A Comparison of Survival after Radiosurgery in Non-Small Cell Lung Cancer Patients with One versus More than Twenty Brain Metastases

Whether the number or cumulative volume of brain metastases affects survival in patients with metastatic non-small cell lung cancer (NSCLC) remains controversial. We sought to compare whether patients with solitary brain disease had better outcomes than patients with ≥ 20 brain metastases. Methods Between 2014 to 2020, 26 NSCLC patients (925 tumors) underwent stereotactic radiosurgery (SRS) for ≥ 20 metastases in a single procedure (median margin dose = 16 Gy, median cumulative tumor volume 56 patients underwent SRS for a single metastasis (median margin dose = volume = The overall survival (OS), local tumor control (LC), (ARE) risk, were


Introduction
Lung cancer is one of the most common malignancies in the world, accounting for 11.4% of all documented cancer cases and 18% of total deaths [1]. Non-small cell lung cancer (NSCLC) is the most common subtype and 30% -40% of NSCLC patients who are screened for brain spread at the time of the primary diagnosis are found to have Stage IV disease [2,3].
Historically, management approaches include both whole brain fractionated radiation therapy (WBRT) and surgical resection in patients who present with symptomatic but limited disease. Over the last two decades accumulating evidence showed that stereotactic radiosurgery (SRS) provides a non-invasive management option for many patients with brain metastatic spread, thereby sparing both the short and long term risks and failure rates of WBRT [4]. The local tumor control after SRS in NSCLC patients consistently exceeds 80% in the literature [5,6].
Debate continues as to whether the number or the cumulative volume of brain metastases (BM) is more important to determine patients' eligibility and tumor response to brain SRS. Current training biases, national guidelines, and third-party insurance reimbursement approvals often determine which option patients can receive [7]. The present study evaluates whether cumulative tumor volume or number (1 vs ≥ 20) affects survival (OS) after SRS. We also assessed local control (LC), adverse radiation effect (ARE) incidence, and new tumor development rates in this volume matched two cohort comparison.

Inclusion criteria and SRS treatment
The details of SRS procedure have been documented in the previous publications [8 -10]. Between 2014 and 2020, 82 NSCLC with either solitary brain metastasis or more than 20 brain metastases underwent SRS. Twenty-six patients with ≥ 20 BMs (cumulative tumor volume = 4.52 cc) were volume matched with 56 patients with solitary tumors (cumulative tumor volume = 4.74 cc). All patients had a Karnofsky performance Score (KPS) of ≥ 70. The histological sub-types of the primary lesion, extent of extracranial metastatic disease, previous brain surgery or WBRT, systemic treatment, and the number of BM were retrospectively reviewed from patients' medical record or radiographic imaging. On the day of SRS, mild sedation with intravenous midazolam and fentanyl along with local anesthetic were used for frame placement. High-resolution contrasted stereotactic magnetic resonance imaging (MRI) was performed. T1-weighted sequence with 1.5-mm slice thickness that was used for target localization and treatment planning. SRS was performed using the Leksell Gamma Knife (AB Elekta, Stockholm, Sweden) [8]. Tumor margin doses ranging from 16 to 20 Gy were prescribed based on tumor volume, location, and histology.
Lower doses were used for larger tumor volumes or in cases with adjacent critical structures such as the brainstem, optic nerves, or cochlea.

Patient Characteristics
For the 26 patients with ≥ 20 BMs, the median age was 64 years with females and males constituting 57% and 43%, respectively. The most common histological type was adenocarcinoma (77%). Nearly eight percent of the patients had received prior WBRT, and 15% of patient had prior craniotomy to remove one or more BMs. Eighty ve percent of patients received chemotherapy and 61.5% of patients received immunotherapy. The median number of metastases treated was 25, and the average total tumor volume per patient was 4.52 cc. During SRS a median margin doe of 16 Gy (range from 13 -20 Gy) was prescribed to a median of 80% isodose. The a median maximum dose received by each tumor of 20 Gy and the median total 12 Gy volume was 22.52 cc.
The median age for the 57 NSCLC patients with solitary tumors was 57 years, ranging from 44 -84 years old (Table 1), with 1:1 female to male ratio. Adenocarcinoma (65%) and squamous cell carcinoma (17.5%) represented the most common histological types. None of the patients received prior WBRT, and 28% of patient had prior surgical resection of tumor followed by tumor bed SRS. Most patients (83%) received chemotherapy and 17.5% received immunotherapy. The median tumor volume was 4.74 cc (ranging from 0.62 cc to 29.65 cc). During SRS the median 12 Gy volume was 13.41 cc and the median maximum dose received by each tumor was 35.5 Gy.  [11] or surgical pathological con rmation of radiation necrosis. Suspicious but indeterminate contrast enhancements (tumor recurrence vs ARE) were followed up with MRI until local recurrence or ARE could be discerned.

Statistical Analyses
Statistical analyses were performed with Prism version 9 (GraphPad, California). Survival, local control, regional control, and adverse radiation effect curves were generated using the Kaplan-Meier method. A p value of < 0.05 was necessary to be considered statistically signi cant. Patients with solitary BM whose volume was less than 0.5 cc were excluded from this study to control the total tumor volume between the two patient groups.

Survival Rates
The median OS for patients with solitary BM was 12 months. By comparison, the median survival for patients with ≥ 20 BMs was 15 months. No statistically signi cant difference in survival rates was observed (p = 0.3) between the two cohorts (Fig. 1).

Discussion
SRS is a safe and effective management modality for the CNS metastases from NSCLC. Bowden et al. reported that the median survival of NSCLC patients with brain metastasis was 19.9 months, with 1-year survivals of 70% and 5-year survival rates of 16% [9]. The initial management of patients with both solitary and multiple management of metastatic NSCLC continues to evolve, despite past guidelines that only patients with 1-4 brain metastases are considered appropriate for initial treatment by SRS. In the current study, we sought to determine whether tumor number or total tumor volume can serve as a better predictive factor for patients' survival after SRS.
Although high de nition MRI can determine the number of brain metastases, simple tumor number is insu cient to determine the best strategy. At present neither medical oncologists, radiation oncologists, or neurosurgeons are able to provide accurate predictions of survival when confronted with a new patient with metastatic brain disease [12]. Routman et al. [13] found that the number of brain metastases was not a signi cant predictor of patient survival for patients who underwent SRS with or without prior WBRT and/or surgery. Chang et al. [14] reviewed 323 brain metastasis patients who underwent SRS and found no signi cant difference in survivals among patient groups that had 1-5, 6-10, 11-15, and ≥ 15 BMs.
Prior to the current analysis, we hypothesized that tumor volume was more prognostically important than the number of brain metastases. We found that the median survival in patients with ≥ 20 metastases were statistically no different than the survival of patients with single metastases (p = 0.3). Our experience compares favorably with other outcome reports in current management of NSCLC patients. In 2013, Rava et al. [15] reported a median survival of 6.5 months in 53 patients with ≥ 10 metastatic brain lesions. Chang and colleagues [14] reported that the median survival times for patients with 11-15 and ≥ 15 BMs was 13 months and 8 months, respectively. Karlsson et al. [16] analyzed 1921 metastatic patients who underwent SRS over the time span of 30 years and found no statistically signi cant survival differences among individuals with 2, 3-4, 5-8, or ≥ 8 metastases.
The model constructed by Ali et al. [17] found brain metastasis number as a continuous variable and only a 4% increase in the hazard of death for every increment of 6-7 metastases. This study showed that the number of brain metastases only serves as a modest prognostic factor in predicting patient survival and treatment recommendations should be evaluated using other factors when trying to decide between in the greater clinical context in the decision making between SRS and WBRT. While tumor number is often evaluated, cumulative tumor volume is rarely included in guidance documents [18].
In the context of active systemic cancer, some patients will experience new tumor development, progression of treated disease, or treatment related side effects. The present study showed a median onset of 8 months until the development of new tumors in patients with initially solitary tumors. In comparison, the interval between SRS and new tumor detection in the ≥ 20 cohort was slightly less at 8 months. Somewhat surprisingly we found that median survivals were slightly longer in the ≥ 20 cohort, although this nding was not statistically signi cant. SRS is an effective management option for patients with 20 or more BMs. Overall survival of patients with 20 or more metastases was no different than survivals in NSCLC patients who had with single BM if the cumulative tumor volume was controlled total tumor volume. This suggests that total tumor volume may be more predictive of survival than number of brain metastases.

Declarations
Funding: The authors declare that no funds, grants, or other support were received during the preparation of this manuscript. Consent to publish: The authors a rm that human research participants' consent for publication of the images in Figure(s) 1 (a-d), and 2 (a-d) was not needed as the data is anonymized.