A matched-pair analysis comparing whole-brain radiotherapy with and without a stereotactic boost for intracerebral control and overall survival in patients with one to three cerebral metastases

Most cancer patients with one to three cerebral metastases are treated with radiotherapy alone. Several radiotherapy approaches exist including WBRT alone, SRS or FSRT alone, and WBRT plus a stereotactic boost (SRS or FSRT). Several randomized trials compared SRS alone to SRS plus WBRT [3‐6]. Aoyama et al. presented a trial of 132 patients with one to four cerebral metastases, 67 treated with SRS and 65 with SRS plus WBRT [3]. WBRT significantly improved IC at 12 months (53% vs. 24%, p < 0.001) but not OS (39% vs. 28%, p = 0.42). In 2009, Chang et al. reported a randomized trial of 58 patients with one to three cerebral metastases that was stopped after an interim analysis, since the decline in neurocognitive function at 4 months was significantly worse in the SRS plus WBRT group than in the SRS alone group (mean posterior probability of decline 52% vs. 24%) [4]. IC rates at 12 months were 73% after SRS plus WBRT and 27% after SRS alone (p < 0.001). Unfortunately, neurocognitive function was not assessed at 12 months. OS rates were not significantly different. In the third randomized trial comparing SRS plus WBRT to SRS alone, both SRS and neurosurgical resection were allowed and compared to the same procedures plus WBRT. In the subgroup of 199 patients treated with SRS, 100 patients received SRS alone and 99 patients SRS plus WBRT. WBRT significantly improved local control of the treated lesions (p = 0.040) and freedom from new cerebral metastases (p = 0.008) at 2 years. Similar to the previous two trials, OS was not significantly improved with WBRT. In the most recent trial of 213 patients with one to three cerebral metastases, cognitive progression at 3 months was significantly more severe in the SRS plus WBRT group. Time to intracranial failure was significantly shorter after SRS alone than after SRS plus WBRT (p < 0.001). Median OS times were 10.4 months vs. 7.4 months (p = 0.92). When summarizing these trials, it appears that WBRT in addition to SRS significantly improves IC but not OS [3‐6]. Unfortunately, WBRT also leads to more pronounced neurocognitive dysfunction. However, an intracerebral recurrence can be associated with impairment of neurocognitive function. Thus, the omission of WBRT may not be best in all situations. A recent meta-analysis of three trials [3‐5] suggested omitting WBRT in patients aged ≤50 years [13].

The risk of developing neurocognitive decline can be reduced by using doses per fraction considerably lower than 3 Gy and avoiding concomitant chemotherapy [14, 15]. Furthermore, hippocampus sparing WBRT offers an additional approach to reduce the risk of neurocognitive decline. In the study of Gondi et al., who used the Hopkins Verbal Learning Test - Revised for Delayed Recall (HVLT-R DR), the rates of decline at 4 months vs. baseline were 7% with hippocampal sparing and 30% in a historic control group (p < 0.001) [16]. In a randomized trial of 508 eligible patients, memantine in addition to WBRT (20 mg/d, starting 3 days prior to WBRT for 24 weeks) showed a strong trend towards less decline in delayed recall at 24 weeks when compared to WBRT without memantine (p = 0.059). Significantly better results were found for several endpoints including executive function at 16 weeks (p = 0.008), processing speed at 24 weeks (p = 0.014) and delayed recognition at 24 weeks (p = 0.015).

When using these new options to lower the risk of WBRT-induced neurocognitive decline, WBRT should still be considered an option for patients with one to three cerebral metastases. In this light, one important question is whether results of WBRT can be improved with a stereotactic boost? Two randomized trials were reported so far that compared WBRT with and without a stereotactic boost [10, 11]. One small trial of patients with two to four lesions was stopped after inclusion of 27 patients [10]. One-year local failure rates were 8% after WBRT plus SRS and 100% after WBRT alone (p = 0.002). Median OS times were not significantly different (11 vs. 7.5 months, p = 0.22). The other randomized trial included 331 eligible patients with one to three cerebral metastases and a KPS of ≥70 (RPA class 1 or 2) [11]. In this trial, SRS in addition to WBRT significantly improved local control of the treated lesions at 1 year (82% vs 71%, p = 0.01). IC, which was defined as increase in size of any lesion, new cerebral lesions or neurological deterioration despite stable disease on MRI, was also significantly improved (p = 0.013). However, IC rates were not explicitly stated. An OS benefit was limited to patients with a single cerebral metastasis. Median survival times were 6.5 and 4.9 months, respectively (p = 0.039), whereas in the entire cohort OS was not significantly different (6.5 vs 5.7 months, p = 0.14).

Similar to the previous trial (p = 0.013), the addition of a stereotactic boost significantly improved IC in the present study (p = 0.005) [11]. OS was not significantly affected in the entire series, which agreed well with the results of the two previous randomized studies [10, 11]. In the RTOG 9508 trial, an OS benefit was found in patients with a single cerebral metastasis. This result was not confirmed in the present study, although a certain trend was observed (p = 0.12). One may speculate whether the number of 120 patients with a single lesion in the present study was too small to provide adequate statistical power to detect a significant difference. The number of patients with a single lesion in the RTOG 9508 trial was greater (n = 186). Furthermore, although this matched-pair study followed very strict matching criteria (individual matching for nine factors) the source of data still is retrospective in nature. In case of retrospective data, the risk of a hidden selection bias always remains. One potential selection bias may have been introduced due to the difference regarding the size of the treated lesions that were slightly smaller in the WBRT plus SRS group.

When reflecting the results of the two randomized trials [10, 11] and the present study, it appears that a stereotactic boost in addition to WBRT significantly improves local and intracerebral control. Since an intracerebral recurrence may lead to severe symptoms and even death, a stereotactic boost can be recommended for patients with one to three cerebral metastases selected for WBRT. This recommendation is supported by other results of the RTOG 9508 trial [11]. The stereotactic boost led to improved functional autonomy in all patients and improved OS in patients with a single lesion. The recommendation applies to the majority of patients with few cerebral metastases. In a study that developed a score to predict the probability of developing new metastases after SRS alone, only 22% (47/214) of patients were considered good candidates SRS alone [17]. When adding a stereotactic boost to WBRT two additional aspects need to be considered. A stereotactic boost will likely be associated with an increased risk of radionecrosis, particularly when combined with modern targeted therapies such as ipilimumab that are increasingly used in cancer patients [18, 19]. Another aspect, which needs to be considered, is cost-effectiveness [18]. In a cost-effectiveness analysis published in 2015 that compared neuro-cognitive sparing radiotherapy programs to WBRT alone in patients with one to three brain metastases, hippocampal sparing WBRT plus a stereotactic boost proved to be cost-effective when compared to WBRT alone in patient groups with a median survival pf 12 months or longer following radiotherapy [20]. Moreover, WBRT plus a boost to the metatatic sites can be safely perfomed with volumetric modulated arc therapy (VMAT) in one course as WBRT plus a simultaneous integrated boost (SIB) [21].