Multicenter analysis of stereotactic radiosurgery for multiple brain metastases from EGFR and ALK driven non-small cell lung cancer

Abstract

Introduction:

Patients with brain metastases (BrMs) arising from EGFR and ALK driven non-small cell lung cancer (NSCLC) have favorable prognoses and evolving treatment options. We evaluated multicenter outcomes for stereotactic radiosurgery (SRS) to multiple (≥4) BrMs, where randomized data remain limited.

Methods:

Data were collected retrospectively from 5 academic centers on EGFR and ALK NSCLC who received SRS to ≥4 BrMs with their first SRS treatment between 2008 and 2018. Analyzed endpoints included overall survival (OS), freedom from CNS progression (FFCNSP), and freedom from whole-brain radiotherapy (FFWBRT).

Results:

Eighty-nine patients (50 EGFR, 39 ALK) received a total of 159 SRS treatments to 1,080 BrMs, with a median follow up of 51.3 months. The median number of BrMs treated with SRS treatment-1 was 6 (range 4–26) and median for all treatments was 9 (range 4–47). Sixteen patients (18 %) had received WBRT prior to SRS treatment-1. The median OS was 24.2, 21.2, and 33.2 months for all patients, EGFR, and ALK subsets, respectively. After multivariable adjustment, only receipt of a next-generation tyrosine kinase inhibitor was associated with OS (HR 0.40, p = 0.005). No differences in OS were observed based on number of BrMs treated. The median FFCNSP was 9.4, 11.6, and 7.5 months, for all patients, EGFR, and ALK subsets, respectively. After multivariable adjustment, the number of BrMs (continuous) treated during treatment-1 was the only negative prognostic factor associated with FFCNSP (HR 1.071, p = 0.045). The 5-year FFWBRT was 73.6 %.

Conclusions:

This multicenter analysis over a >10-year period demonstrated favorable OS, FFCNSP, and FFWBRT, in patients with EGFR and ALK driven NSCLC receiving SRS to ≥4 BrMs. These data support SRS as an option in the upfront and salvage setting for higher burden CNS disease in this population.

1. Introduction

Stereotactic radiosurgery (SRS) has displaced whole-brain radiotherapy (WBRT) in the treatment of limited brain metastases (BrMs) after multiple randomized trials demonstrated similar overall survival (OS) and improved cognitive and quality of life (QOL) outcomes with SRS alone [1,2]. However, these trials only included patients with limited BrMs, defined as up to 3 or 4 lesions; the optimal management of multiple BrMs (≥4) remains controversial [2].

In patients with EGFR and ALK driven NSCLC, >50 % may develop BrMs during their disease course [3]. Due to their relatively favorable prognosis and expanding treatment options, these patients may uniquely benefit from multidisciplinary strategies that avoid the long-term sequelae of WBRT in favor of treatments utilizing a combination of SRS and tyrosine kinase inhibitors (TKIs) [3].

In this multi-institutional retrospective study, we report data over a >10-year period characterizing the outcomes for patients with EGFR and ALK driven NSCLC who were treated with SRS to ≥4 BrMs, where randomized data remain limited. The aim of this study was to provide a benchmark for SRS outcomes in EGFR and ALK driven NSCLC patients with multiple BrMs.

2. Methods

Following institutional review board (IRB) approval, data were retrospectively collected from the University of Colorado, Yale University, MD Anderson Cancer Center, University of Virginia, and Fox Chase Cancer Center. We included patients with EGFR and ALK NSCLC who had ≥4 BrMs treated in SRS treatment-1. Key endpoints included OS and freedom from CNS progression (FFCNSP) after SRS treatment-1. Because first SRS delivery could be given at varying times after BrM identification, OS from the time of BrM identification was also reported. Survival outcomes were estimated using the Kaplan-Meier (KM) method. Factors associated with OS and FFCNSP were analyzed with univariable and multivariable Cox proportional hazards models controlling for age (continuous), sex, driver oncogene, BrMs at metastatic diagnosis, prior WBRT, number of BrMs (continuously and dichotomously analyzed 4–10 vs >10) treated in SRS treatment-1, and receipt of a next-generation TKI (defined as osimertinib for EGFR and alectinib, brigatinib, lorlatinib, and ceritinib for ALK) following the diagnosis of BrM. Freedom from WBRT (FFWBRT) analyses were limited to patients who were WBRT naïve at SRS treatment-1. Analyses were conducted by the University of Colorado Cancer Center Biostatistics Core.

3. Results:

Eighty-nine patients were analyzed who received a total of 159 SRS treatments to 1,080 BrMs from 2008 to 2018. Fifty patients (56 %) had EGFR and 39 (44 %) had ALK driven NSCLC, with a median age of 58.5 years at SRS treatment-1 (Table 1). Forty-three patients (48 %) had synchronous BrMs at metastatic presentation. The median number of BrMs treated in treatment-1 was 6 (range 4–26). Forty-seven patients (53 %) had 1 SRS treatment, while 42 patients had 2–5 total treatments. The median number of BrMs treated in all SRS treatments was 9 (range 4–47). Sixteen patients (18 %) had received WBRT prior to SRS treatment-1. Thirty-eight patients (43 %) received a next-generation TKI after the diagnosis of BrM.

Table 1.

Patient characteristics.

Characteristic		Total = 89, N (%)
Age, years	Median (range)	58.5 (27.3–88.7)
Sex	Female	54 (60.7 %)
	Male	35 (39.3)
Driver oncogene	EGFR	50 (56.2 %)
	ALK	39 (43.8 %)
Brain metastases at metastatic presentation	Yes	43 (48.3 %)
	No	46 (51.7 %)
Number of brain metastases treated in SRS treatment-1	Median (range)	6 (4–26)
	4–10	76 (85.4 %)
	>10	13 (14.6 %)
Total number of SRS treatments	1	47 (52.8 %)
	2	24 (27.0 %)
	3	9 (10.1 %)
	4	8 (9.0 %)
	5	1 (1.1 %)
Total number of brain metastases irradiated in all SRS treatments	Median (range)	9 (4–47)
	4–10	51 (57.3 %)
	11–20	26 (29.2 %)
	>20	12 (13.5 %)
Receipt of any TKI after diagnosis of BrM	Yes	82 (92.1 %)
	No	7 (7.9 %)
Receipt of next-generation TKI after diagnosis	Yes	38 (42.7 %)
of BrM	No	51 (57.3 %)
WBRT	No WBRT	65 (73.0 %)
	WBRT before SRS	16 (18.0 %)
	WBRT after SRS	8 (9.0 %)
Vital status at last follow-up	Alive	29 (32.6 %)
	Deceased	60 (67.4 %)

Recorded data on tyrosine kinase inhibitors (TKIs) includes receipt after diagnosis of brain metastases (BrM) through last contact. Next-generation TKIs were defined as osimertinib for EGFR mutations and alectinib, brigatinib, loralatinib, and ceritinib for ALK rearrangements.

At last follow up, 60 patients (67 %) were deceased. From BrM diagnosis, the median OS for all patients, ALK, and EGFR subsets were 29.7 (95 % CI 24.5–43.4), 40.8 (95 % CI 28.8 to not reached), and 25.2 (95 % CI 19.7–39.6) months, respectively. The median follow up from SRS treatment-1 was 51.3 months. From SRS treatment-1, the median OS for all patients, ALK, and EGFR subsets were 24.2 (95 % CI 15.9–37.0), 33.2 (95 % CI 18.1 to not reached), and 21.2 (95 % CI 11.3–37) months, respectively. ALK positive status was associated with improved OS compared to EGFR on univariable analyses (Fig. 1A), but this was non-significant after multivariable adjustment. Receipt of a next-generation TKI following the diagnosis of BrM was the only factor associated with improved OS on multivariable analyses (Fig. 1C, Table 2). No differences in OS were observed by age, sex, number of BrMs treated by continuous or dichotomized (4–10 vs >10, Fig. 1C) analyses, BrMs at metastatic presentation, or prior WBRT.

Fig. 1.

Table 2.

Factors associated with overall survival (OS) after SRS treatment-1, Cox proportional hazards analysis.

Factor		Univariable				Multivariable
		HR	95 % CI lower	95 % CI upper	p-value	HR	95 % CI lower	95 % CI upper	p-value
Sex	Male	Reference				Reference
	Female	0.937	0.553	1.587	0.808	0.868	0.504	1.493	0.608
Driver oncogene	EGFR	Reference				Reference
	ALK	0.571	0.333	0.981	0.042	0.922	0.483	1.760	0.806
Age	Continuous	1.015	0.993	1.037	0.177	1.010	0.985	1.035	0.434
Brain mets at stage IV presentation	No	Reference				Reference
	Yes	1.062	0.636	1.773	0.819	0.952	0.534	1.697	0.867
No. of brain mets treated with SRS treatment-1	Continuous	1.023	0.952	1.099	0.532	1.021	0.941	1.107	0.617
WBRT prior to SRS	No	Reference				Reference
	Yes	1.052	0.557	1.988	0.876	1.281	0.610	2.689	0.512
Receipt of next-generation TKI	No	Reference				Reference
	Yes	0.375	0.213	0.659	<0.001	0.399	0.211	0.752	0.005

FFCNSP after SRS treatment-1 for all patients, ALK, and EGFR subsets were 9.4 (95 % CI 6.8–11.9), 7.5 (95 % CI 5.1–9.9), and 11.6 (95 % CI 8.1–15.1) months (Fig. 2 A). Median FFCNSP for SRS treatment-2 and treatment-3 were 9.3 (95 % CI 4.1–14.5) and 3.6 (95 % CI 0.5–6.8) months, respectively (Fig. 2D). The only negative prognostic factor associated with FFCNSP on univariable and multivariable analyses was the number of BrMs irradiated in treatment-1, analyzed as a continuous variable (Table 3).

Fig. 2.

Table 3.

Factors associated with freedom from CNS progression (FFCNSP) after SRS treatment-1, Cox proportional hazards analysis.

Factor		Univariable				Multivariable
		HR	95 % CI lower	95 % CI upper	p-value	HR	95 % CI lower	95 % CI upper	p-value
Sex	Male	Reference				Reference
	Female	0.927	0.551	1.561	0.777	0.975	0.557	1.706	0.930
Driver oncogene	EGFR	Reference				Reference
	ALK	1.114	0.665	1.866	0.682	0.964	0.502	1.852	0.912
Age	continuous	0.988	0.967	1.011	0.275	0.984	0.959	1.010	0.218
Brain mets at stage IV presentation	No	Reference				Reference
	Yes	0.897	0.536	1.500	0.679	0.936	0.526	1.667	0.823
No. of brain mets treated with SRS treatment-1	Continuous	1.071	1.003	1.144	0.042	1.071	1.001	1.145	0.045
WBRT prior to SRS	No	Reference				Reference
	Yes	1.100	0.582	2.078	0.770	1.038	0.507	2.126	0.918
Receipt of next-generation TKI	No	Reference				Reference
	Yes	0.941	0.563	1.572	0.816	0.884	0.496	1.574	0.674

Symptomatic radiation necrosis (RN) occurred in 10 patients, for a per patient rate of 11.2 % and a per lesions treated rate with SRS treatment-1 (667) of 1.5 %. Seven patients were managed medically, while 3 required surgery. One patient experienced grade-4 optic neuropathy and vision loss. Data on neurological mortality were available for 82 patients; 10 patients (12.2 %) were classified as “likely” dying of neurological causes and 4 patients (4.9 %) were classified as “possible,” due to active concurrent intracranial and extracranial disease. Sixty-nine patients (84.1 %) were classified as alive or having died of non-neurological causes. The 5-year FFWBRT was 73.6 % (Fig. 1D).

4. Discussion

Contemporary BrM guidelines recommend SRS for 1–4 lesions and recognize SRS as a potential option for patients with 5–10 lesions, where optimal therapy remains an area of active research. In EGFR and ALK driven NSCLC the CNS activity of newer generation TKIs has been increasingly acknowledged both in combination with CNS radiation and in emerging strategies involving upfront TKI alone in select patients with small asymptomatic BrMs. The prolonged OS measured in years and increasing systemic therapy options in EGFR and ALK NSCLC may magnify the benefits of strategies that avoid the long-term sequela of WBRT. Previously we reported our single-institution experience of SRS for ≥4 EGFR and ALK NSCLC BrMs [4], which demonstrated a favorable OS, comparable outcomes in patients regardless of the number of BrMs treated or SRS treatments delivered, and the since-replicated finding of favorable hippocampal and WBRT doses with SRS alone compared to WBRT (+/− hippocampal-avoidance) for multiple BrMs [5].

In this larger multicenter study, the OS, CNS disease control, and toxicity outcomes for SRS to ≥4 EGFR and ALK BrMs remain highly encouraging. Although we exclusively analyzed patients with ≥4 BrMs and, thus, a relatively high burden of CNS disease, the OS from BrM diagnosis of 40.8 (ALK) and 25.2 months (EGFR), are comparable to the literature for EGFR and ALK driven BrMs. A clinically important observation is the nuanced effect of treating an increasing number of BrMs, which was associated with shortened FFCSNP when analyzed as a continuous variable but did not affect OS. Despite the inclusion of the number of BrMs in prognostic calculators, [6] this series builds on recent observations which suggest that the number of BrMs may not be prognostic for OS in EGFR and ALK driven NSCLC [7]. This series also supports repeated SRS treatments, as the FFCNSP was similar between treatment-1 and treatment-2 (9.4 and 9.3 months), and FFWBRT remains relatively high at 74 % after 5 years. Reassuringly, rates of RN and neurological mortality were comparable to series including patients with fewer BrMs arising from various histologies [8].

The only factor associated with improved OS in this series was receipt of a next-generation TKI, underscoring the prognostic importance of improvements in targeted therapy. The CNS efficacy of newer TKIs has expanded multidisciplinary treatment options and raised the question of optimal administration with timing of SRS [3]. Overall, data from the era of first-generation TKIs suggest that early administration of radiation improves CNS control and, in some studies, OS [9,10], whereas the tradeoffs associated with upfront versus salvage SRS with next-generation TKIs with enhanced CNS activity are an open question [10] and the subject of ongoing trials (NCT03497767, NCT03769103).

To date, the highest level of fully published evidence to support SRS for >4 BrMs is a single-arm prospective trial of >1000 patients treated with SRS alone to 1–10 mixed tumor histology BrMs [8]. This trial demonstrated comparable OS, disease control, and neurological mortality outcomes with SRS for 2–4 vs 5–10 BrMs. More recently, a single-institution randomized trial of SRS vs WBRT in 72 patients with 4–15 non-melanomatous BrMs was presented in abstract form and reported improvement in cognitive preservation with SRS alone and no difference in OS [12]. A multicenter cooperative group trial of SRS vs hippocampal-avoidant WBRT for 4–15 BrMs, CCTG CE.7, is ongoing (NCT03550391). These and similar ongoing trials (NCT03075072, NCT03775330) highlight the substantial interest in studying SRS alone for increasing CNS disease burdens.

The strengths of this series include its multicenter design, relatively large patient numbers for this specific disease presentation (i.e., ALK/EGFR with ≥ 4 BrM treated) and SRS treatment paradigm, long follow up, and the tracking of multiple SRS treatments to numerous lesions. Limitations include the retrospective study design, absence of data on TKI timing relative to SRS, and lack of a comparison cohort of either upfront WBRT or upfront TKI alone. Evolving clinical questions regarding the optimal sequencing of TKIs and SRS, which are the subject of other dedicated analyses and ongoing clinical trials, are not addressed in this analysis. However, it is also relevant to note this study spans a >10-year period coinciding with the clinical introduction of first-generation TKIs, and the growing acceptance of upfront TKI monotherapy for BrMs represents a more recent development [3,11].

In conclusion, in a multicenter study spanning >10 years, SRS for multiple BrMs from EGFR and ALK driven NSCLC was associated with encouraging OS, FFCNSP, and FFWBRT. These data support upfront and salvage SRS in as an alternative to WBRT for this patient population.