HT. Starting February 2014, VMAT was introduced and became the most utilized modality option due to its faster treatment time, replacing the SF modality completely.

Statistical Analysis

Characteristics of the sample were summarized using mean and/or median as well as percentages. One-way ANOVA or Kruskal-Wallis tests were used to assess whether central tendency in distributions amongst groups were significantly different or not. Fisher’s exact test was used to assess whether proportions were different. Kaplan-Meier estimates of time-to-local failure were produced and significance of distributions across groups were assessed using the log-rank test. A two-sided level of significance of 5% was used to address statistical significance of testing. SAS V9.4 (SAS Institude, Cary, NC) was used to facilitate data management as well as data analyses.

Results

A total of 176 patients were eligible for analysis, including 106 who received SF SBRT, 34 who received HT SBRT and 36 who received VMAT SBRT. Mean follow-up times for SF, HT and VMAT SBRT were not significantly different at 21.3, 12.9 and 16.8 months respectively. There were no statistically significant differences between the SBRT modality groups with respect to age, stage, tumor stage, smoking status or tumor pathology. Significantly more males than females received HT SBRT (p=0.0051) (Table 1).

Table 1.

Patient demographic and tumor specific data by SBRT modality

	Static Fields Mean (stderror)	Helical Tomotherapy Mean (stderror)	VMAT Mean (stderror)	p-value
Age	70.6 (0.89)	72.3 (1.57)	70.8 (1.51)	0.6401
Sex (% male)	50	79	68	0.0051
T stage (% stage 2)	26	26	27	0.9973
Location of the tumor (%)
LLL	10.4	14.7	13.5	0.4113
LUL	31.1	26.5	35.1
RLL	20.8	11.8	18.9
RML	9.4	2.9	13.5
RUL	28.3	44.1	18.9
Smoking Status (% current)	36	48	38	0.4453
Tumor pathology
Adenoca	38	35	57
NSCLC	27	23	3
SCC	35	42	40	0.0753

Dosimetric comparison was made among SF SBRT, HT SBRT and VMAT SBRT (Table 2). There were no statistically significant differences between the SBRT modality groups with respect to total dose, number of fractions, gross tumor volume (GTV), GTV minimum dose, PTV size, PTV minimum dose, V20, maximum rib dose and D2cm. Mean biologically effective dose (BED) was 133.5 for SF SBRT, 126.9 for HT SBRT and 129.4 for VMAT SBRT. HT had a small but significantly higher percentage of PTV coverage at 96.2% for HT, 95.7% for SF and 95.3% for VMAT (p=0.0166). However, there was also significantly higher V5 to lung with HT at 23.6%, compared to 18.1% for SF and 15.2% for VMAT (p=0.0001). Maximum esophagus dose was significantly higher with HT at a mean of 17.6 Gy compared to a mean of 12.63 Gy with SF and 10.16 Gy with VMAT (p=0.0032). Parameters such as R30Gy, R50%, and CI were significantly better with VMAT SBRT. R30Gy was 3.0 with VMAT versus 4.0 for SF and 4.8 for HT (p<0.0001). R50% was 4.0 with VMAT versus 5.2 with SF and 6.4 with HT (p<0.0001). CI was also the best with VMAT with a CI of 1.0 versus 1.2 for SF and 1.1 for HT (p<0.0001).

Table 2.

Dosimetric data by SBRT modality.

	Static Fields Mean (stderror)	Helical Tomotherapy Mean (stderror)	VMAT Mean (stderror)	p-value
Total Dose	5135.2 (48.21)	5120.6 (84.73)	5118.9 (81.22)	0.9794
Number of fractions	3.44 (0.08)	3.65 (0.14)	3.54 (0.13)	0.4261
GTV Vol (cc)	8.85 (1.350)	9.26 (2.067)	10.16 (1.909)	0.8546
GTV Min Dose (cGy)	5670.1 (83.16)	5609.9 (127.35)	5434.4 (117.61)	0.2639
PTV Vol (cc)	35.26 (3.49)	35.7 (5.34)	36.7 (4.93)	0.9734
PTV Min Dose (cGy)	4438.4 (61.90)	4649.9 (94.79)	4456.9 (87.54)	0.1362
PTV Coverage (%)	95.7 (0.14)	96.2 (0.22)	95.3 (0.21)	0.0166
V20 (%)	4.1 (0.33)	5.0 (0.51)	4.2 (0.48)	0.2993
V5 (%)	18.1 (0.93)	23.6 (1.42)	15.2 (1.33)	0.0001
Max Rib Dose (cGy)	3903.3 (182.8)	3836.2 (265.39)	4282.8 (240.84)	0.3642
Max Esophagus Dose (cGy)	1263.2 (106.12)	1760.6 (160.09)	1016.7 (147.86)	0.0032
2cm Max Dose (cGy)	3070.3 (58.72)	2845.0 (91.41)	2927.8 (79.93)	0.0879
2cm Max Dose (%)	59.0 (1.24)	55.6 (1.93)	57.0 (1.68)	0.2958
R30Gy (unitless)	4.0 (0.17)	4.8 (0.26)	3.0 (0.23)	<0.0001
R50% (unitless)	5.2 (0.19)	6.4 (0.29)	4.0 (0.26)	<0.0001
CI (unitless)	1.2 (0.02)	1.1 (0.02)	1.0 (0.02)	<0.0001

Clinical outcomes were also measured and there were no statistically significant differences between the modalities, including pneumonitis, grade of pneumonitis, toxicity other than pneumonitis, and local control (Table 3). For patients followed for more than one year, 13/137 patients (9.4%) had evidence of local failure which was often accompanied by regional and distant disease. There was a trend toward statistical significance (Figure 1) with time to local failure with VMAT being the most favorable group on the Kaplan-Meier curve (p=0.0733). Actuarial analysis showed that local control was not significantly different among the groups for any timepoint. At one year, local control was 95.9% for SF SBRT, 85.1% for HT SBRT and 100% for VMAT (p=0.7948). At three years, local control was 81.2% for SF SBRT, 69.6% for HT SBRT and 100% for VMAT (p=0.0892). At five years, local control was 81.2% for SF SBRT, 69.6% for HT SBRT and 100% for VMAT (p=0.0884). There were 158 patients with BED > 100Gy and only 18 with BED ≤100Gy. Neither rates of nor time to local failure between these groups was statistically significant (p=0.7464). When dosimetric parameters were compared with clinical outcomes, Grade 2+ radiation pneumonitis was associated with a larger median GTV of 21.39 cc versus 7.65 cc (p=0.0016). Similarly, Grade 2+ radiation pneumonitis was associated with a larger median PTV of 65.62 cc versus 31.75 cc (p=0.0030). A higher V20 was also associated with Grade 2+ radiation pneumonitis 6.62% versus 4.08% (p=0.0408).

Table 3.

Clinical patient outcomes by SBRT modality.

	Static Fields Mean (stderror)	Helical Tomotherapy Mean (stderror)	VMAT Mean (stderror)	p-value
Toxicity other than pneumonitis (% yes)	10.5	7.7	3.5	0.4969
Pneumonitis (% yes)	11.6	7.7	13.8	0.7693
Grade of Pneumonitis	1.5 (0.28)	2 (0.65)	1.5 (0.46)	0.7432
Local failure (%yes)	10.6	16.7	0	0.1057

Figure 1.

Kaplan- Meier curve for time to local failure for different SBRT modalities

Discussion

By definition, SBRT delivers doses greater than or equal to 6 Gy per fraction in one to five fractions to a relatively small (typically less than or equal to five centimeters) mass. Given the small treatment margins and high local radiation doses, most stereotactic setups must ensure reproducible immobilization during treatment, control of or allowance for respiratory movement of the target, and appropriate treatment planning in order to enable reproducible submillimeter treatment accuracy. Robust quality control parameters should be a major consideration to ensure precision treatment delivery to achieve the high local control along with minimal toxicities that SBRT promises. SBRT was first validated as a reasonable treatment mechanism for early-stage NSCLC when the Radiation Therapy Oncology Group 0236 study was published. Results from this study showed that only 13% of patients experienced a Grade 3 toxicity and 4% of patients experienced a Grade 4 toxicity. Three-year tumor control was 98% and three-year locoregional control was 87%[6]. As new technologies emerge, careful assessment of their strengths and limitations is necessary. The current study looked at the long-term outcomes of intensity modulated or three-dimensional (3-D) SF, HT and VMAT for SBRT treatment of early stage NSCLC.

Several studies which have compared SF versus VMAT SBRT showed an improved CI with VMAT [7,8,9]. A study by Palma et al. found that there was no significant difference between Grade 3 or 4 pneumonitis or radiological changes with VMAT versus SF SBRT[11]. The majority of SBRT dosimetric comparison studies concentrate on the differences between VMAT and SF SBRT, with only a few studies comparing HT to other modalities. Baisden et al. demonstrated that HT is a well-tolerated method of SBRT[12]. A comparison was made between HT, intensity modulated fixed-gantry SBRT and 3-D SF SBRT, which found that the D2cm was more favorable in HT and intensity modulated fixed-gantry SBRT[13]. Few studies have compared SF, VMAT, and HT SBRT[14,15,16]. Unfortunately, these studies do not report clinical outcomes and are plagued by a small sample size (between 6-10 lung cancer patients).

SF SBRT, HT SBRT and VMAT SBRT have never been compared to each other with regard to dosimetric parameters or clinical outcomes such as toxicity, local control and survival in early stage NSCLC. Other studies have been hampered by low patient numbers, lack of clinical outcomes and dosimetric only studies. In the current study, dosimetric parameters from larger numbers of treatment plans of each modality were compared to observed clinical outcomes. The compared dosimetric parameters include V5, V20, D2cm, GTV size, PTV size and the CI. Clinical outcomes that were collected include local control, Grade 2+ radiation pneumonitis and other toxicities to detect any previously unknown associations between each dosimetric planning parameter and observed clinical outcome.

This study found the HT SBRT had a slightly better PTV coverage when compared to SF and VMAT SBRT. Similarly, Chi et al. found a better D95 with HT as opposed to VMAT[17]. The study found significantly higher V5 and maximum esophagus doses with HT. Conceptually this makes sense because of the helical treatment nature of tomotherapy. Other studies have found no difference in the V5 to the lungs or maximum esophagus doses between SF SBRT or HT SBRT[13,16]. A lower V5 with VMAT as compared to SF SBRT has been described in other studies[8,10,18]. In the current study, the R30Gy and R50% were significantly lower with VMAT SBRT as compared to SF SBRT or HT SBRT. This is at variance from what was found by Weyh et al. who found that HT provided the lowest R50%[16]. Chiu et al. showed better R50% with VMAT compared to SF SBRT[19]. Similarly, this study found the best CI was given by VMAT SBRT with a CI of 1.0, but Weyh found the best CI for HT [16]. Possible reasons for this discrepancy include that the Weyh study was a dosimetric study that only used 8 patients with simulated plans using SF SBRT, HT SBRT and VMAT SBRT for only peripheral lesions. This study has a larger sample size. These plans have actually been delivered to patients. This study included tumors which were in any lobe of the lung and it included various dose and fractionation schemes. Some papers comparing SF SBRT to VMAT SBRT found VMAT SBRT to have better CI [8,9,19,20,21].

The current study confirmed that for the entire cohort of patients, overall local failure was documented in <10% of patients. The 3 and 5 year actuarial local control rates appear lower than other reports, but this may be due to the fact that most patients were medically inoperable with significant comorbidities, the sample size was small and patients with local recurrences were more likely to re-present for further therapy. Very few patients failed only locally with most developing combined local, regional and distant failure. There was a trend for better local control with flattening filter free (FFF) VMAT SBRT when compared to SF or HT SBRT. With regard to outcomes, one study by Navarria et al. looked at local control and found that VMAT delivered in FFF mode had a statistically significant higher local control than SF beams treated with a flattening filter[10]. The improvement in local control may be due to the higher dose rate when using FFF VMAT beams. Evidence has shown that higher dose rates result in reduced clonogenic survival in certain cell lines[22]. A caveat is that both of these studies are small and that this is preliminary data. Generally, treatment times and V5 are less with VMAT[7,8,9]. Shorter treatment times reduce the risk of motion after the initial set-up. Improved local control also may be due to better visualization of the tumor volume when using kilovoltage cone beam CT scans. These findings must be interpreted with caution.

This study found that a higher median GTV, a higher median PTV and a higher median V20 (6.62% versus 4.08%) significantly correlated with a higher rate of pneumonitis. All three of these factors are likely related. While normally a V20 of 6.1% would be thought to be associated with a relatively low risk of pneumonitis of any grade using standard fractionation, this may need to be revisited with regard to SBRT. Indeed, other studies have found that a V20 of 6-10% is associated with higher rates and higher grades of pneumonitis[23,24]. Additionally, Matsuo et al. showed a significant relationship between radiation pneumonitis and PTV size[25].

There were a few limitations of this study. This was a retrospective study which is subject to bias by its nature. Also, the specific treatment algorithms used to plan with each modality were not carefully examined in order to see if that accounted for some of the differences seen. As always, there were planner-dependent differences which could not be perfectly accounted for.

As SBRT becomes more generally accepted as a primary modality for treatment of all early stage NSCLC patients, we must gain confidence that treatment platforms at different institutions can deliver the treatment with precision and accuracy. While there are differences in dosimetric characteristics demonstrated in this study, there was minimal clinical difference in outcomes which is reassuring. The present study goes beyond the typical dosimetric studies to plans used to treat patients with varying SBRT doses, tumor sizes and location within the lungs. This study also reported retrospective clinical outcomes which are normally absent in conventional dosimetric studies. It still remains that robust quality assurance and care with respect to patient setup, treatment planning and attention to detail in all aspects of SBRT treatment is of supreme importance to gain the results promised by SBRT. This study confirms this promise.