The aim of this study was to evaluate the efficacy and safety of HF-SRT for patients with resected BM.
Brain metastases affect a significant percentage of cancer patients, and their prevalence is increasing as emerging systemic therapies keep improving long-term survival outcomes. Multiple treatment alternatives are available for patients with BM including surgery, WBRT, SRS, systemic therapy, or some combination of the above.
Surgery plays a vital role in the management of patients diagnosed with limited BM, improving overall survival, local control, and quality of life in patients with solitary brain mets [6]. In addition, surgery also offers the benefit of establishing a histological diagnosis, immediate mass reduction of large lesions and management of neurological symptoms refractory to steroids. Surgical resection alone, however, is associated with increased risk of local recurrence in the absence of adjuvant radiotherapy [16]. Although historically WBRT was a cornerstone therapy for patients with BM, new concerns regarding declines in neurocognitive toxicity and QOL led to the increasing use of SRS as a safer alternative in the definitive and postoperative setting [7, 17-18].
Brown et al [17] published a randomized trial comparing adjuvant SRS with WBRT for resected BM cases. At a median follow-up of 11 months, postoperative SRS showed no difference in OS compared to WBRT (median OS 3.2 months vs 3 months), with better 1-year cognitive deterioration free survival (median 3.7 months vs 3 months. Another randomized phase III trial compared gross total resection with or without SRS to surgical cavity in patients with 1–3 brain metastases [19]. 132 patients were randomized to receive either total resection of BM plus SRS boost (volume-based dosing, 12–16 Gy in a single fraction) or surgery alone. In this study, the addition of SRS led to a significant improvement in local control rate compared to observation (72% vs 43% at 12 months, respectively), while OS and neurological death were unchanged.
One significant disadvantage of postoperative SRS is difficulty of accurate delineation of the surgical cavity. This seems to translate into a relatively high rate of local recurrence, especially in large lesions compared with adjuvant WBRT [7, 19]. Indeed, Choi et al. demonstrated that a small PTV expansion of 2 mm improved local control without increasing toxicity [20].
Rationale for postoperative hypofractionated SRS
Currently, there is an ongoing debate whether single fraction SRS or HF-SRT should be the preferred radiation approach for patients with resected BMs. Potential advantages of hypofractionation include beneficial radiobiological effects (repair of normal tissues, reoxygenation and redistribution), reducing the risk of radiation necrosis (ref) and allowing for safe total dose escalation in candidates with large operative cavities or target volumes close to critical structures.
Overall, there is scarce literature comparing postoperative single vs hypofractionated stereotactic radiotherapy, although the published retrospective data suggests improved LC and lower risk of RN with HF-SRT versus single fraction SRS to risk cavity. A recent meta-analysis of over 3400 patients reported an improvement of 12-months LC and decreased risk of RN for patients treated with HF-SRT compared with single fraction SRS [21]. In addition, there is a phase III trial (Alliance A071801) currently randomizing patients with HF-SRT versus single fraction SRS to treat postoperative resection cavities.
A summary of selected studies of hypofractionated SRS for resected brain metastases is presented in Table 5. In our series, 12-month LC, DBF and RN rates were comparable to other HF-SRT studies reported in the literature. It was worth noting that the 12-month LC of 82.1% in our cohort exceeded the 12-months LC of the SRS arms of the two phase 3 randomised controlled trials, 72% in Mahajan et al. [19] and 61.8% in Brown et al [17], both of which employed single fraction SRS to the resection cavities.
The optimal dose fractionation protocol for HF-SRT for resected BM remains unknown. Kumar et al. reported that BED10 ≥48 Gy to the resection cavity is associated with improved LC for multifractionated SRS [39]. Meanwhile, Wiggenraad et al. published a systematic review recommended a BED12 of at least 40 Gy for local control rates above 70% [9]. In our trial, we achieved good LC result, despite only 52% patients (n=28) receiving at least 40 Gy BED12. One explanation is that our results reflect a lower median PTV volume of our series of 15.3cc which falls in the lower end of the range reported in Table 1.
Table 5: Summary of published HF-SRT studies
Reference
|
n
|
GTR
|
Total dose/ Fractionation
|
Margin (mm)
|
Median PTV/Range (cc)
|
1yr-LC
|
1yr-DBF
|
LMD
|
Median OS (months)
|
RN
|
Choi et al.[20]
|
112
|
90%
|
20 Gy/1-5#
|
0-2
|
11.1 (0.1-66.8)
|
90.5%
|
54%
|
|
17
|
3.6%5
|
Wang et al.[22]
|
37
|
|
24 Gy/3#
|
2-3
|
28.84 (11.07-81)
|
80%2
|
20%3
|
|
5.5
|
2.9%5
|
Steinmann et al.[23]
|
33
|
|
30-40 Gy/5-10#
|
4
|
25.59 (4.87-93.56)
|
71%
|
33%
|
|
20.2
|
0
|
Minniti et al.[24]
|
101
|
100%
|
27 Gy/3#
|
2
|
29.5 (18.5-52.7)
|
93%
|
50%
|
|
17
|
9%6
|
Brömme et al.[25]
|
42
|
83.3%
|
24-40 Gy/1-10#
|
3
|
22 (14–44)
|
77%
|
67%
|
4.7%
|
15.9
|
2.4%5
|
Al-Omair et al.[26]
|
20
|
80.9%
|
25-37.5 Gy/5#
|
2
|
34.5 (5-179.8)
|
71%
|
37%
|
|
23.6
|
0%
|
Ahmed et al.[27]
|
65
|
|
20-30 Gy/5#
|
1-2
|
16.88 (4.8-128.4)
|
87%
|
50.9%
|
9.2%
|
10.1
|
1.5%6
|
Ammirati et al.[28]
|
36
|
|
30 Gy/5#
|
3
|
20.1 (1.1-109.2)1
|
84%3
|
44%3
|
|
16
|
7.75
|
Ling et al.[29]
|
99
|
81%
|
10-28 Gy/1-5#
|
0-1
|
12.9 (0.6-51.1)
|
71.8%
|
64.1%
|
6%
|
12.7
|
9%5,6
|
Eaton et al.[30]
|
36
|
55.6%
|
21-30 Gy/3-5#
|
1-10
|
37.7 (12.7-80)
|
74.4%
|
|
|
11.2
|
13.8%6
|
Vogel et al.[31]
|
30
|
58%
|
15-35 Gy/1-5#
|
2-3
|
25.1 (4.7–90.9)
|
76%3
|
62%3
|
34%3
|
10.1
|
10%6
|
Dore et al.[32]
|
95
|
94.8%
|
23.1 Gy/3#
|
2
|
12.9 (0.8–64.7)
|
84%
|
56%
|
28%
|
25
|
20.6% 6
|
Lima et al.[33]
|
41
|
100%
|
20-30 Gy/5-10#
|
2
|
26.4 (14.1-38.4)
|
89.4%
|
10.4%
|
|
28.2
|
0%
|
Specht et al.[34]
|
46
|
63%
|
35 Gy/7#
|
2
|
26.19 (3.45-63.9)
|
88%
|
52%
|
|
25
|
0%
|
Pessina et al.[35]
|
69
|
100%
|
30 Gy/3#
|
3
|
55.2 (17.2-282.9)
|
100%
|
19.6%
|
|
24
|
0%
|
Abuodeh et al.[36]
|
75
|
94.8%
|
25 Gy/5#
|
1-2
|
13.8 (1.9-128.4)
|
88.8%
|
42.9% 3
|
2.7%
|
|
3%5,6
|
Keller et al.[37]
|
181
|
94.2%
|
33 Gy/3#
|
2
|
14.15 (0.8-65.8)
|
88%
|
61%
|
14.4%
|
17
|
18.5%5,6
|
Lockney et al.[38]
|
143
|
83.9%
|
30 Gy/5#
|
2-5
|
|
84%3
|
45*
|
|
5.4
|
11.8%
|
Kumar et al.[39]
|
39
|
100%
|
18-30 Gy/3-5#
|
|
|
77%2
|
7.7%3
|
13%
|
|
0
|
Navarria et al.[40]
|
101
|
64.4%
|
30 Gy/3#
|
3
|
52.9 (7.6-282.9)
|
98.9%
|
31.4%
|
13.9%
|
22
|
25.7%
|
Bilger et al.[41]
|
60
|
11.6%
|
30 Gy/5#
|
2
|
34.5 (3.6-307)
|
81.5%
|
55.8%3
|
|
15
|
0%
|
Martinage et al.[42]
|
160
|
92%
|
25-30 Gy/3-5#
|
2
|
15.2 (2.2-129.8)
|
88%
|
52%
|
20%
|
15.2
|
8.9%
|
Soliman et al.[43]
|
122
|
89%
|
25-35 Gy/5#
|
21
|
|
84%
|
45%
|
22%
|
17
|
6%5,6
|
Minniti et al.[44]
|
95
|
|
27 Gy/3#
|
1
|
22.4 (6.3-67.4)
|
83%
|
50%
|
7%
|
13.5
|
15%7
|
Eitz et al.[45]
|
558
|
46%
|
18-35 Gy/2-7#
|
|
23.9 (13.5-36.3)
|
84%
|
|
13.1%
|
21.2
|
8.6%6
|
Our Study
|
54
|
91%
|
21-30 Gy/1-5#
|
2
|
15.3 (3.7-87.2)
|
82.1%
|
25.9%
|
13.8%
|
14.93
|
7.4%
|
1: 5-10 mm expansion for dura contact 2: LC at 6 months 3: Crude rate 4: Only for postoperative fractionated stereotactic radiotherapy
5: RN confirmed by histology 6: RN radiological rate 7: 1-year cumulative incidence rates of RN
|
Toxicity
Brain radiation necrosis is one of the most concerning late adverse effects in the treatment of brain malignancies and can have a significant impact in patient's quality of life. RN usually occurs 6 months to 2 years after radiotherapy and risk factors include large target volume, increased fraction size, previous brain irradiation and concomitant use of systemic therapy/immunotherapy [46].
One of the main concerns of adjuvant SRS is that a PTV margin expansion around the surgical cavity can lead to higher rates of neurotoxicity. Although Choi et al reported that adding a 2 mm PTV expansion was not correlated with an increase in late toxicity [20], recent retrospective studies have linked margin expansion to worsening number of RN events [47, 48].
Postoperative HFSRS is one of the strategies to mitigate this problem, allowing DNA damage repair between fractions while delivering a smaller biological effective dose for normal brain tissue. However, a recent meta-analysis showed no significant reduction of RN rates with the use of adjuvant HF-SRT [49].
In our series, we observed a RN crude rate of 7.4% (n=4), which is in line with previously published results (As per table 5). This relative low rate of RN occurred despite 75% of our patients exceeded the recently published HyTEC threshold for brain including PTV (V20 and V24 <20cc for 3 and 5 fractions) [12]. Among the 4 patients with RN, 2 cases were symptomatic and required low dose steroid administration (n=2) and bevacizumab (n=1).