Comparison of Survival Outcomes of Single- and Five-Fraction Schedules of Stereotactic Body Radiation Therapy for Early-Stage Central or Peripheral NSCLC

Simple Summary Patients with early-stage non-small cell lung cancer (NSCLC) treated with stereotactic body radiation therapy (SBRT) often receive different treatment regimens based on the location of the tumor to minimize the risk of severe side effects. The aim of our study was to examine survival outcomes of patients treated with single-fraction SBRT for peripheral tumors and five-fraction SBRT for central tumors. In a cohort of 265 patients with NSCLC, we found no differences in patients treated for peripheral versus central tumors in progression-free survival, overall survival, local failure, nodal failure, or distant failure. These findings were confirmed upon propensity score matching. Our study demonstrated survival outcomes of patients treated with SBRT for early-stage NSCLC were equivalent for central and peripheral tumors. Abstract Background: The treatment of early-stage non-small cell lung cancer (NSCLC) with stereotactic body radiation therapy (SBRT) frequently involves different fractionation schemes for peripheral and central tumors due to concerns with toxicity. We performed an observational cohort study to determine survival outcomes for patients with peripheral and central NSCLC treated with SBRT. Methods: A single-institutional database of patients with early-stage NSCLC treated with SBRT from September 2008 to December 2018 was evaluated. Outcomes were progression-free survival (PFS), overall survival (OS), local failure (LF), nodal failure (NF), and distant failure (DF). Cox multivariable analysis (MVA), Kaplan–Meier plotting, Fine–Gray competing risk MVA, and propensity score matching were performed. Results: A total of 265 patients were included with a median follow-up of 44.2 months. There were 191 (72%) and 74 (28%) patients with peripheral and central tumors treated with single-fraction SBRT to a dose of 27 Gy and five-fraction SBRT to a dose of 50 Gy, respectively. On Cox MVA, there was no difference in OS (adjusted hazards ratio (aHR) of 1.04, 95% CI of 0.74–1.46) or PFS (aHR of 1.05, 95% CI of 0.76–1.45). On Fine–Gray competing risk MVA, there were no differences in LF, NF, or DF. Propensity matching confirmed these findings. Conclusion: The survival outcomes of patients treated with SBRT for early-stage NSCLC were equivalent for central and peripheral tumors.

Due to proximity to critical structures such as the proximal bronchial airway, heart, and esophagus, SBRT for central tumors is currently thought to increase the risk of normal tissue toxicity. In 2006, significant excessive toxicity was reported for early-stage, central NSCLC lung tumors treated with three-fraction SBRT [8]. Further analysis of this trial showed that the excess toxicity associated with central tumors was no longer statistically significant and that survival was not changed by peripheral versus central location [9]. Multiple subsequent studies have identified central lung tumors to be associated with high rates of toxicity [10,11].
Despite the updated findings of the Timmerman trial, concerns about toxicity with central lung SBRT persist. Since the initial finding, later studies of SBRT have made the central lung a "no-fly zone" and focused on either peripheral or central lung tumors. Consequently, little is known about the long-term outcomes of central versus peripheral tumors treated with SBRT. To compare central versus peripheral NSCLC SBRT on survival and recurrence outcomes, we performed a single-institution, observational cohort study.

Methods
The cohort study was approved by the Roswell Park Comprehensive Cancer Center Institutional Review Board (EDR 171710). The Strengthening the Reporting of Observations Studies in Epidemiology (STROBE) reporting guideline was adhered to.
The patient database was selected from patients diagnosed with early-stage NSCLC (T1-2N0M0) who underwent 27 Gy in 1 fraction for peripheral tumors or 50 Gy/5 fractions for central tumors with heterogeneity correction at Roswell Park Comprehensive Cancer Center between September 2008 and December 2018. Tumor centrality was defined per RTOG 0813 as tumors within or touching the zone 2 cm around the proximal bronchial tree (PBT) or immediately adjacent to the mediastinal or pericardial pleura [11]. Candidacy for surgical resection was determined at the discretion of the thoracic surgeon.
All patients underwent SBRT as previously described [5]. Patients underwent CT simulation in the supine position with arms above their head using a thoracic Medical Intelligence BodyFIX ® immobilization system (Elekta, Stockholm, Sweden). Tumor motion management included either abdominal compression or respiratory gating, as previously described [12,13]. Dose delivery techniques employed included non-coplanar 3-dimensional conformal fields (3DCRT) or volumetric modulated arc therapy (VMAT). Heterogeneity corrections were used only for patients treated with intensity-modulated radiation therapy. Normal tissue dose constraints from RTOG 0915 were utilized for patients treated with 1 fraction and constraints from RTOG 0813 were used for patients treated with 5 fractions [2,11].
Clinically relevant variables, including age, gender, race, Karnofsky Performance Status (KPS), histology (adenocarcinoma, squamous cell carcinoma, NSCLC not otherwise specified), T stage, smoking status, operability, and tumor location were obtained retrospectively. Race was categorized as White, African American, Asian, American Indian/Alaska Native, unknown, or declined to answer. Toxicity data were unavailable. Due to the small sample size, non-White patients were grouped as a single category. There were no missing values in our database in such variables when analyses were performed.
Overall survival (OS) and progression-free survival (PFS) were the primary outcomes. OS was determined from time between the start of SBRT to the last known follow-up or death from any cause. PFS was determined from the time between the start of SBRT to any tumor recurrence or, if none, the last known follow-up or death. Secondary outcomes were local failure (LF), nodal failure (NF), and distant failure (DF). LF was defined per RTOG 0813 as local enlargement confirmed by positron emission tomography (PET) scan or biopsy, marginal failure, or involved lobe failure [11]. NF was defined as tumor recurrence in any thoracic nodal station and DF as any extra-thoracic or contralateral lung recurrence. All recurrences were evaluated in a multidisciplinary setting based on radiographic findings and, if available, biopsy results of metastatic sites.

Statistical Analysis
A Fisher exact test and Mann-Whitney U test were performed to compare baseline characteristics among the cohorts with different fractions. All categorical variables were handled as unordered. To evaluate OS and PFS, Cox multivariable analysis (MVA), Kaplan-Meier plotting, and a log-rank test were performed. Fine-Gray competing risk MVA was used to evaluate LF, NF, and DF with death as a competing event. All MVA models included previously listed clinically relevant variables.
To reduce selection bias, propensity score matching was performed between central and peripheral tumor cohorts. This technique attempts to balance treatment group confounding factors to make them comparable so conclusions can be drawn about the casual impact of a treatment on outcomes using observational data. Cohorts were matched based on the variables previously listed above. Matching was based on the nearest-neighbor method in a 1:1 ratio without replacement with a caliper distance of 0.2 [14]. After matching, Kaplan-Meier plot, Cox, and Fine-Gray regression models were performed to evaluate OS, PFS, LF, NF, and DF. All statistical tests were two-sided and p < 0.05 was considered statistically significant. Data analyses were conducted using R (version 4.1.2, R Project for Statistical Computing, Vienna, Austria).
After propensity score matching, 68 matched pairs were identified ( Table 4). The

Discussion
This study is the largest to report outcomes of central versus peripheral early-stage NSCLC cohorts treated with SBRT. We found no significant difference in OS or PFS between peripheral versus central tumors. When looking at secondary outcomes, we also observed no differences in LF, NF, or DF.
Two prospective randomized trials demonstrated single-fraction SBRT had the same local control, OS, and PFS outcomes as three-and four-fraction SBRT in peripheral earlystage NSCLC patients [2,12]. Retrospective analysis has shown that a single-fraction regimen is equivalent to five-fraction regimens for early-stage peripheral NSCLC [6]. Based on RTOG 0813, central early-stage NSCLC was treated with at least five fractions of SBRT at our institution [3,11].
Unfortunately, toxicity profiles among patients with central NSCLC tumors who underwent five-fraction SBRT were unavailable. High-grade toxicities have been reported among patients with such tumors [10,15,16]. In our study, central tumor SBRT outcomes were comparable to peripheral NSCLC tumors, consistent with prior retrospective studies [17,18]. These discrepancies may suggest a heterogeneous population of patients with central NSCLC tumors.
A further distinction exists for ultracentral early-stage NSCLC tumors, where the planning target volume (PTV) is abutting the proximal bronchial tree or touching mediastinal structures, such as the esophagus or great vessels. The Nordic HILUS trial reported a 18% grade 5 toxicity rate treating tumors ≤ 1 cm from the mainstem bronchus with SBRT [10]. A systematic review of SBRT for ultracentral lung lesions found treatment to be associated high local control rates but the potential for severe toxicity in patients receiving high doses to the proximal bronchial tree, those with endobronchial disease, and those receiving bevacizumab or anticoagulants around the time of SBRT [19]. In a matched cohort comparison between central and ultracentral lung tumors, a report found ultracentral tumor location correlated with reduced non-cancer-associated survival [20]. This could be due to increased cardiac toxicity from SBRT for ultracentral tumors, as another analysis of the same patient cohort found increased dose to the right atria to be associated with non-cancer-associated survival [21]. Despite these risks, another retrospective study found no difference in survival outcomes or toxicity between patients with central or ultracentral lung tumors [22].

Limitations
Our study has multiple limitations due to its retrospective nature. Without toxicity data for the cohort, it is unclear whether patients with central tumors had differences in treatment toxicity that contributed to survival outcomes. Additionally, the reasons for undergoing SBRT, such as significant medical comorbidities or refusing surgery, were unavailable for analysis. Although the performance status variable was well balanced among different fractionation schemes even prior to propensity score matching, it is notable that patients treated with single-fraction SBRT had a higher number of patients with a KPS < 80.

Conclusions
Survival and tumor recurrence outcomes of patients treated with SBRT for early-stage NSCLC were equivalent for central tumors and peripheral tumors.  Informed Consent Statement: Patient consent was waived because many subjects are no longer following-up at our institution or may be deceased, and attempts to contact subjects pose a greater risk than obtaining the data needed for this study.
Data Availability Statement: Ma and Singh had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. The data underlying this article cannot be shared publicly for the privacy of the individuals that participated in the study. Research data are stored in an institutional repository and will be shared upon request to the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.