Long term results of single high dose Stereotactic Body Radiotherapy in the treatment of primary lung tumors

Stereotactic body radiotherapy (SBRT) is a standard treatment for inoperable early-stage NSCLC, with local control rates comparable to surgical series. Promising results have been achieved utilizing a high single-dose schedule. The aim of our study was to evaluate long-term local control and toxicity in a series of patients treated with SBRT delivered in a single dose of 30 Gy. 44 patients affected by early stage NSCLC were treated with SBRT delivered in a single dose of 30 Gy. Survival and prognostic factors were retrospectively evaluated. Median follow-up was 34 months (range 3–81). Three- and 5-year local progression-free survival (LPFS) were 87.8% and 87.8% respectively (median 30 months; range 6–81 months), 3- and 5-year OS and CSS were 64.9% and 36.9%, 80.9% and 65.5%, respectively. Two (4.6%) cases of grade 3 pneumonitis occurred. At the univariate analysis lesion diameter ≤ 25 mm was predictive of better 5-year LPFS (95.8% versus 56.3%; p = 0.003) and 5-year PFS (69.8% versus 27.8%; p = 0.002). The results of our study indicated a high local control, survival and tolerability after a long-term follow-up with the use of SBRT 30 Gy single dose. Further prospective studies could better define the role of this regimen.

The current study was carried out according to the Declaration of Helsinki (1964) and was approved by the Internal Review Board (Department of Radiation Oncology, Sant' Andrea Hospital, "Sapienza" University, Rome). Written informed consent was obtained by all patients. treatment. Details of SBRT planning and delivery at our Institution have been extensively described in previous publications from our Department 5,10 . Briefly, all patients underwent a 4-dimensional (4D) pre-treatment planning CT. The maximum intensity projection was reconstructed using software (Advantage 4D, General Electrics Company, Waukesha, WI) from the 10-phase 4D-CT images and was used to delineate the internal target volume (ITV) from the gross tumor volume (GTV). Planning CT images were matched with diagnostic PET/ CT images for the GTV delineation. The planning treatment volume (PTV) was determined by adding 4-5 mm in all directions to the ITV.
The prescribed dose to the PTV was 30 Gy in one single dose (biological equivalent dose 10 [BED10] = 120 Gy) at the 95% isodose with normalization to the maximal dose, for all cases.
Patients' positioning was verified before treatment using an in-room cone-beam (Kilo-Voltage) CT scan. The treatment was delivered with a Linear Accelerator with 6-MV photons, using 7-9 static non-opposing coplanar fields.
Follow-up and statistics. Treatment-related adverse effects were assessed at each follow-up according to CTCAE v 4.0. The first follow-up was performed 6 weeks after SBRT with a Chest-CT scan. The following follow-up was performed with a CT scan with contrast medium or FDG/PET-CT every three months for the first two years after SBRT and every six months afterwards.
Local recurrence was defined by the dynamic enlargement of the local tumor on follow-up CTs that continued for at least 6 months and by the increasing of metabolic values at the FDG-PET, routinely used. Moreover, in-field recurrence was defined as any recurrence occurred within the 95% isodose curve and marginal recurrence as any recurrence occurred within the 50% isodose curve.
Survivals were defined as follows: LPFS as the time to occurrence of in-field or marginal regrowth of the disease; PFS as the time to local/distant progression or death; MFS as any site of distant progression (including the ipsilateral lung); CSS was defined as the time to death by cancer or last follow-up; OS as the time to death or last follow-up. Survivals were estimated using the Kaplan-Meier method. Prognostic factors such as age, sex, primary histology, lesion diameter, GTV size, PTV size, type of response and severe toxicity were included in the statistical analysis. Univariate analysis was performed to determine significant prognostic factors using the log-rank test or the Cox method for continuous variables. The multivariate analysis was performed with the multiple logistic regression method and the log-rank test to identify predictive factors; we included in the analysis all the clinically relevant variables. Variables were included in the multivariate analysis according to the correlation at the univariate analysis (P = ≤0.2). The threshold of lesion size related to SBRT response rate and survival was determined with the ROC curve method, calculating the highest product of (sensibility*specificity) 16 . The Statistical analysis was performed using the SPSS statistical software package version 22.0 (SPSS Inc, Chicago, IL). A p-value ≤ 0.05 indicated a significant association.

Results
patients' characteristics. We treated 44 patients with primary lung tumor. Twenty-nine (66%) patients were male and 15 (34%) were female. Initial stage of disease, patients' and tumor characteristics are reported in Table 1.
pattern of relapse and treatment. Thirteen (32.5%) patients developed local and/or distant progression. Local progression occurred in 4 (9%) patients after a median time of 14 months (range 7-22 months), and was treated as follows: 2 patients received reirradiation using stereotactic technique (50 Gy/5 fractions) and two patients received systemic therapy. One (2.3%) patient developed a relapse to the mediastinal lymphnode and was treated with systemic therapy. Distant progression occurred in 11 (25%) patients after a median time of 22 months (range 2-40 months) and was treated as follows: 6 patients received systemic therapy, 3 patients had oligoprogression other than the previous one site (lung, brain and liver, respectively) and were treated with SBRT plus systemic therapy, and 2 cases received only best supportive care.
Toxicity. Grade 2 pneumonitis had been detected in 7 (15.9%) patients. Two (4.6%) cases had grade 3 pneumonitis, one of which had a known history of chronic obstructive pulmonary disease and both cases required pharmacological treatment and recovered without sequelae (Table 4); one (2.3%) patient developed grade 2 esophagitis. Moreover Dose-Volume Histograms (DVH) of patients with severe pneumonitis were analyzed and constraints were respected. See dose to critical normal structures in Table 5.

Discussion
SBRT is a standard treatment for early stage NSCLC, for patients not suitable for surgery or in the case of refusal. SBRT can achieve rates of local control at 5 years of 83.9%, comparable to the 80% of surgical series, as widely demonstrated in many studies 7,8 . On the other hand, SBRT treatment may be characterized by a lower rate of mediastinal nodal control, as compared to surgery, even though the evidences is herein conflicting [17][18][19][20] . This might be related to the nodal dissection performed during surgery. Nevertheless, SBRT presents several advantages such as the possibility to be performed also in impaired patients, lower costs and toxicities, low engagement in an outpatient situation, a rapid integration with systemic therapies and parenchymal lung preservation. The optimal schedule is not yet standardized, since many prospective ongoing trials are evaluating it. We previously   www.nature.com/scientificreports www.nature.com/scientificreports/ published a large retrospective series of single dose of 30 Gy SBRT, delivered to lung metastases from different primary tumors, demonstrating a good outcome, especially in small lesions from primary NSCLC 5 . Therefore we analyzed the outcome of a series of early-stage NSCLC treated with stereotactic technique delivered in a single dose of 30 Gy. Long term survival and prognostic factors were evaluated.
Single dose SBRT presents the particular advantage to avoid the intrafraction uncertainties, as compared to multifraction regimens, but 4D simulation and pre-treatment cone beam CT are mandatory, as recommended by the ESTRO ACROP guidelines 1 .
Ma et al. 21 compared 65 small early stage NSCLC treated with 30 Gy via single dose with 94 patients treated with three-fraction schedule SBRT (48-60 Gy), reporting no differences in 2-year local control (87.7% and 86.2% for single-fraction and three-fraction regimen, respectively) and 2-year OS (63.2% and 61.6%, respectively). Only one case of grade 3 pneumonitis occurred in the three-fraction group, with no differences in toxicity rates. Finally the single dose was established as the standard regimen at their Institution.   www.nature.com/scientificreports www.nature.com/scientificreports/ Videtic et al. 13 treated 82 Stage I medically inoperable NSCLC: 80 patients received single doses of 30 Gy (n = 55) and 34 Gy (n = 25) delivered with SBRT. Unexpectedly the results at 1-year seemed to favor 30 Gy versus 34 Gy in terms of local failure (2% vs. 13.8%), occurrence of distant metastases (10.6% vs. 20.9%), OS (75% vs. 64%) and lung cancer-specific mortality (2.1% vs. 16%). Moreover, no grade 3 toxicity was reported in both arms. Due to the retrospective nature of the study, the authors themselves conclude that these results should be considered with caution.
The phase II trial RTOG 0915 randomized 84 patients affected by stage I peripheral NSCLC to receive SBRT in a single dose of 34 Gy or 48 Gy in 4 fractions. The study fulfilled the primary end-point of safety, with 7.9% of severe acute toxicity in the single dose arm, compared to the 15.8% of multi-fraction schedule. Preliminary results of local control, the secondary end-point of the study, showed 97% and 92.7% at 1 year, respectively, while 1-year OS was 84.6% and 91.1%, respectively. Due to these encouraging results the single dose schedule was considered better tolerated and more effective than the multi-fraction regimen 7 . Nevertheless no factors were evaluated as useful in the selection of patients. Cummings et al. 22 compared 65 patients affected by early stage NSCLC who were treated with SBRT 30 Gy single dose, with 98 patients who were treated with SBRT 50 Gy in 5 fractions. The results of the propensity-matched analysis showed again no differences in 2-year LC (92% versus 82%, p = 0.38; for single dose and five-fractions, respectively, and in the 2-year OS (68% versus 74%, p = 0.18). There was one case of grade 3 pneumonitis in the single dose group, but no differences were observed when compared with the five-fraction regimen.
The long-term results of the current study and the related LPFS at 5 year of 87.8% are in line with previous publications. We observed that most of recurrences occurred during the first 2 years of follow-up that may justify an intensive and multimodal approach.
One of the major concerns regarding the use of SBRT is the possibility of a higher mediastinal nodal relapse, as compared to surgery 4,6 . The cumulative incidence of regional recurrence after SBRT for early stage NSCLC, diagnosed with PET-FDG ranged from 0 to 28.6% with a median incidence of 9.6% 23 . We reported only one case of mediastinal nodal relapse, which can be related to the systematic use of PET in the staging that allowed for a more accurate selection of patients, and the enhancement of detection rates of mediastinal relapse after SBRT 24 .
Local progression in our series occurred in 4 (9%) tumors after a median time of 14 months: 3 out of 4 of these occurred in lesions >25 mm, which can represent a threshold of effectiveness and can be used to select early-stage NSCLCs' best population suitable for the treatment with 30 Gy SBRT. It is known that smaller lesions respond better to SBRT 1,25-28 , but no studies, to data, have assessed dimensional prognostic factors of response after 30 Gy SBRT in this setting. Moreover some controversies exist regarding the definition of local failure, since some authors define it as any relapse within the same lobe and/or of the treated lesions 7,29 , while others consider it as the sole failure of the treated lesion 30,31 and this can complicate the comparison and interpretation of the results.
The cut-off of 25 mm represents a novel finding, since patients' selection is crucial for maximizing treatment outcome. Larger tumors with hypoxic areas may benefit more from a multi-fractions regimen that can directly induce cell death and take partial advantage of the reoxygenation, while smaller tumors may benefit the most from a high dose treatment in a single dose which affects the tumor vessels and indirectly determines cell death 11,31 .
Toxicity of our series was mild and in line with previous publications on 30 Gy single dose treatments, with pneumonitis occurring in 1.7-3.6% of cases 5,13,14,[32][33][34][35] . We reported two cardiac deaths in patients with severe cardiac and pulmonary diseases. By the DVH analysis and considering the location of the treated lesions away from central organs we did not ascribe these events as treatment-related death, also considering the recent literature on the correlation between dose to cardiac structures and cardiac event 36 .