Lung Stereotactic Radiation Therapy: Intercomparison of Several Irradiation Devices in Terms of Outcome and Predictive Factors

Background: Stereotactic body radiotherapy (SBRT) is recommended for the treatment of inoperable early stage non-small-cell lung cancer and lung oligometastases. The radiation oncology department of the Lausanne University Hospital (CHUV) gathers three different radiotherapy devices able to treat pulmonary lesions in SBRT conditions: CyberKnife® (CK), Helical Tomotherapy® (HT), and volumetric modulated arc therapy (VMAT). The aim of this study is to dene the patients’ outcome in terms of irradiation ecacy and toxicities after lung SBRT depending of the choice of the SBRT technique. Methods: We retrospectively analyzed the clinical, radiological, and dosimetric data of patients with primary lung tumor or pulmonary oligometastases treated with SBRT between January 2016 and February 2020. We analyzed descriptive data using the Chi-2 test for proportions and the T-test for means comparisons, survival data by the Kaplan-Meier method and comparisons between groups by the Log-rank test and Cox-regression. Results: We identied 111 patients mostly in good condition (82.9% PS 0-1) with a median age of 71.4 years. One hundred forty-two lesions were treated with a typical fractionation of 55 Gy in 5 fractions, delivered by CK (59.9%), VMAT (38.0%), or HT (2.1%). Compared to other techniques, CK technique allowed to treat comparable gross tumor volume (GTV; 2.1 vs 1.4cc, p = 0.84) with smaller planning treatment volume (PTV; 12.3 vs 21.9 cc, p = 0.013), and was associated with a lower mean lung dose (MLD; 2.6 vs 4.1 Gy, p < 0.001), a lower V5 (13.5 vs 19.9 cc, p = 0.002) and a lower V20 (2.3 vs 5.4 cc, p < 0.001). Local control rates at 2 years were not different depending on the irradiation device, respectively of 96.2% (range, 90.8-100) and 98.1% (range, 94.4-100), p = 0.68. Toxicity incidence was signicantly increased with V5 value > 17.2% (56.0 vs 77.4%, p = 0.021). Conclusions: Compared to other SBRT techniques, CK treatments permitted to treat comparable GTV with reduced PTV, MLD, V20, and V5. The dosimetric benet of CK SBRT was not associated with a clear clinical benet, with comparable outcome in terms of control rates and toxicity. Toxicity incidence was less frequent when reducing the V5. The use of CK is particularly attractive in case of multiple courses of lung SBRT or in case of local relapse requiring lung re-irradiation. on


Background
Lung cancer is the most common cause of cancer-related death (1). The introduction of pulmonary stereotactic body radiotherapy (SBRT) for early malignancies has been correlated to a reduction in the number of patients who remain untreated (2,3). SBRT is recommended for early stage non-small-cell lung cancer (NSCLC), either in medically or surgically inoperable patients, or in well-selected operable patients, according to clinical guidelines (4)(5)(6)(7). In addition, various SBRT indications for lung oligometastases are well known(8).
Synergy® is a linear accelerator (linac) that uses volumetric modulated arc treatment (VMAT). The second device, termed helical Tomotherapy® (HT), combines a megavoltage computed tomography (MV-CT) and a miniaturized linac to provide image-guided radiation therapy. During helicolidal irradiation, the machine slowly spins around the patient while the table travels longitudinally. When using Synergy® and HT, an internal target volume is either determined, on a 4D CT or in obstructed inspiration owing to a respiratory servo system called Active Breath Control® (ABC). When using a 4D-CT for planning, the gross tumor volume (GTV) of each phase is combined into an internal target volume (ITV). This method is the most commonly used when irradiation is done using VMAT or HT. The planning target volume (PTV) corresponds to an ITV with a 3-mm margin in this con guration. The CyberKnife® (CK) is a linear accelerator with a head and a robotic arm, totally dedicated to SBRT. Several tracking systems allow for alternative irradiation treatments with this device: i) Synchrony®: respiratory monitoring with placement through ducial markers; ii) X-sight Spine®: monitoring of the spine associated with a ITV delineated based on a 4D CT-scanner; and iii) X-sight Lung® with monitoring of the tumor itself (9,10).
When using the CK tracking method, the GTV is immediately expanded by 3 mm to the PTV. As a result, CK may be able to lower GTV to PTV margins while avoiding ITV-related volume expansion.
Within the same radiation oncology department, only a few institutions have different devices able to perform pulmonary SBRT. The aim of this study is to determine the patients' outcomes in terms of irradiation e cacy and toxicities after lung SBRT according to the different SBRT techniques.

Methods
Between January 2016 and February 2020, we reviewed the clinical, radiological, and dosimetric data of consenting patients with primary lung tumors or pulmonary oligometastases treated by SBRT at CHUV. All patients were selected for SBRT after a multidisciplinary thoracic oncological tumor board discussion, and signed an informed consent form for irradiation. Patients having a lung tumor treated with a method other than SBRT, such as surgery or conventional radiotherapy, were not included.
Clinical data such as the patient's age, sex, WHO performance index, smoking status, comorbidities, previous medical history including previous irradiation or thoracic surgery, cancer histology, pulmonary location of the tumor (affected lobe, central or peripheral), and the reason for radio-oncological treatment rather than the surgical one, were recorded. The reason for choosing SBRT modality was at the discretion of the physician.
The fractionation schedule, GTV, PTV, PTV dose, dose received by the lungs, i.e. the mean lung dose (MLD), the percent of the volume of the lung receiving 5 Gy (V5) and 20 Gy (V20), and composite dosimetry in the event of multiple courses of irradiation were all part of the dosimetric data record.
Radiological data included tumor response after SBRT (stable disease, complete or partial response to treatment), and any recurrence (local, locoregional or distant metastatic) were described.
Clinical and radiological data were extracted from the institutional clinical software Soarian® (Cerner, United States) and Archimède® (CHUV) software; dosimetric data were recorded from RayStation® (RaySearch Laboratories, Sweden) and VelocityTM (Varian Medical Systems, USA) radiotherapy software. Data were then anonymized by assigning to each patient a neutral identi er of 4 digits, and the study was registered on February 24th 2021, ID 2021-00267, with the authorization of the CER-VD ethics committee (Switzerland). Statistical analyses were performed using the JMP® 15 software. We analyzed descriptive data using the Chi-2 test for proportions and the T-test for mean comparisons. Survival data were analyzed by the Kaplan-Meier method, and

Patients' and treatment characteristics
We identi ed 111 patients, mostly smoker (64.9%) men (65.8%) in good condition (82.9% PS 0-1) with a median age of 71.4 years. Patients presenting primary pulmonary tumors (n = 57) and lung metastases (n = 54) were found in equal numbers. Patients, who were medically inoperable due to comorbidities, were the most common cause for not selecting for surgery, and instead choosing SBRT (44.2%). We treated 142 pulmonary lesions, the majority being peripheral tumors (73.9%). More than half of the patients had a history of thoracic surgery (52.3%), and about a quarter had a history of thoracic irradiation (23.4%). The prescribed dose ranged from 34-60 Gy over 1-8 fractions with a typical fractionation of 55 Gy in 5 fractions (66.7%), delivered by CK (59.9%), VMAT (38.0%), or HT (2.1%). A large majority of patients bene ted from a tumor tracking through ducials (51.4%), and many of others from an ITV Plani cation with 4D-CT (33.8%). The median MLD was 3.7 Gy, with a median V5 at 17.2%. In case of iterative irradiation, the median composite MLD and composite V5 reached 7.9 Gy and 37.9% respectively.
Other patients and treatment characteristics are presented in Table 1. comparisons between groups by the Log-rank test. Con dence intervals (CI) were computed from standard errors.
If p-value was ≤ 0.15 in univariate analysis, the variable was tested in multivariate analysis through Coxregression test.  Table  2. LC was not in uenced by the other treatment characteristics: the choice of radiotherapy device, neither by the SBRT modality, nor by the dose/fractionation scheme. We observed 2 mediastinal relapse, corresponding to a loco-regional control rate of 96.2% at 1 year and 94.4% at 2 years, respectively.  rst recurrence at brain (13.5 vs 5.5 months, p = 0.007 and 0.010, respectively). MFS and PFS were signi cantly longer when treated lesions were primary lung tumors compared to metastases, while median MFS was not reached and PFS nearly doubled in case of primary lung cancer. However this last factor was not con rmed after multivariate analysis. Median survival, signi cantly decreased in case of rst metastatic relapse occurred before 13.1 months. Median OS was 37.8 months when rst metastatic relapse occurred before 13.1 months compared to a median OS of 64.5 months when rst metastatic relapse occurred after 13.1 months (p = 0.035), and remained predictive after Cox-regression (p < 0.0001). Other predictive factors for survival (OS, MFS, and PFS) are depicted in Table 3. Toxicity is described in Table 4.     notice that this study did not report the dosimetrical parameters speci c to the device choice. The study of Claude et al, represents the only prospective trial other that confronted the two modalities; however, we can underline that the treatments were not performed in the same institution (multicentric study), and could induce bias for this direct comparison (13). Even retrospective, our study represents the rst one directly comparing the outcomes of patients receiving lung SBRT using three different modalities in the same radiotherapy department.
In our study, all patients that presented a local relapse were older than 71 years old. We identi ed that oldest patients presented a signi cantly lower local control compared to younger ones. A retrospective study of 219 patients also reported that age signi cantly in uenced the local outcome after SBRT (14). On the other hand, Watanabe, et al depicted a preserved local control in a cohort of 64 patients older than 80 years old, which was 98.4% at 3 years (15).
In our cohort about 40% of patients were treated without histological proof of malignancy and indication for irradiation was based on clinical and radiological criterias and this subset of patients presented similar outcomes compared to those with a histological proof of malignancy. In this context, several publications already showed comparable outcomes in patients after empiric SBRT, supporting the hypothesis that patients did have lung cancers(16-18). Loco-regional control was good and comparable to other studies of lung SBRT and also in compare to surgical outcomes after video-assisted thoracoscopic surgical lobectomy with mediastinal lymph node dissection (19).
OS was signi cantly decreased in case of early metastatic recurrence (< 13.1 months), and PFS was signi cantly shorter when the rst identi ed relapse site was the brain corresponding to the usual clinical factors and in uencing the outcome in metastatic patients (20,21).
More than half of our patients presented at least one toxicity (68.5%), and the most common form was Grade 1 acute toxicity (79.3%). About three quarter (73.9%) of patients have not developed any late toxicity. Incidence of toxicity was comparable with the literature (22)(23)(24)(25). In our cohort, several dosimetric parameters were analyzed in order to identify predictive factors for toxicity such as PTV, V5, V20, MLD, and composite MLD. Compared to other techniques, and even though comparable GTV, CK treatments signi cantly permitted to decrease PTV, MLD, V5, and V20; which are dosimetric parameters usually found to be related with toxicity(26, 27). In our cohort, an increased V5 value correlated with a higher incidence of toxicity. Reducing the V5 value could be particularly interesting in case of multiple courses of SBRT lung irradiations (27). Hence, even if we not found a signi cant difference regarding toxicity in relation to the delivery device employed, it can be appropriate to favor CK irradiation in order to improve dosimetric parameters. Indeed, our study showed that cumulative MLD and cumulative V5 values are approximately doubled in case of multiple courses of lung irradiations, and techniques able to spare healthy lung reducing these dosimetric parameters could be preferred from the rst lung SBRT.
Moreover, CK could be particularly of interest in the event of local relapse re-irradiation, situation known to be at risk of high rates of severe toxicity(28, 29).

Conclusion
Compared to other SBRT techniques, CK treatments are able to treat comparable GTV with reduced PTV, MLD, V5, and V20. The dosimetric bene t of CK irradiation was not associated with a clinical bene t, with comparable outcome in terms of control rates and toxicity. Toxicity incidence was less frequent when reducing V5. Altogether, it seems appropriate to favor the use of CK, particularly in the event of multiple courses of irradiations or in case of re-irradiation. Registered on February 24th 2021, ID 2021-00267, with the authorization of the CER-VD ethics committee (Switzerland).

Consent for publication
Not applicable for that section.

Availability of data and materials
The datasets used and analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
No speci c funding was received for this investigation. The investigators have no con ict of interest in this project.
Authors' contributions ELR, EMO, and RK collected and analyzed the patient data, and were major contributors in writing the manuscript.
AC, AD, HB, ERO, AL, CVG and TK contributed in writing the manuscript. All authors read and approved the nal manuscript.