Adjuvant WBRT has been considered the standard of care after surgery for many decades, with the rationale of improving both local surgical cavity and distant intracranial control. More recently, the advent of highly sensitive neuroimaging, such as multi-parametric brain MRI, made it reasonable to omit adjuvant WBRT, and its related early (hair loss and fatigue) and late (neurocognitive deterioration) complications, when a complete resection is achieved.

Some multicenter RCTs were conducted to investigate the role of WBRT in this setting.

The first one by Patchell and colleagues (20) enrolled 95 patients (60% with a primary diagnosis of NSCLC), receiving a complete-surgical resection of a single BM. WBRT resulted in better intracranial tumor control compared to observation (anywhere recurrence: 18% vs 70%, p<0.001). Furthermore, WBRT determined less neurological deaths (14% adjuvant WBRT vs 44% surgery alone, p=0.03) but no significant differences in functional independence and OS (approximately 10 months in both arms) were observed.

The EORTC 22952-26001 study investigated the role of adjuvant WBRT after surgery or SRS for 1-3 BMs, randomizing patients to WBRT or observation: 359 patients (NSCLC was the primary tumor in 53%) with a limited number of BMs (1–3) were randomized to receive WBRT or to close observation after local therapy, either surgery or SRS. In the surgery group (160 patients) almost all the patients had a single (96%), large (median size 30 mm) metastasis. For these patients, postoperative WBRT significantly reduced the probability of local relapse by 32% (from 59% to 27%, p <0.001) and the probability of distant intracranial relapse by 19% (from 42% to 23%, p=0.001). Again, no difference was found in functional independence and OS (19).

More recently, the Japanese JCOG0504 non-inferiority trial randomly assigned to WBRT or observation and salvage SRS 271 patients (47% diagnosed with NSCLC) with 4 or fewer surgically resected BMs (73% single lesion).

Observation with salvage SRS was not inferior to WBRT in terms of OS (median OS: 15.6 months in both arms, p=0.027), although median intracranial Progression Free Survival (PFS) was shorter (10.4 months for WBRT vs 4 months for salvage SRS). In addition, a deferred brain irradiation with a highly conformal technique resulted in more than half reduction of severe cognitive deterioration for salvage SRS patients (G2-G4 cognitive dysfunction: 16.4% WBRT vs 7.7% salvage SRS) (21).

Conclusions: WBRT is usually avoided as adjuvant treatment, considering the absence of survival benefit and the severe impact on neurocognitive functions. Anyway, it remains the best strategy to prevent the onset of new BMs.

Given the suboptimal LC of surgery alone, with an unacceptable high risk of local recurrence (around 50%) (19, 20), and the inescapable severe neurocognitive sequelae of whole brain irradiation, the modern trend is to offer focal postoperative RT to significantly improve the remission rate of the surgical cavity. In the last two decades, several retrospective studies investigated the role of adjuvant SRS to the surgical cavity and a recent meta-analysis, including 3458 patients from 50 studies, showed high LC (1-y LC: 83.7%) and low toxicity profile (RN rate: 6.9%) (37).

Two randomized trials confirmed the efficacy and safety of this treatment. In the study by Mahajan and colleagues (38), the 12-months freedom from local recurrence was 43% in the observation group and 72% in the SRS group (p=0.015), but with no difference in OS. In the NCCTG N107C/CEC.3 multicenter trial 194 patients (59% NSCLC) were randomly assigned to adjuvant SRS or WBRT, with the latter one providing a better time to intracranial tumor progression (median time: 27.5 months vs 6.4 months, p<0.0001), considering both local and distant recurrence. On the other hand, neurocognitive preservation was significantly higher in the SRS group (median cognitive-deterioration-free survival: 3.7 months vs 3 months, p<0.0001) with significant difference in terms of immediate and delayed memory, processing speed and executive function (39). As predictable, OS was comparable between the two arms (12.2 months for SRS vs 11.6 months for WBRT, p=0.70).

Further evidence from the ESTRON German trial (NCT03285932), presenting the same interventional arms, are awaited to provide more robust evidence on the role of SRS in the adjuvant setting after the resection of BMs (40). Although the availability of high quality data in support of the clinical role of adjuvant SRS to the surgical cavity, some practical and technical challenges remain to be solved. First, the optimal dose and fractionation to obtain the ideal balance between LC and risk of RN. Considering that surgery is generally performed for large, symptomatic lesions, it is very common for the treating radiation oncologist to face with large post-operative volumes to be irradiated. The considerations for cavities greater than 2-3 cm in diameter are the same as for large unresected BMs, with SRT representing the best solution. To date, several retrospective and meta-analysis data are available and confirm an excellent risk-benefit balance of focal SRT to the resection cavity (41–45), with a trend for better LC and lower rates of RN compared to SRS (29, 37). The prescribed dose of SRT usually ranges 24-30 Gy in 3-5 fractions in the published series (41, 46). Based on the available data, the HYTEC TCP and NTCP papers provide useful indications on doses and constraints to orientate in the clinical practice, as previously done for the radical and upfront setting (30, 31). The validation of these results in a randomized setting are awaited from the ongoing Alliance trial (NCT04114981), which is randomizing ≥2 cm-sized-operated brain lesions with post-surgical cavities smaller than 5 cm to be irradiated with single fraction SRS or 3-5 fractions SRT (Table 1).

The second issue is related to the uncertainties in the target delineation. The dynamic adaptation of the surgical cavity and of the surrounding tissues after surgical resection, cause significant changes in the shape of the target area, with an average cavity/volume reduction after surgery estimated in a range of 15-43% (47). This volume shrinkage and the timing needed to obtain a reasonable post-surgical clinical recovery, influence the timing for the administrations of adjuvant RT, generally performed within maximum 4-6 weeks after surgery.

In 2017 a consensus contouring guideline for adjuvant SRS in completely resected BM cavity was generated, with the aim of standardizing this highly variable procedure. This paper recommended to use pre- and post-operative contrast-enhancing T1-weighted MRI to guide the contouring process and to include in the clinical target volume (CTV) the entire surgical cavity as well as the surgical tract, with the inclusion of any site of preoperative dural or venous sinus involvement (48).

The last important consideration on post-surgical SRS/SRT is about the risk of leptomeningeal dissemination (LMD). Surgical procedures may disseminate cells in surrounding tissues causing the neoplastic seeding of the cerebrospinal fluid, and focal adjuvant RT seems to be less effective than whole-brain irradiation to prevent this complication. In retrospective series LMD incidence ranges from 8 to 13% (49, 50). In the “Leptomeningeal Metastasis” section this scenario will be further explored.

Conclusions: focal RT is currently the preferred adjuvant treatment after surgical resection of BM. The excellent LC and a good toxicity profile make this strategy preferable to WBRT. The ideal timing, doses, fractionations and volumes for post-operative stereotactic irradiation are currently under investigation.

The new frontier in the treatment of single or few BMs is preoperative SRS. The rational is to sterilize any microscopic disease before the macroscopic resection of a brain lesion and to prevent all critical issues related to adjuvant SRS. Generally, neoadjuvant RT is performed as a single fraction of 16 Gy, given a few days before surgery.

Neoadjuvant SRS allows a better definition of the target, with a potentially reduced risk of LMD (sterilizing microscopic disease) and RN compared to the adjuvant setting. Figures 2, 3 illustrate volume and dose-volume histogram comparison between neoadjuvant and adjuvant focal RT.

Several single arm studies have collected retrospective and prospective data on neoadjuvant SRS, showing interesting efficacy and toxicity profile (including low rates of RN and LMD), but the small sample sizes are still inadequate to provide a robust evidence (51–56).

The recent PROPS-BM, a multicenter cohort study from 5 American institutions, represents the largest series of patients treated with neoadjuvant SRS to date. The outcomes of 242 patients (43.4% diagnosed with NSCLC) with 253 lesions (62.4% had a single BM) were analyzed. With a median time between SRS and surgery of 1 day (range 1-3 days), the median prescribed dose at the 80% isodose line was 15 Gy to a median GTV of 9.9 cm³. The treatment was effective (1-year LC: 85%; median survival time 16.9 months; LMD rate: 7.9%) and safe (postoperative complications: 7%; RN rate: 7.1%), with subtotal resection as strong independent predictor for local recurrence (57).

Ongoing studies, particularly the phase III trial from MD Anderson Cancer Centre (MDACC) (NCT03741673), Mayo Clinic (NCT03750227) and Alberta (NCT04474925), will provide further information on this promising approach (Table 1).

Conclusions: For resectable BMs neoadjuvant SRS is a promising approach, with potential advantages if compared to adjuvant RT. Further evidence is needed for a greater implementation of this treatment into clinical practice.

The demonstrated local efficacy of focal SRS has raised the question on the feasibility of omitting WBRT after SRS not only in patients with a single lesion but also in those with a limited number (2–4) of BMs, in order to keep WBRT as a salvage option and to delay or possibly avoid neurocognitive dysfunctions.

Four historical randomized trials and an individual patient data meta-analysis investigated this scenario with different endpoints (16–19, 22).

Overall, WBRT demonstrated an increase in local and distant intracranial control (by approximately 15-30% and 50% respectively), but without any survival benefit and at the price of a worsening in neurocognitive functions and QoL. These results generated a new trend to omit WBRT also in patients with few BMs, despite some controversial aspects.

In fact, it is known that intracranial relapse could be the primary cause of death in these patients, and WBRT has proved to be an effective strategy to achieve both local and distant intracranial control. At the same time, the progression of metastatic brain disease involving critical areas can cause rapid neurocognitive deterioration. The correct patient stratification, according to modern prognostic scores, could help to identify those patients that could benefit from the combination of WBRT and SRS in this peculiar setting.

With this purpose, 3 secondary analyses of the JROSG (58), EORTC (59) and Alliance (60) trials were conducted, post-stratifying the NSCLC populations according to the Disease Specific-GPA (DS-GPA). Among the 88 NSCLC patients of the Japanese study, better OS was observed with SRS + WBRT in the 47 with DS-GPA 2.5-4.0 (WBRT+SRS: 16.7 months vs SRS: 10.6 months, p = 0.04). No difference was observed in the unfavorable DS-GPA group (0.5-2.0), probably due to a less significant impact of intracranial control on patient’s prognosis (HR: 3.57 p =0.04 vs HR: 8.31 P < 0.001) (58). By contrast, the other two exploratory analyses of the EORTC and Alliance trials, did not demonstrate a survival benefit in favor of the addition of WBRT in any risk category according to DS-GPA (59, 60). Table 2 resumes these findings.

Conclusions: SRS alone could be safely administered to patients with 2-4 BMs. The omission of WBRT preserves the neurocognitive functions without compromising survival in the overall NSCLC population. Some secondary analyses suggest a potential survival advantage combining WBRT and SRS in NSCLC patients with 2-4 brain lesions and a favorable GPA score.

Historically, systemic therapy played a secondary role in the management of BMs, considering the CNS as a “sanctuary site” for traditional chemotherapy (ChT) due to the blood brain barrier.

In NSCLC, the development of immune checkpoint inhibitors (ICIs) and targeted agents [particularly Epidermal Growth Factor Receptor (EGFR) and Anaplastic Lymphoma Kinase (ALK) Tyrosine Kinase Inhibitors (TKIs)], redesigned this scenario, for the ability of these novel drugs to permeate in the CNS. The new generation of targeted agents, such as osimertinib for EGFR-mutated (61, 62) and alectinib or lorlatinib for ALK-rearranged (63, 64) NSCLC, significantly improved intracranial response rate compared to traditional ChT or first-generation TKIs, with responses durable over time.

Today, these drugs may be integrated with focal therapies in a multimodal approach. The best sequence of treatment is not clear, and the choice must be individualized according to different clinical aspects, such as the time of BMs onset (synchronous vs metachronous), the time to previous treatments for oligoprogressive or oligorecurrent disease and the presence of neurologic symptoms. In patients with NSCLC and synchronous brain oligometastases, particularly with asymptomatic presentations, upfront systemic therapy is one of the most adopted strategy, deferring focal RT upon evaluation of clinical and radiological response.

In 2015 a Korean phase III trial evaluated the role of upfront ChT in de-novo oligometastatic disease from NSCLC. This study randomized 105 patients with 1-4 asymptomatic BMs to receive SRS + ChT vs ChT alone. Upfront ChT alone resulted in good response rates (37%) with no difference in intracranial control (median time: SRS 9.4 months vs ChT 6.6 months, p=0.248) and OS (SRS 14.6 months vs ChT 15.3 months, p=0.418) between the two arms (65). Similar studies with the use of new generation TKIs are expected to even improve these outcomes.

Another solution is the combined administration of SRS and targeted treatments. Data from retrospective series in oncogene-addicted NSCLC, preliminarily showed promising results in favor of the synergistic effect of this regimen in improving the intracranial disease control (66, 67) and this approach is under investigation in ongoing RCTs.

In patients with oligorecurrent or oligoprogressive NSCLC presenting metachronous BMs, the focal approach with SRS is even helpful to delay the start of a systemic treatment or the switch to a new regimen (66).

Conclusions: Upfront systemic therapy with the delay of RT administration is a viable option for asymptomatic synchronous brain oligometastases, particularly interesting in oncogene-addicted NSCLC. Conversely, focal RT can be the way to delay systemic treatment in case of metachronous oligorecurrence or oligoprogression to the brain.

The irradiation of the brain is associated with a dose-dependent radiation-induced leukoencephalopathy as a result of demyelination, vascular compromise and direct damage to neurons (78, 79). Patients affected by leukoencephalopathy may develop some degree of cognitive dysfunction which can compromise the QoL and affect memory, executive function, attention and concentration, and could lead to learning disorders and dementia (80). This entity is associated with diffuse supra-tentorial white matter abnormalities and cerebral atrophy. Apart from leukoencephalopathy, cranial irradiation can damage the hippocampus which has a fundamental role in the memory function.

The use of a tumor-selective agents that enhances the effects of radiation in tumors but spares normal brain tissue might extend the therapeutic ratio of WBRT, improving LC without increasing radiation toxicity.Certain agents target tumors selectively, generating reactive oxygen species intracellularly and lowering the apoptotic threshold to radiation and chemotherapy (81).

One of the most accredited theories suggests vascular damage and mineralizing microangiopathy, with subsequent small vessel insufficiency and infarction as it is seen in vascular dementia, as the leading cause of RT-related neurotoxicity (82, 83). Therefore, neurotransmitter regulators commonly used to treat vascular dementia such as memantine have also been taken into consideration. In the RTOG 0614, a placebo double-blind RCT, the use of daily memantine for 24 weeks during and after WBRT resulted in better cognitive function over time, delayed time to cognitive and memory decline, executive function, and processing speed compared to placebo (84).

A recent drug, donepezil (an acetylcholinesterase inhibitor), is being tested in a phase 3 study among long-term adult brain tumor survivors after a course of fractionated WBRT or PCI, with the aim of improving cognitive impairment associated with brain cancer and its treatments. Despite modest improvements in several key cognitive functions, especially among patients with greater pre-treatment cognitive impairment, treatments with donepezil did not significantly improve the overall composite score (85).

The addition of Motexafin Gadolinium (MGd) to WBRT did not produce a significant overall improvement compared to WBRT alone (86). Anyway, in the intent-to-treat analysis, MGd exhibited a favourable trend in neurologic outcomes, significantly prolonging the interval to neurologic progression in NSCLC patients with BMs receiving prompt WBRT (87). Other molecules have been tested but promising results have been obtained for other histologies (88) and not for NSCLC.

Preliminary studies have demonstrated that BMX-001 provides protection of normal tissues from radiation-induced injury. Ongoing clinical trial (NCT03608020) will provide information on safety, tolerability and neurocognitive preservation of this drug.

Conclusions: Of the several molecules available to limit neurocognitive toxicity, only memantine prolongs time to neurocognitive and neurologic progression in patients with BMs from NSCLC.

As anticipated, the hippocampus has an important role in learning, memory and mood regulation (89). Preclinical studies have demonstrated that relatively modest doses of radiation cause an early and significant decline in neurogenesis in the subgranular zone of hippocampi associated with suppression of new memory formation and impaired recall (90). In addition, recent clinical studies have observed a dose-response related risk of postradiotherapy decline in learning delayed recall caused by the dose of radiation absorbed by the hippocampus (91, 92). More in general, cranial irradiation plays a crucial role for memory decline. Memory function, specifically recall and delayed recall, as assessed with the Hopkins Verbal Learning Test Revised (HVLT- R), have a statistically significant neurocognitive decline at 4 and 6 months from WBRT (17), with a simultaneous impairment of patient-reported QoL (93).

Recent technological improvements in radiotherapy, including helical tomotherapy, LINAC-based IMRT or volumetric-modulated arc therapy (VMAT), may be adopted to selectively spare anatomical structures involved in memory and learning during cranial irradiation (Figure 4) (92, 94).

Gondi et al. first proved that HA-WBRT is able to reduce both maximum and mean dose per fraction delivered to the hippocampus (94). For a prescription dose of 30 Gy in 10 fractions to the whole brain, HA-WBRT is able to reduce the mean dose per fraction to the hippocampus (normalized to 2-Gy fractions) by 87% using helical tomotherapy and by 81% using LINAC-based IMRT. The maximum dose received by the hippocampus is 12.8 Gy (Dmean 5.5 Gy) using helical tomotherapy, and 15.3 Gy (Dmean 7.8 Gy) using LINAC based IMRT, with acceptable target coverage and homogeneity (94) (Figure 5).

On the basis of these feasibility analyses, randomized prospective trials were designed to prove the efficacy of HA-WBRT as a viable option for polymetastatic disease. The single-arm phase II RTOG 0933 trial on HA-WBRT for BMs reported significant memory preservation (assessed through reduction of HVLT-R Delayed Recall decline), compared with historical data of patients treated with standard WBRT (92). Further confirmation of the effectiveness of HA-WBRT came from a prospective randomized phase III trial (NRG-CC001) which evaluated the potential combined neuroprotective effects of hippocampal avoidance in addition to prophylactic memantine during WBRT. HA-WBRT + memantine prolonged time to neurocognitive failure; decrease in neurocognitive function at 6 months was 59.5% vs. 68.2% (HR: 0.76, p = 0.03) favouring the combination of HA-WBRT + memantine, without any difference in intracranial recurrences (p = 0.208) or OS (p = 0.307) (95).

Ongoing research on the pathophysiology of brain irradiation damage may reveal other possible important brain structures or hippocampal subregions (e.g. cornu ammonis) whose sparing could also contribute to better preservation of neurocognitive function (NCT04801342, NCT03223922).

It should be noted that sparing the hippocampus and the peri-hippocampal region determines the theoretical risk of intra-cranial disease progression/relapse in these anatomical regions. However, it is estimated that this volume accounts for approximately 2% of the whole brain and the incidence of the development of metastases in this area is an uncommon event with a detection rate of less than 10% in previous reports (94).

Conclusions: HA-WBRT is an alternative solution and should be considered in selected patients with good KPS to better preserve cognitive function and patient-reported symptoms. The combination with memantine proved to improve the neurocognitive endpoints in a RTOG trial and is now approved in the United States in combination with HA-WBRT.

Although current evidence confirms WBRT as the standard treatment for patients with multiple BMs (75), focal irradiation is increasingly used in daily clinical practice and many institutions extend its utilization beyond the oligometastatic setting, mostly in fear of the neurocognitive side effects of WBRT. It remains controversial which subpopulation of multiple BM patients may obtain the greater benefit from local treatments like SRS. In published series, predictive factors for LC comprise delivered dose, total volume of treated metastasis, and histology of the primary tumor (96–100) without a clear correlation with the total number of BMs. To date, only two prospective studies have been published. Thus, clinical data is currently quite limited. The multicenter phase 2 study by the Japanese Leksell Gamma Knife Society (JLGK0901) reported the outcome of patients treated with SRS alone upfront, regardless of the number of BMs, and reported the significant inferiority in OS for patients with 5-10 BMs (largest tumor <10 mL in volume and <3 cm in longest diameter; total cumulative volume ≤15 mL), compared with patients with only 2-4 BMs, although the difference in median survival time among the two groups was not clinically meningful (7.0 vs 7.9 respectively). On the other hand, cumulative incidences of neurological death at 6, 12, and 24 months after SRS did not differ significantly between patients with 2-4 BMs, and those with 5-10 BMs, nor did the cumulative incidence of neurological deterioration after SRS (11.6% vs 13%, p=0.54). At 12 months after SRS, neurocognitive function was maintained in 91% of patients with 2-4 BMs and 88% of patients with 5-10 BMs (p=0.60) (101).

Subsequently, Yamamoto et al. conducted a dedicated case-matched study comparing SRS in patients with 2–9 BMs and in those with 10 or more BMs. In this study, median survival time for the two groups was not significantly different, likewise neurological death-free survival, cumulative incidence of local recurrence, use of salvage SRS for new lesions, neurological deterioration and SRS-related complications. They concluded that even patients with 10 or more BMs may be suitable candidates for SRS after careful selection (e.g., low intracranial tumor burden) (100). A multicentre, single-arm, phase 2 study by Nichol et al. also reported the effectiveness and tolerability of SRS for patients with 1 to 10 BMs (102).

Other retrospective studies have confirmed SRS as appropriate in patients with polymetastatic disease with LC rate and toxicity comparable to those observed in patients with a limited number of BMs (103–105).

Recently, the first phase III randomized controlled trial (Dutch Trial) investigating WBRT versus SRS for patients with 4–10 BM suggested that SRS is a valid palliative treatment option for patients with polymetastatic disease for the reduced incidence of toxicity and for the preservation of QoL. Anyway, the trial was closed prematurely, mainly as a result of patients’ and doctors’ preference for SRS, and no definitive results were produced on the non-inferiority of SRS in term of OS and brain failure-free survival (106).

Ongoing phase III trials directly comparing WBRT to SRS in patients with multiple BMs will hopefully bring more relevant data and evidence to guide the treatment selection.

A randomized phase III trial at the MDACC has randomized 72 patients with 4–15 BMs from a non-melanoma primary tumor to SRS (15-24 Gy) or WBRT (30 Gy) (107). The results presented at the annual ASTRO meeting in 2020 showed similar LC, new onset of BMs and OS (approximately 8 months) among the two cohorts, while patients receiving SRS had a shorter time to systemic therapies (2 weeks vs 4) and a better preservation of neurocognitive function. A clinical trial led by the National Cancer Information Center, is randomizing patients with 5-15 BMs to SRS or WBRT in combination with memantine. OS and neurocognitive deterioration free survival are the primary endpoints (NCT03550391). Another trial from Dana-Farber Cancer Institute (NCT03075072) is randomizing patients with 5-20 BMs to SRS or WBRT plus hippocampal sparing and QoL is the primary endpoint (108). Table 3 resumes the ongoing studies investigating this innovative setting.

Conclusions: WBRT remains the standard treatment for patients with >4 metastases. However, there is growing evidence to support the role of SRS for patients with 4-15 metastases to improve the risk-benefit ratio of these patients, reserving WBRT as salvage treatment in case of rapid and progressive intracranial disease. It remains controversial which subpopulation of multiple BM patients benefits most from local treatments, including SRS.

In the past, dose escalation with WBRT plus sequential boost was employed to improve LC within the context of oligometastatic disease (11, 16).

The new technical frontier is to combine WBRT with a simultaneous integrated boost (SIB) in patients with multiple BMs. Modern RT techniques, such as dynamic IMRT and VMAT, are able to generate steep dose gradients between close volumes and to simultaneously delivery different doses to the whole brain and to visible BMs. SIB is then the technological evolution of the sequential boost, with several technical and logistical advantages. SIB deliveries/includes WBRT and a boost on visible BMs in the same session, with an optimized dose distribution, a single simulation protocol, an improvement in patients’ compliance and a reduction in overall treatment time and costs (109). The first feasibility studies involved patients with a limited number or volume of BMs (110, 111). The same results emerged even in patients with a larger number and volume of BMs (109, 112). WBRT+SIB on large lesions appears to be safe and effective even for patients with 4-10 BMs, without significant cognitive decline (113); therefore, SIB is frequently employed for the treatment of poly-metastatic disease. Recently it has also been shown to provide a significantly longer median intracranial PFS (9.1 vs 5.9 months, p=0.001) and median OS (14 vs 11 months, p=0.037) compared to WBRT + sequential boost (114, 115). There is no consensus on the most appropriate hypofractionation schedule and each institute bases the choice on clinical judgment and experience. The most frequently used schedule is in 5 daily fractions with a dose of 20 Gy (4 Gy per day) to the whole brain and a SIB of 40 Gy (8 Gy per day) to the visible BMs (110).

Modern technologies allow for a further innovative step, represented by the addition of hippocampal avoidance in this integrated approach (HA-WBRT+SIB). Through a VMAT or, even better, helical tomotherapy planning, it is possible to simultaneously spare the hippocampi, deliver high doses of radiation to multiple metastases and treat the remaining brain volume with a homogeneous dose distribution, all in a single treatment plan/session (116, 117) (Figure 6). HA-WBRT+SIB can be an effective therapeutic option for patients with multiple BMs: it shows improved LC of treated metastases (98% vs 82% at 1 year; P = 0.007) and improves overall intracranial PFS in comparison with WBRT alone (13.5 vs 6.4 months; P = 0.03) (118).

The potential of HA-WBRT+SIB to prevent neurocognitive effects and to concomitantly improve LC is currently under investigation in randomized, multicenter trials (119, 120).

Conclusions: WBRT+SIB improves LC for the treatment of multiple BMs. The combination of HA-WBRT with SIB could reduce deterioration in neurocognitive function and further improve the therapeutic index for these patients.

Early studies investigating the RT-TKI combination have compared patients treated with WBRT + erlotinib or WBRT alone, revealing a statistically significant increase in ORR, median PFS and median OS (137, 138). These results were not confirmed in a recent phase III trial, showing that concurrent erlotinib with WBRT does not improve intracranial PFS or OS for NSCLC patients with BM, but the combination therapy is well tolerated with no unexpected neurotoxicity (139). A retrospective study by Chen et al, comparing the combination of EGFR-TKI and RT (WBRT and SRS) and EGFR-TKI alone, revealed no statistically significant differences in PFS and OS but only in median time to intracranial progression (21.5 vs 15 mo, p=0.036) (140). In view of the inconclusive data published until now, ongoing prospective trials are examining the effects of the SRS/EGFR-TKI combination in patients with known EGFR-status (Table 4).

Regarding the association of ALK inhibitors and RT, the combination of crizotinib and brain RT brought a statistically significant increase in ORR and median time to tumor progression (7 vs 13.2 months, respectively) (141). Another cohort of ALK-rearranged NSCLC patients treated with RT and TKIs showed favorable intracranial PFS and OS rates (142). However, the ASCEND-4 trial presented no statistically significant differences in terms of outcomes for the addition of RT to ceritinib (143). No relevant results and acute toxicity data are available to justify the implementation in clinical practice of the combination ALK-TKI- RT. To date most of the ongoing trials evaluate the combination EGFR-TKI- RT.

Conclusions: The combination of EGFR and ALK TKIs + RT is still under investigation and preliminary results suggest a possible benefit, particularly in terms of intracranial control, regardless of the radiotherapy technique (SRT or WBRT) and number of brain lesions. The ideal timing for the combination, the type of RT (focal/WBRT) and possible toxicities must be investigated.

The discovery of a lymphatic drainage system in the brain and the ability of T-cells to cross the blood-brain barrier has led to believe that the combination of RT and IT may improve antigen presentation in T-cells (144).

Data regarding the IT-RT combination in driver negative NSCLC patients, mostly from retrospective studies, is contradictory.

Clinical data revealed an increase in OS for the combination of RT and ICIs, starting immune checkpoint inhibitors (ICIs) at least 30 days earlier and continued throughout the RT treatment (145).

The retrospective study by Ahmed et al. presented no additional toxicity for NSCLC patients with BMs who received SRS and anti-PD1/PDL1 therapy (146).

So, there is reassuring data regarding the safety profile and efficacy of the combination of anti-PD1/anti-PDL1 agents and various RT regimens (SRS, WBRT) (147), with only a potential warning on an increased risk of RN (M155). In general practice, at least 50% of physicians do not interrupt ICIs when patients require SRS or WBRT (148).

A randomized study on ICIs with or without SRS in patients with asymptomatic BMs is still lacking to date. The Table 5 shows various ongoing trials examining the use of IT together with different brain radiation techniques.

Conclusions: The role of the RT-IT combination is still unclear. The current trend favors focal treatments such as SRS alone, or adjuvant SRS over WBRT, so further investigation of this combination is still required.