In our series, we demonstrate favorable outcomes for MBM patients who undergo WBRT with a median OS of 7 months. This is much higher than expected for these patients, many of whom were offered WBRT as a last resort, after they had failed other options. Even with WBRT, survival for MBM patients has been exceedingly limited in the literature, with a recent retrospective series on 198 MBM patients documenting a median OS of 3.6 months with WBRT. [12] As a result, there has been a recent paradigm shift towards treating MBM with systemic therapies alone, such as concurrent ipilimumab/nivolumab, which have shown promising intracranial penetrance and survival outcomes in MBM patients. A recent phase II trial testing ipilimumab/nivolumab in 94 MBM patients showed a 12-month OS of 82% and an intracranial response rate of 56%. [13] However, it is important to note that only ECOG 0–1 patients with asymptomatic BM were included in this trial and that 76% of the patients had < 3 BM. Also, treatment toxicities were not insignificant, with a 55% rate of grade 3–4 adverse events, and one patient had a grade 5 adverse event. Given the pre-treatment characteristics of our patient population, which included 25% of patients with ECOG 2–3, 68% of patients with extracranial metastases, and 69% of patients with ≥ 5 brain metastases at the time of WBRT, our results indicate that WBRT still has potential to be a viable treatment option with comparatively minimal side effects, even for patients with unfavorable baseline characteristics.
There is limited contemporary data on WBRT outcomes in the MBM patient population. One retrospective series analyzed 61 MBM patients receiving WBRT with or without systemic therapies. In this study, where 92% of patients had ≥ 3 BM, patients were divided into cohorts depending on whether they received WBRT for newly diagnosed BM or intracranial disease progression. Both groups had a median OS of 3 months and the overall study population had a 59% rate of radiographically evident intracranial disease progression based on available post-WBRT MRI scans. [14] However, our series demonstrates more favorable outcomes with WBRT, with a median OS of 7 months, which did not differ significantly based on whether WBRT had been used for de novo or progressive BM. Six-month DFFS and LFFS were also 52% and 43%, respectively, indicating that the majority of our patients were experiencing more prolonged durations of CNS control with WBRT. These discrepant findings can be partially explained by the co-utilization of systemic therapies; in the Fuente et al. study, only 13% of patients received additional systemic therapy such as Ipilimumab or Temozolomide, compared to our patients, the majority of whom received systemic therapies. Although most individual systemic therapy options, except for BRAF inhibitors, were not noted to be prognostic in our analysis, there is likely some synergistic effect with multimodality combination treatment that contributes to improved patient outcomes. However, our study size may be too small to detect statistically significant improvements with the use of WBRT in the modern cohort of patients. There is already emerging data that suggest that BRAF inhibitors are a potent radiation sensitizer, and low dose radiotherapy can enhance T cell infiltration within the tumor micro-environment. [15] Thus, future trials should evaluate the role of BRAF inhibitors and emerging immunotherapy agents in combination with WBRT.
Another retrospective study by Rauschenberg et al. also reported on WBRT outcomes in 92 MBM patients undergoing radiation and/or systemic therapies. In this study, the WBRT population experienced a similar median OS of 7.1 months. [16] Patients undergoing WBRT had a mean number of 5 BM, and the majority also had extracranial metastases. The addition of anti-PD1, anti-CTLA4, and BRAF inhibitor +/- MEK inhibitor failed to significantly impact OS, although OS was notably above historical estimates of 2–4 months. [16] This was largely consistent with our study; except for combination therapy with BRAF inhibitors, which were given in 13% of patients, PFS and OS failed to significantly improve with the addition of other systemic therapy agents to the WBRT regimen. This could partially be explained by the limited number of patients receiving certain individual systemic therapy options, such as concurrent ipilimumab/nivolumab (3%).
The association between combination WBRT with BRAF inhibitors and improved outcomes in our study adds support to previous studies that identified BRAF mutant status as a positive prognostic factor. [17] A single-arm phase II trial of 172 MBM patients receiving Dabrafenib monotherapy resulted in an overall intracranial response rate of 35% in Val600Glu mutant patients with previously untreated or treated BM, suggesting good intracranial effect of this therapy. [18] Combining BRAF inhibitors with radiotherapy seems to further enhance this effect. In a retrospective pilot analysis of 12 MBM patients treated with Vemurafenib, the 3 patients who also received WBRT had either partial response (66%) or complete response (33%) intracranially. Their six-month OS of 92% was higher than what seen in our study, although SRS patients with limited BM were included in this analysis. The median number of BM in WBRT patients was 11 (range 6–12), suggesting that this response could be explained by the potential of BRAF inhibitors to serve as a radiation sensitizer. [19] This hypothesis has not fully been explored in the literature and warrants further investigation, especially in the era of more contemporary systemic therapy treatment options. Timing of WBRT and BRAF inhibitors was not explicitly stated in this study, but at our institution, it is our practice to pause BRAF inhibitors up to 3 days before WBRT to avoid skin toxicities such as cutis verticis gyrata, as per consensus guidelines. [20]
Our study also identified performance status as prognostic for OS and PFS. This is consistent with existing prognostic models such as the recursive partitioning analyses classes, disease-specific Graded Prognostic Assessment (ds-GPA), and melanoma marker GPA, which predict outcomes for patients with brain metastases. [21–23] However, other previously identified prognostic variables such as the number of brain metastases, age, or the presence of extracranial disease were not identified as prognostic in our analysis. This could be attributed to the limited size of our study population, changes in newer systemic agents, or more frequently timed brain MRI surveillance intervals allowing for earlier detection given the high proportion of patients with previously existing extracranial metastases.
Regarding WBRT-associated side effects, our patient series shows a 32% rate of post-treatment intralesional hemorrhage and a 16% rate of radionecrosis. With the development of multimodality treatment, there has been increasing concern over the possibility of a synergistically-motivated increase in radiation toxicities. However, the rate of radionecrosis in our patients, of whom > 50% also received systemic therapy, is consistent with the 4%-24% rate of radionecrosis reported after radiotherapy alone. [24, 25] Nevertheless, the limited follow-up duration of our study precludes a complete assessment of radionecrosis incidence, which continues to occur over 12 months after radiotherapy. [26] Our institutional rate of post-WBRT intralesional hemorrhage is also consistent with a retrospective study performed by Klein et al., which reported a 31% rate of post-radiosurgery intralesional hemorrhage. [27] Reassuringly, neither the development of radionecrosis nor hemorrhage contributed to worsened OS or PFS, although the presence of intralesional hemorrhage significantly decreased DFFS. This correlation could be related in part to limitations in radiographic sensitivity in discerning between blood products and disease progression.
Notably, post-WBRT memory deficits were reported in under 20% of our patients. Due to particular concern over WBRT and associated neurocognitive toxicities, recent studies are evaluating the feasibility of an SRS-based approach for increasing numbers of MBM; for example, a recent study on 143 patients with 10 + BM reported a 96.8% local control rate in those treated with upfront SRS. However, new BM developed in 81.2% of patients, suggesting that DFFS remains a major limitation to SRS. [28] Our study results indicate that WBRT, even in combination with systemic therapy, can produce an advantageous median DFFS of over 6 months without inducing widespread neurocognitive deficits. Recent advances to WBRT, such as hippocampal-sparing contouring techniques or concurrent administration of Memantine, have shown promise in further sparing neurocognitive toxicities. [29, 30] An ongoing phase III trial comparing survival outcomes and quality of life in patients with 5–15 BM randomized to SRS or hippocampal-avoidant WBRT with Memantine should provide further information on these interventions (NCT 03550391).
Finally, our outcomes suggest that systemic disease progression continues to play a major role in outcomes. Previous studies have shown that death from extracranial progression is a predominant cause of death in MBM patients receiving WBRT. [5] In our analysis, neither the number of BM nor the presence of intracranial progression prior to WBRT impacted OS or PFS. Also, median DFFS and LFFS were notably above the median PFS of 2.2 months seen in our study. Taken together, this suggests that WBRT is helping control intracranial disease whereas systemic disease control continues to be a problem. Thus, the value of WBRT should not be discounted, despite the emergence of newer systemic options for MBM.
Limitations of this study include the small study size and retrospective single-institution nature of the study, which increases the likelihood of biases in factors such as patient selection. Another significant limitation arose from the lack of available death date records on all patients, which led to the designation of the date of transition to hospice or hospice discussions with providers as the date of death for those with otherwise incomplete information. As a result, the median OS and PFS experienced by our study population may be longer than reported, which would increase the favorability of WBRT. Also, the presence of memory deficits was identified through a review of patient charts, and information was lacking as to whether such deficits were attributed to WBRT or intracranial progression since formal neurological testing was not performed. Nevertheless, our study presents compelling evidence that MBM patients with multiple adverse characteristics can still benefit from WBRT. WBRT also seems to play a beneficial role in the multimodality setting, and future studies can further elucidate optimal treatment strategies for these late-stage patients.