Lung stereotactic body radiation therapy using simultaneous integrated BED-escalation for peripherally located non-small cell lung cancer

Abstract

Purpose

Report the outcomes of patients with non-small cell lung cancer (NSCLC) and peripheral tumors treated with simultaneous integrated biologically equivalent dose (BED)-escalation (SIBE) lung stereotactic body radiation therapy (SBRT) to achieve dose escalation.

Materials/methods

Patients with NSCLC within 5 mm of the chest wall treated with a SIBE approach were eligible. Patients received 60 Gy in 5 fractions, with dose decreased to 50 Gy based on proximity to the chest wall. Dosimetry, oncologic outcomes, and toxicity were evaluated.

Results

Twenty-four patients met inclusion criteria. Median BED to the PTV was 135.4 Gy. Median chest wall V30 was 18.7 cc. The 3-year LC, OS, and PFS of the non-metastatic cohort was 93%, 35%, and 39%, respectively. The crude rate of chest wall toxicity was 12.5%, with no rib fractures.

Conclusions

SIBE lung SBRT appears to be well tolerated and achieves favorable local control rates and survival.

INTRODUCTION

Although surgical resection remains the standard-of-care for medically operable early-stage non-small-cell lung cancer, stereotactic body radiotherapy (SBRT) has made radiation an increasingly viable modality for inoperable tumors or in cases where patients decline surgery. Its efficacy has been demonstrated in multiple prospective trials, with 2-year overall survival (OS) and local control (LC) rates of approximately 77% and 89%, respectively, a significant improvement from 3DCRT (65% and 59%, respectively) [1-3]. This benefit is likely driven by SBRT delivering ablative doses of radiation in a highly conformal manner over 1-5 fractions, leading to higher cumulative biologically effective dose (BED) while maintaining tolerable toxicity levels.

Evidence has suggested that BED is the most significant predictor of local control in early-stage NSCLC treated with SBRT, with doses greater than 100 Gy and less than 146 Gy appearing to be optimal [4-9]. However, a competing factor is proximity to critical structures such as chest wall, due to an increased risk of severe toxicity [10-12]. Therefore, SBRT dose and fractionation is dependent on tumor location. National guidelines indicate a dose of 50 Gy in 5 fractions (BED=100 Gy) should be given for tumors adjacent to chest wall, largely due to normal tissue tolerance of the chest wall. One problem with this approach is that it can deliver a sub-optimal BED to a significant portion of the tumor that lies further away.

Starting in 2013, we developed an institutional standard to treating peripheral lung lesions with SBRT that were in close proximity to the chest wall, using a simultaneous integrated BED-escalation (SIBE) approach of 60 Gy/5 fractions while limiting 50 Gy/5 fraction to a limited peripheral edge of the tumor. We herein report our approach and clinical outcomes using this technique.

MATERIALS AND METHODS

Patient selection

After institutional review board approval, we conducted a retrospective review using an institutional database of patients with cT1-3 NSCLC (either early stage or metastatic) who were treated with SIBE lung SBRT between 3/2013 and 12/2020. This included patients who had treated tumors within 5 mm of the chest wall. All patients were staged using chest computed tomography (CT) and positron emission tomography (PET), as well as brain imaging and pathologic mediastinal evaluation where indicated. Clinical, pathologic, and dosimetric characteristics were collected. Follow-up dates were extracted from medical records.

Radiation treatment planning

All patients were initially scanned on a large-bore (85 cm) CT simulator (Philips Medical System) and 2.5-mm slices were obtained. Patients were simulated in the head-first supine position using a whole-body VakLok (CIVCO Radiation therapy, Orange City, Iowa) for immobilization. Patients were planned using either a free-breathing or end-expiratory breath hold technique. All patients were planned using the Eclipse treatment planning system (Varian Medical Systems, Palo Alto, CA). The chest wall for optimization was defined as a 2 cm expansion off the ipsilateral lung until 2014, at which point it was contoured as adjacent ribs and intervening intercostal muscles. All patients were prescribed a total dose of 60 Gy in 5 fractions, with the dose decreased to 50 Gy based on proximity to the chest wall using SIBE. Dose was delivered using an SBRT technique with conformal arcs. For patients treated on a free-breathing scan, the gross tumor volume (GTV) was initially defined as the visible mass on lung window on the average CT scan. The iterative target volume (ITV) was defined based on all gross tumor visible on all phases of a 4-dimensional CT. For patients treated on end-expiratory breath hold, the GTV was defined as all gross tumor visible on the end-expiratory breath hold CT. Initial prescription target volumes (PTV) were defined as a geometric 5 mm expansion from the IGTV (ITV for patients treated on free-breathing or GTV for patients treated on end-expiratory breath hold), except for 3 mm in the direction of the chest wall. The 60 Gy PTV was defined as the PTV cropped to be 8 mm away from the chest wall, with the 50 Gy PTV being defined as the volume within 8 mm of the chest wall. The 8 mm distance serves a dual function of allowing for dose fall-off to achieve acceptable chest wall dosimetry as well as to account for respiratory variation. A target dose constraint of less than 30 cc of chest wall receiving 30 Gy or greater was utilized for all cases, along with a goal of max 0.03 cc chest wall dose <55 Gy. Dosimetric data was extracted from final radiation treatment plans. Example treatment volumes and an example treatment plan are visualized in Figure 1.

Figure 1.

Example radiation treatment volumes (A,B; red: chest wall; ITV: blue; green: PTV60; yellow: PTV50) and treatment plan (C,D) of patient undergoing SIBE lung SBRT, including dose volume histograms target volumes (D; yellow: PTV50; red: PTV60) and chest wall (F)

Study endpoints and statistical methods

Descriptive statistics were performed on clinical, pathologic, and dosimetric characteristics. All outcome measures were defined from the of SBRT initiation. Patients underwent disease surveillance with CT chest every three months. Post-treatment response was retrospectively assessed using the Delphi consensus guidelines [13]. Chest wall and pulmonary toxicity was scored based on the Common Terminology Criteria for Adverse Events (CTCAE) version 4.0. Overall survival (OS) was defined as time to death resulting from any cause. Progression-free survival (PFS) was defined as time to disease progression, relapse, or death resulting from any cause. Local control (LC) was defined as time to disease progression within the radiation treatment volume. Patients were censored at time of last follow-up. The Kaplan-Meier method was used for survival time estimations [14]. The cutoff for data analysis was August 31, 2021(analytic date). Statistical analyses were performed using open-source libraries in Python 3.8 (PSF, Wilmington, DE).

RESULTS

Patient characteristics

A total of 24 patients met inclusion criteria. The median age of the patients was 77 years old (range: 52-93 years). Nineteen patients (79.2%) were undergoing definitive treatment while five (20.8%) were metastatic undergoing palliative/consolidative treatment. The median tumor diameter was 22.5 mm (range: 10-44 mm). Twenty patients (83.3%) were treated with a free breathing technique while 4 (16.7%) were treated with end expiration breath hold. Complete patient characteristics are listed in Table 1.

Table 1.

Patient characteristics (n=24)

Characteristic	N (%)
Age (years) (median [range])	77.0 (52.0-93.0)
Gender
Female	14 (58.3%)
Male	10 (41.7%)
Race/Ethnicity
Asian	3 (12.5%)
Black	1 (4.2%)
Hispanic White	3 (12.5%)
Native Hawaiian or Other Pacific Islander	1 (4.2%)
Non-Hispanic White	16 (66.7%)
Eastern Cooperative Oncology Group     Performance Status
0	13.0 (54.2%)
1	10.0 (41.7%)
2	1.0 (4.2%)
Smoking History
Current	3 (12.5%)
Former	20 (83.3%)
Never	1 (4.2%)
Lobe
Left lower lobe	4 (16.7%)
Left upper lobe	6 (25.0%)
Right lower lobe	5 (20.8%)
Right middle lobe	1 (4.2%)
Right upper lobe	8 (33.3%)
Histology
Adenocarcinoma	18 (75.0%)
Squamous cell carcinoma	6 (25.0%)
T Stage
T1a	1 (4.2%)
T1b	7 (29.2%)
T1c	11 (45.8%)
T2a	4 (16.7%)
T2b	1 (4.2%)
Primary Tumor Size (mm) (median [range])	22.5 (10.0-44.0)
Grade
2	15 (62.5%)
3	6 (25.0%)
Unknown	3 (12.5%)
Metastatic	5 (20.8%)
Breathing Technique
End-Expiration Breath Hold	4 (16.7%)
Free-Breathing	20 (83.3%)

Dosimetry

An example patient case with treatment volumes, chest wall, and treatment plan is visualized in Figure 1. The median PTV volume was 32 cc (range: 10.1-54.8 cc), with the 60 Gy PTV volume comprising a median of 73.2% of the total PTV volume (range: 30.9-91.2%). The mean BED to the IGTV was 141.2 Gy (range: 125.8-154.0 Gy) and the mean BED to the PTV was 135.4 Gy (range: 122.7-146.1 Gy). The median IGTV heterogeneity index was 25.6 (range: 19.0-28.4) and median PTV heterogeneity index was 35.2 (range: 24.9-61.7). The median overlap between the PTV and the chest wall was 0.4 cc (range: 0-4.8 cc). The median chest wall V30 was 18.7 Gy (range 3.1-29.6 Gy). The D0.03 cc to the chest wall was 54.7 Gy (range: 46.8-65.1 Gy). Complete dosimetry data is visualized in Table 2.

Table 2.

SIBE lung SBRT dosimetry

Parameter	Median (Range)
Chest Wall
V30 (cc)	18.7 (3.1-29.6)
V50 (cc)	0.8 (0-5.4)
V60 (cc)	0 (0-0.1)
Maximum (Gy)	54.7 (46.8-65.1)
PTV Overlap (cc)	0.4 (0-4.8)
PTV Overlap Minimum (Gy)	46.4 (39.4-49.5)
Target Volumes
IGTV Minimum (cGy)	102.1 (97.3-114.4)
IGTV Maximum (cGy)	160.1 (148.1-174.6)
IGTV Mean (cGy)	141.2 (125.8-154)
IGTV Heterogeneity Index	25.6 (19-28.4)
PTV Volume (cc)	32 (10.1-54.8)
PTV60 Volume (cc)	21.9 (7.3-41.8)
PTV60 Ratio (%)	73.2 (30.9-91.2)
PTV Minimum (Gy)*	85.9 (42-96.8)
PTV Maximum (Gy)*	162.8 (148.7-174.6)
PTV Mean (Gy)*	135.4 (122.7-146.1)
PTV Heterogeneity Index	35.2 (24.9-61.7)

*Radiation Biologically Effective Dose (BED)

Abbreviations: PTV; prescription target volume. IGTV; internal gross target volumes

Tumor control and survival

Median follow up was 1.0 years (range: 0.3-7.4 years) for all patients and 1.85 years (range: 0.4-7.4 years) for living patients. The median and 3-year LC of the patient cohort was not reached [NR] years (95%CI: 2.0-NR years) and 77% (95%CI: 39-93%), respectively. The median and 3-year LC of the non-metastatic cohort was NR (95%CI: NR-NR years) and 93% (95%CI: 61-99%), respectively, with only a single local failure. The median and 3-year OS of the non-metastatic cohort was 1.49 years (95%CI: 0.74-5.56 years) and 35% (95%CI: 13-59%), respectively. The median and 3-year PFS of the non-metastatic cohort was 0.97 years (95%CI: 0.49-NR years) and 39% (95%CI: 16-62%), respectively. Four non-metastatic patients died of progressive disease. Kaplan Meier curves are visualized in Figure 2.

Figure 2.

Kaplan Meier curves of patients receiving SIBE lung SBRT showing OS (A*), PFS (B*), LC (C,D*) *Metastatic patients excluded

Toxicity

The crude rate of chest wall toxicity was 12.5%, with one patient experiencing early grade 1 chest wall pain and two patients experiencing late grade 1 chest wall pain. No patients developed rib fractures. No patients experienced early pulmonary toxicity and only one patient experienced late symptomatic grade 2 pulmonary toxicity requiring steroids. No grade 3+ toxicities were observed.

DISCUSSION

This report of our institutional technique suggests that a SIBE technique prescribed based on proximity to the chest wall is a feasible means of increasing BED to primary tumors in patients with NSCLC. In this cohort of patients, this technique led to durable local control, particularly in patients who were receiving SBRT as definitive treatment. The approach still permitted traditional dosimetric parameters for lung SBRT to be met, without increased rates of toxicity.

The benefits of increasing BED have been a focus of extensive research, with the original threshold of 100 Gy being established by Onishi et al and Guckenberger et al [4, 9]. However, there is growing evidence that even higher doses may be beneficial. Tateishi et al report on a retrospective study that stratified 433 patients by having a maximum BED of greater than 200 Gy or a mean BED of greater than 150 Gy (HighBED), compared to patients who did not reach those thresholds (LowBED). Patients with a HighBED had lower rates of local recurrence, with 1.7% at 5 years compared to 8.3% (p=0.001) [15]. However, this additional benefit has conflicting data. In a meta-analysis of thirty-four studies, Zhang et al divided BED into low (<83.2 Gy), medium (83.2-106 Gy), medium-high (106-146 Gy), and high (>146 Gy) [8]. In this study, 2-year OS was highest in the medium (76.1%) or medium-high (68.3%) BED groups compared to the low (62.3%) and high (55.9%) BED groups (p≤0.004). Notably there was no difference in 2-year LC between the medium-high and high BED arms (94.2% vs 92.8%; p=0.453), although medium dose had significantly lower LC relative to the medium high (87.4 vs 94.2%; p=0.028) and high (87.4 vs 92.8%; p=0.031) BED groups. Therefore, while the inferior OS is likely not driven by inferior LC, it is possibly related to a slightly higher rate of Grade 3-5 toxicity (9.3% in the high cohort vs 7.8% in the medium-high cohort). Regardless, these data are still suggestive that escalation beyond a BED of 100 Gy, where feasible, could be beneficial. Our study provides a possible way of doing that, while keeping toxicity acceptable, as our mean PTV BED of 135.5 Gy would be classified as medium-high, and achieved comparable rates of control (3-year LC of 93%) and acceptable rates of toxicity.

With regards to the dose fractionation used in our study, there has been limited data directly comparing 50 Gy/5 fractions to 60 Gy/5 fractions, but in one small retrospective study of 43 patients, Kim et al report that a dose of 60 Gy was not associated with a significantly improved response rates or local control, although the 60 Gy cohort was numerically favored (complete response rate: 31.4% vs 12.5%; crude local control: 94.2% vs 87.5%) [16]. Notably on that study, the cases of tumor progression solely occurred in tumors larger than 2.5 cm. This study was not limited to peripheral tumors, and therefore chest wall toxicity was not reported.

This is not the first study to tailor radiation dose based on proximity to the chest wall, although such approaches have optimized overall prescription dose and have not used an SIBE technique. Lagerwaard et al applied a risk-adapted schema, with 3x20 Gy for T1 tumors, 5x12 Gy for T2 tumors or T1 tumors with broad contact with the thoracic wall, and 8x7.5 Gy for tumors adjacent to the heart, hilum, or mediastinum. Outcomes were excellent, with crude local failure of 3% and a chest wall pain rate of 12% [17]. Similarly, Coroller et al applied risk adapted lung SBRT, where plans that could achieve a chest wall constraint of V30≤30 Gy and minimal chest wall contact were treated to 54 Gy in 3 fractions, while tumors with broad chest wall contact where the chest wall constraint could not be achieved were treated to 50-60 Gy in 5 fractions [18]. This approach led to favorable outcomes, with a 2-year estimate of local control of 83.0. Furthermore, they reported an overall 8.3% risk of chest wall pain, a 6.9% risk of rib fracture, and a 13.9% risk of any chest wall toxicity. A benefit of our approach compared to ones used in these studies is that it also allows for better dose uniformity within the volume overlapping with chest wall; an approach using a single 60 Gy volume would have a sharper dose-fall off to the chest wall and does not take into account motion uncertainty. The net result of a single volume is therefore higher heterogeneity within the target volume close to the chest wall, which could result in higher than expected doses being given to the abutting chest wall due to respiratory uncertainties, and thereby serves as a theoretic source of increased chest wall toxicity risk.

Table 3 details the chest wall toxicities in our study relative to those in other published series, including Lagerwaard et al and Coroller et al [6, 17-22]. Our study overall compares favorably to these studies, with better or comparable rates of chest wall pain, and no Grade 3 chest wall pain or rib fractures. Notably, the studies with higher rates of toxicity than our study did not utilize chest wall dosimetric constraints, as they provide the basis for modern chest wall dose constraints (30 cc≤30 Gy, 5 cc≤40 Gy, and max dose≤50 Gy) [19, 22]. Additionally, our results are consistent with results from Ma et al, who performed a meta-analysis of 57 studies and 5985 cases [23]. This study identified the risk of developing grade ≥1, grade ≥2, and grade ≥3 chest wall pain as 11.0%, 6.2%, and 1.2%, respectively, which is similar to the rates observed in our study. Also of note, our study and Coroller et al were the only studies limited to peripheral tumors, which would be expected to be at a higher risk of chest wall toxicity, although this was not observed, suggesting a risk adaptive approach achieves its desired outcome. Regardless, our study in combination with existing literature suggests that a higher mean-BED (>100 Gy) can be given for tumors close to the chest wall with excellent long term local control and low rates of late toxicity.

Table 3.

Comparison of chest wall toxicity from our study versus previously reported studies

Study (n)	Chest Wall Pain (%)	Chest Wall Grade (1/2/3)	Fractionation Schemas
Dunlap et al [19] (n=61)	32.8	2/1/17	3-5x7-12 Gy
Stephans et al [20] (n=48)	22.2	4/6/0	3x20 Gy
Woody et al [21] (n=102)	18.9	6/13/1	3x20 Gy, 4x12 Gy, 5x10 Gy, 10x5 Gy
Andolino et al [22] (n=203)	15.8	Unknown	2-5x6-24 Gy
Lagerwaard et al [17] (n=206)	12.0	Unknown	3x20 Gy, 5x12 Gy, 8x7.5 Gy
Stephans et al [6] (n=662)	11.0	29/40/4	1x30 Gy, 3x20 Gy, 5x10 Gy, 8x7.5 Gy
Coroller et al [18] (n=72)	8.3	3/2/1	3x18 Gy, 5x10 Gy
Present Study (n=24)	12.5	2/1/0	5x10-12 Gy

A notable strength of this study is that it includes a cohort of patients uniformly treated based on an institutional standard. Limitations include the small sample size, precluding any univariate or multivariable models to account for known confounders, such as age and performance status. This sample size limits the conclusions that can be made about local control rates and toxicity rates in particular, given that it may be underpowered to detect these events.

In conclusion, we show that lung SBRT using a SIBE approach is a well-tolerated treatment option for patients with NSCLC who have tumors in close proximity to the chest wall, and permits a higher mean BED than would otherwise be achievable with a single uniform 50 Gy/5 fraction prescription dose. The technique is associated with high rates of local control consistent with historical controls. As ongoing efforts are dedicated to dose escalation trials for central tumors, similar prospective research is needed for chest wall-based lesions.