Novel radiotherapeutic strategies in the management of brain metastases: Challenging the dogma

Abstract

The role of radiation therapy in the management of brain metastasis is evolving. Advancements in machine learning techniques have improved our ability to both detect brain metastasis and our ability to contour substructures of the brain as critical organs at risk. Advanced imaging with PET tracers and magnetic resonance imaging-based artificial intelligence models can now predict tumor control and differentiate tumor progression from radiation necrosis. These advancements will help to optimize dose and fractionation for each patient’s lesion based on tumor size, histology, systemic therapy, medical comorbidities/patient genetics, and tumor molecular features. This review will discuss the current state of brain directed radiation for brain metastasis. We will also discuss future directions to improve the precision of stereotactic radiosurgery and optimize whole brain radiation techniques to improve local tumor control and prevent cognitive decline without forming necrosis.

Brain metastasis are the most common intracranial neoplasm in adults occurring in 20–40% of all cancer patients.^1,2 The frequency of brain metastasis is increasing with the most common primary tumors including lung, breast, melanoma, and renal cell carcinoma.² Management of these tumors requires a multidisciplinary approach. Treatment options include systemic therapy, surgery, radiosurgery (SRS), fractionated radiosurgery (fSRS), and whole brain radiation treatment (WBRT).^1,2 In this review we will focus on the management of brain metastasis with radiation treatment. Machine learning algorithms are now able to perform automated detection of metastasis, map brain substructures with interconnected functions and predict tumor response.^3–11 We will highlight the current state of radiation therapy, future directions, and novel technologies.

Modern Role of Radiosurgery

Although considered a straightforward diagnosis, treatment of brain metastases is complex with continued evolution of treatment options. For multiple metastases, traditional WBRT is more commonly replaced with the modern approach of hippocampal sparing whole brain radiotherapy.^12–16 This may be further enhanced in the future with the addition of a simultaneous integrated boost to targetable metastases.¹² A boost delivered simultaneously during WBRT may improve local control and lower of the likelihood of neurologic death in those patients who may not have the opportunity for radiosurgery in the future.

Whenever possible, it is largely accepted that the optimal radiation paradigm for brain metastases is stereotactic radiosurgery. This can be performed in the definitive setting in the absence of surgery and/or whole-brain radiotherapy,^13,17 in the adjuvant setting to bolster local control and minimize the neurocognitive side effects of adjuvant WBRT,^18,19 or in the neo-adjuvant setting to minimize the likelihood of nodular leptomeningeal recurrence and decrease the risk of surgical cavity necrosis.^20,21 Several prospective randomized trials are currently underway to assess the outcomes of neoadjuvant versus adjuvant radiosurgery (NRG BN012—NCT05438212, NCT04474925).

In the setting of multiple metastases, there is growing evidence that the quantity of metastases should not be the driver of whether a patient receives whole brain radiotherapy.^22,23 Rather, factors which need be considered include the patient’s performance status, diagnosis, molecular status with CNS penetrating drugs, and brain metastasis velocity. Clinical logistics are also a consideration with a growing availability of multi-metastasis delivery platforms with either linear accelerator or dedicated radiosurgery platforms such as the Leksell GammaKnife© and CyberKnife ©.

When treating increasing numbers of metastases, it is important to consider the ideal distribution of how to deliver the treatments, both spatially as well as temporally. There is the potential for overlapping intermediate doses within normal brain when targets are close to one another. This could increase the V12 above the desired risk for symptomatic radiation necrosis, which may increase to 10%, 15%, and 20% with volumes of 5 cc, 10 cc, or >15 cc, respectively.²⁴ There also are patient and clinical logistics, such as comfort in a mask and scheduling, which need be considered and may necessitate the distribution of metastases over multiple days. Distributed radiosurgery can be performed in several ways. Adjacent metastases can be treated on the same day versus distributing adjacent metastases across separate days.^25,26 The latter may have the advantage of giving intervening brain time for recovery and lowering the risk of necrosis. Another option may be to treat only the largest, eloquently located metastases in an initial course with re-assessment after a short interval scan for completion of remaining targets.^25,27

A second historical limitation for radiosurgery which needs reconsidering is the size of a metastasis. With a larger metastasis (>2 cm), single fraction radiosurgery is associated with a lower control and higher risk of radiation necrosis.^24,28 Fractionated stereotactic radiotherapy improves the control but albeit with a moderately elevated risk of radiation necrosis.^29,30 An emerging paradigm is adaptive radiotherapy but that alone may not be enough. One example of this is staged fSRS, By allowing a break of 2–4 weeks between a second and third fSRS dose, the volume of surrounding brain tissue decreases and time is allowed for normal tissue repair.^27,31–33 Similarly, personalized ultra-fractionated stereotactic ablative radiotherapy (PULSAR) aims to further enhance the effects of adaptation by introducing strategic time delays between fractions, or rather pulses, of radiosurgery. By allowing time between pulses, volume reductions of up to 40% can be seen.^32,34–36 Based on response, a provider now has the flexibility when adapting. For responsive metastases, volume reduction and/or dose de-escalation could be considered. For non-responsive metastases, dose escalation or surgical intervention may be warranted. This paradigm can also be considered for smaller metastases with significant edema and/or in eloquent locations. Thus, the complexity of brain metastasis radiotherapy will continue to evolve in a personalized fashion.

Postoperative Stereotactic Radiosurgery/Fractionated Radiosurgery

Surgery is an important part of the management of brain metastases. Indications include large metastases greater than 2 cm, symptoms due to mass effect, or to obtain tissue diagnosis. It is usually reserved for patients, with good performance status and 1–3 brain metastases in non-eloquent regions of the brain. Historically, patients were given WBRT after resection to reduce the risk of both local and distant brain failure.^37,38

With the emergence of evidence showing the detrimental effects of WBRT on quality of life and neurocognition in the management of patients with intact brain metastases, interest in utilizing SRS in the post-operative setting grew. Four randomized trials have since evaluated the role and impact of postsurgical WBRT, SRS, or just observation.^18,19,39,40 Two of these trials compared post-operative WBRT with SRS and found lower rates of intracranial progression with WBRT.^18,40 In the NCCTG N107C/CEC.3 trial, there was better quality of life and functional independence, and less cognitive deterioration in the SRS arm both at 3 months and in long term survivors.^18,41 The 1-year tumor bed control was inferior with postoperative SRS (61% compared with 81% in the WBRT arm, P < 0.001). However, with central imaging review of the N107C/CEC.3 patients no longer show a difference in tumor bed control.⁴²

Surgery alone even with a gross total resection has a high 1-year tumor bed failure of up to 65%. Adjuvant radiotherapy halves this risk.^19,37,39 A randomized trial of 132 patients to either observation or single-fraction SRS after surgery and reported a 1-year local control (LC) of 43% in the observation arm and 73% in the SRS arm (P = 0.02). One-year local control (91%) was significantly better in tumors less than 2.5 cm demonstrating that single-fraction SRS is an excellent treatment for small resection cavities.¹⁹ However, larger tumors had significantly higher rates of local failure (LF) and maybe be better managed with fSRS.⁴³ Hypofractionation is hypothesized to be able to maintain high rates of LC with an acceptable risk of radiation necrosis for larger targets. Non-randomized series with 3 or 5 fraction SRT show greater than 80% 1-year LC and low rates of symptomatic radiation necrosis with this strategy in tumors with a pre-operative tumor size greater than 3 cm.^29,44,45

In addition to dose, target delineation may play an important role in LC with cavity SRS. Expert consensus contouring guidelines have been developed that recommend including the surgical window and a margin beyond the initial area of contact in cases with pre-operative tumor abutting or involving the dura or venous sinus.^44,45 Although these guidelines were not based on a pattern of failure analysis there is at least one series that demonstrates the higher risk of local failure in patients with dural contact.⁴⁶

With the increase in cavity SRS in clinical practice, a distinct pattern of recurrence coined nodular leptomeningeal disease (nLMD) has been observed.^47,48 This pattern has been reported in up to 12% of cases at 1-year and in contrast to the sugar-coating appearance seen with classical LMD, nLMD is likely iatrogenic and usually begins with a nodular pachymeningeal appearance. Although some patients can be salvaged with focal radiation or WBRT, this pattern of recurrence is still associated with poor prognosis.^21,48 This along with the larger volume irradiated with post-operative cavity SRS has increased the interest in pre-operative SRS. This approach appears to have low rates of LMD (8% at 2 years) as well as symptomatic radiation necrosis (4% at 2 years) and no difference with surgical complications.^21,49,50

The role of adjuvant radiation treatment following surgical resection for a brain metastasis is based on several trials.^37,51 The earliest study randomized 48 patients with a single brain metastasis to surgical resection followed by adjuvant WBRT or needle biopsy followed by adjuvant WBRT.⁵¹ There was a large survival benefit (40 weeks versus 15 weeks; P < 0.01) as well as improved functional independence (38 weeks versus 8 weeks; P < 0.01), both favoring the addition of surgery. An additional study which randomized 95 patients with a single brain metastasis to surgical resection followed by WBRT or surgical resection with no further treatment.³⁷ There was significant improvement of both local and distant brain control and a lower risk of death due to neurologic cause were favored in the adjuvant WBRT arm. However, there was no significant difference in OS.

While adjuvant WBRT is associated with improved tumor control rates following resection, it is associated with increased risk of neurocognitive decline.^18,52,53 Several recent trials have improved upon conventional 3D conformal WBRT to improve cognition with the addition of memantine and hippocampal avoidance.^15,16,54 Randomized trials have not shown a survival difference with postoperative SRS compared to WBRT, and there are cognitive benefits of SRS alone, thus, SRS is frequently utilized in the postoperative setting.^18,19,42 In 2017, a phase 3 trial that randomized 128 patients underwent resection of a brain metastasis to adjuvant SRS or observation. SRS dosing was volume-based.^3,19The LC for the entire cohort was 43% versus 72% (P = 0.02), favoring SRS. On multivariable analysis, tumors larger than 2.5 cm had a markedly higher risk of local recurrence. On a subgroup analysis, lesions ≤2.5 cm had a 100% LC rate at 1-year. These results suggest larger lesions may require dose escalation for better control. This is in part the hypothesis for the ALLIANCE A071801 (NCT04114981) clinical trial.

A study of 101 patients with a resected brain metastases with a cavity measuring >3 cm were treated with fSRS to a dose of 27 Gy in 3 fractions.⁵⁵ Local control was 93% at 1-year and rates of RN were 9% (any grade) and 5% (symptomatic). A dosimetric analysis revealed that the V24 Gy (not including target volume) was most predictive of RN with rates of 16% and 2% for ≥ 16.8 cm³ and < 16.8 cm³, respectively. A similar study of 187 received postoperative SRS to a dose of 33 Gy in 3 fractions.⁵⁶ The 1-year LC was 88.2% with rates of RN were 19%. An additional study of 46 patients that utilized 3 different fSRS dosing schemes to the postoperative bed either 24 Gy in 3 fractions, 27 Gy in 3 fractions or 30 Gy in 3 fractions.⁵⁷ Rates of RN at 6-months were 0%, 13%, and 37% for the 24, 27, and 30 Gy groups, respectively. Importantly there was worse local control with lower doses. The best control and risk of RN was 27 Gy and this dose is utilized in the ALLIANCE A071801 (NCT04114981).

Five fraction SRS regimens have also been explored in the postoperative setting. In 2019, a study of 122 patients who received fSRS to a dose of 30 Gy in 5 fractions to the resection bed.⁴⁴ The 1-year LC rate was 84% and symptomatic RN occurred in 7% of patients, which is notably lower than the 1 and 3 fraction regimens. Similar findings were observed in a multi-institution study of 558 patients who received postoperative fSRS to a median dose of 30 Gy in 5 fractions.⁵⁸ The rates of LC and RN at 1-year were 84% and 8.6%, respectively.

While there is a lack of randomized data validating the fSRS approach in the postoperative setting, the ALLIANCE A071801 (NCT04114981) trial was closed to accrual in September of 2022.⁵⁹ This phase 3 trial randomized patients with a resected brain metastases to single fraction SRS to a dose of 12–20 Gy or fSRS to a dose of 27 Gy in 3 fractions or 30 Gy in 5 fractions. Dosing in each arm is based on tumor size and volume using the prior ALLIANCE N107C dosing. The primary endpoint is surgical bed recurrence free survival, and the study is estimated to be reported in 2024.

Fractionated Stereotactic Radiosurgery of Intact Brain Metastasis

Fractionated SRS has been increasingly utilized in the intact setting, commonly in patients with large brain metastases who are not surgical candidates.^29,43 In 2016, a single-institution comparative study of 289 patients with previously untreated brain metastases with a tumor diameter >2 cm. Patients underwent single fraction SRS to a dose of 18 or 15–16 Gy or fSRS to a dose of 27 Gy in 3 fractions. The 1-year LC was 77% versus 91% for single fraction SRS and fSRS groups, respectively; RN rates were 20% versus 8% for the single fraction SRS and fSRS groups, respectively. Additionally, the most significant dosimetric factor predicting RN was the V18 Gy (excluding the gross target volume), with RN rates of 5% and 14% for V18 Gy ≤ 30.2 cm³ and >30.2 cm³, respectively.²⁹ A large study utilizing 3 fraction SRS (424 patients with 539 radiosurgery courses) found the optimal dosimetric factors predicting radiation necrosis were V20 < 20 cc and V23 < 15 cc (Target + 1cm-PTV).³⁰

A small single institution retrospective study of 36 patients with previously untreated brain metastases measuring >3 cm in diameter, patients were treated to a dose of 24 Gy delivered over 2–5 fractions.⁶⁰ The 1-year LC and RN rates were 63% and 0%, respectively. However, prognosis was poor for these patients with a 1-year OS was 13%. Importantly, a dose of 24 Gy was subsequently found to be associated with worse tumor bed control in the postoperative setting.⁵⁷

A retrospective series of 102 patients with previously untreated brain metastases ranging in diameter from 2.1–5.0 cm to a dose of 27 Gy in 3 fractions or 32 Gy in 4 fractions.⁶¹ Local control at 1- and 2-years was 96% with a grade 3 RN rate of 6%. A meta-analysis comparing single fraction SRS and fSRS regimens in the upfront setting with tumors measuring 2–3 cm in diameter observed 1-year LC of 77.6% versus 92.9% in the single fraction SRS HSRS groups, respectively (p = 0.2). Additionally, RN rates were 23.1% versus 7.3% for the single fraction SRS and fSRS groups, respectively (p = 0.003).⁴³

An ongoing study examining radiation fractionation schedules is a phase 2 trial being conducted at Mayo clinic sites (NCT05222620). In this trial patients with intact metastases measuring 2–4 cm in diameter to single fraction SRS or fSRS. An additional trial is begin developed in the NRG Oncology randomizing patients between SRS and fSRS (NRG BN013).

Preoperative Versus Postoperative Stereotactic Radiosurgery

Surgical resection with postoperative radiation is the current treatment standard for patients with resectable brain metastases. Patchell et al. showed how surgical resection prior to WBRT improves overall survival (OS).⁵¹ The Patchell group also showed how radiation after surgical resection of a brain metastasis reduces the risk of neurologic death.³⁷ However, WBRT is a toxic treatment that can injure normal brain tissue and cause cognitive toxicity; postoperative stereotactic radiosurgery (SRS) has become the new standard of care to reduce cognitive deterioration through treatment of less normal brain tissue.^18,19 However, postoperative SRS leads to increased rates of radiation necrosis (RN), meningeal disease (MD), and LF. Improved radiosurgery strategies are needed to mitigate these complications.

Postoperative SRS has disadvantages including risk of tumor seeding during surgery and treating a larger tumor cavity which exposes a larger amount of normal brain tissue to radiation.^62,63 Additionally, in JCOG0504 over 30% of patients enrolled to receive SRS post-operatively did not receive their prescribed radiation course.⁶⁴ It has been theorized that pre-operative SRS could sterilize tumor cells prior to seeding and thus minimize the risk of MD, a condition with a dismal prognosis. Through delivering radiation before surgery to an intact tumor, the risk of RN should be lower, and compliance should be higher.

Multiple recent publications have examined preoperative SRS to mitigate these toxicities.^20,21,50,65 In one study at Ohio State, 279 patients with brain metastases who had surgery and SRS were examined.⁶⁵ Patients who had pre-operative radiosurgery (n = 80) had a 6% rate of Grade 2 or higher RN, 3% rate of MD, and 2% rate of LF. This compares favorably to the N107C and MD Anderson studies with a 0–4% rate of Grade 2 or higher RN, 7–28% rate of MD, and 24–38% rate of LF.¹⁹ The updated PROPS-BM publication included 404 patients and documents a 2-year MD rate of 8%, 2-year RN rate of 7.4%, and a 2-year cavity LF rate of 13.7%. The PROPS-BM cohort has an excellent medial OS of 17.2 months with systemic disease control and extent of resection being strong prognostic factors.²¹ These data suggest that preoperative radiosurgery has great potential to improve outcomes for patients with brain metastases, but randomized data is needed. A phase 3 randomized controlled trial examining preoperative versus postoperative SRS (NRG BN012) is currently open and enrolling patents.

Additional strategies are being examined to optimize stereotactic radiosurgery for patients with brain metastases. It is hypothesized that the higher local failure rates seen in the N107C and Mahajan studies are due to the dose de-escalation needed to safely treat large tumors with single fraction radiosurgery.⁴³ For this reason, there has been increased interest in delivering fractionated radiosurgery treatments to deliver a higher biological effective dose (BED) and improve local control. ALLIANCE-A071801 is a phase 3 randomized trial examining postoperative single fraction versus fractionated stereotactic radiosurgery; the trial has finished enrollment and should be reporting results shortly. The RADREMI trial (NCT04047602) examines dose reduced SRS for patients who received an immune checkpoint inhibitor within 30 days of SRS with the goal of reducing symptomatic RN rates.

Cognitive Preservation for Patients With Brain Metastases

A recently published National Cancer Institute directive for brain metastases emphasized the importance of identifying patients at high risk for functional and cognitive impairment and developing strategies to mitigate these toxicities.⁶⁶ Recent radiotherapy brain metastasis clinical trials have had two major focuses to address these needs: identifying an appropriate population to deliver stereotactic radiosurgery (SRS) and developing new techniques to deliver WBRT. For the former, N107C and the Mahajan studies have shown how after surgical resection of a brain metastasis, SRS improves cognitive deterioration free survival compared to WBRT without impacting overall survival.^18,19 For the latter, NRG CC001 showed how hippocampal avoidance WBRT (HA-WBRT) plus memantine can improve cognition without increasing the risk of intracranial progression.¹⁶ However, SRS and HA-WBRT reduce but do not eliminate cognitive decline. Oncologists treating patients with brain metastases have searched for new strategies to further improve cognition.

Increasing utilization of SRS to spare normal brain tissue is one strategy to improve cognition. The CROSS-FIRE and FIRE-SCLC studies show how some small-cell lung cancer (SCLC) patients may benefit from SRS; traditionally, SCLC with brain metastases receive WBRT.^67,68 JLGK0901 showed that overall survival does not differ between patients receiving SRS for 2–4 versus 5–10 brain metastases and therefore patients with numerous brain metastases may benefit from the cognitive sparing of SRS.^22,69 More recent data show how carefully selected patients with up to 15 brain metastases may benefit from SRS.^69,70 Single isocenter multi target (SIMT) techniques utilizing volumetric modulated arc therapy (VMAT) have been shown to be safe, effective, and efficient when treating patients with multiple brain metastases.^71,72 With a median beam-on time of 4 min, SIMT treatments can be performed without reducing the treatment capacity of a department or consuming resources as a part of a longer stereotactic procedure.⁷²Figure 1 demonstrates a patient with 28 brain metastases treated with a SIMT-28 plan, highlighting the precision achieved with the prescription dose and very little normal brain dose with a mean of 578 cGy over 3 fractions. The most recently published ASTRO Clinical Guidelines for patients with brain metastases conditionally recommend SRS for up to 10 brain metastases, but the quality of evidence remains low and prospective clinical trials to further investigate this topic.¹⁴ Current cooperative group trials are examining the role of SRS to treat brain metastases patients with small cell lung cancer (NRG CC009) and patients with a relapsed intracranial disease and a high brain metastasis velocity (NRG BN009).

Figure 1.

Fifty-eight year old male patient with metastatic lung cancer with 28 brain lesions. Panel A and B are coronal and sagittal MRI post contrast volumetric sequences, respectively, with the radiation plan superimposed (yellow isodose line is 24 Gy, red isodose line is 12 Gy, blue isodose line is 8 Gy). Panel C is the dose volume histogram (green—Brain, GTV_total and PTV_total labeled in red).

Developing WBRT sparing techniques beyond the hippocampus is another focus for patients with brain metastases. The amygdala is a structure involved with emotional and memory processing pathway.^73,74 Radiation dose to the amygdala is associated with atrophy, suggesting a role of the amygdala in radiation-induced cognitive decline.⁷⁵ Lesions in the fornix can impact memory recall, and corpus callosum injury is associated with radiation-induced attention and processing speed decline.^76,77 The corpus callosum and fornix are among the most radiation sensitive brain structures with significant radiation dose-dependent changes.⁷⁸ A novel memory-avoidance WBRT (MA-WBRT) has been developed to avoid the hippocampus, amygdala, corpus callosum, and fornix, and quality of life and cognitive data is pending for this novel treatment technique.⁷⁹Figure 2 depicts a memory avoidance plan, demonstrating the memory avoidance contours with prescription dose able to precisely confirm to this central brain volume.

Figure 2.

Panels A, B, and C are axial, coronal and sagittal MRI post contrast volumetric sequences, respectively. The orange contour is the memory avoidance structure (includes bilateral amygdala, bilateral hippocampi, fornix, corpus callosum, hypothalamus, pituitary). The dose color wash depicts a minimum dose of 3000 cGy, demonstrating excellent coverage of the brain while sparing the central memory structures.

In conclusion, there are a plethora of interventions and tools to help our patients treated for brain metastases in the radiation oncology clinic.^80,81 For patients with brain metastases, focus on quality of life and cognition is paramount when discussing goals of care and providing treatment recommendations. The next wave of cooperative group trials for patients with brain metastases will need to involve these endpoints to further advance the care of our patients in clinic.

Salvage Repeat SRS

Despite excellent rates of local control with SRS, tumor recurrence is still seen in 10–30% of cases, particularly in radioresistant and larger tumors.^14,82 Salvage options include surgery, laser interstitial thermal therapy (LITT), systemic therapy, or repeat radiation.⁸³ Surgical resection and LITT may be preferred for symptom alleviation and pathological confirmation of recurrence, however, feasibility is dependent on technical and patient factors such as tumor location, expected survival and performance status.^48,84–87 There may be synergy between LITT and SRS, which is currently being studied (NCT05124912). Systemic therapy is only effective in the minority of patients who have tumors that are responsive to drugs that cross the blood–brain barrier, such as with targeted or immunotherapy. This leaves a growing group of patients left with re-irradiation as the preferred or only feasible option.

One of the biggest challenges associated with retreatment is differentiating tumor progression from RN.^88,89 Standard anatomical MRI and reliance on volumetric changes is inaccurate and insufficient in determining tumor progression. Advanced imaging techniques such as perfusion MRI, spectroscopy or CEST are among the several methods that have been used to improve the sensitivity and specificity of differentiating tumor necrosis from tumor progression.^90–94 More recently, there have been several studies utilizing F-18 Fluciclivine PET imaging to determine true progressive tumor from pseudoprogression.^95,96 The phase 2 study (NCT04410367) utilizing F-18 Fluciclivine PET in addition to MRI imaging for brain metastasis targeted patients with radiation treatment changes post SRS. Among this cohort of 23 patients with 43% having confirmation of recurrent tumor demonstrated an SUV_max threshold of 4.8 conferred a sensitivity of 80% with a specificity of 85% (AUC 0.87).⁹⁵ This is a novel PET agent for brain metastasis given the agent is FDA approved for prostate cancer but demonstrates promising results which offers the best discrimination of pseudoprogression from true progression with an imaging modality to date. Figure 3 demonstrates an example of a patient with an F-18 Fluciclivine PET with SUV_MAX of 2.7 which was found to be radionecrosis and also demonstrated decreased vascular perfusion.

Figure 3.

Sixty-three year old female patient with metastatic leiomyosarcoma status post prior fSRS to 9 brain lesions. Panel A demonstrates an F-18 fluciclivine PET revealing a right frontal mass with SUV_MAX of 2.7. Panel B depicts a heterogeneously enhancing mass with central necrosis in the frontal lobe. Panel C depicts a right frontal lesion with no elevated cerebral blood volume on T1 post-contrast imaging with dynamic contrast enhanced (DCE) perfusion imaging overlaid.

The literature for re-irradiation with SRS has been mostly limited to small institutional series reporting on cohorts of 20–40 patients with LC rates of 50–80% with RN observed in approximately 20–25% of patients.^97–101 The largest study by Kowalchuk et al. pooled results from 8 institutions and included 123 lesions in their analysis. SRS re-irradiation was performed at a median of 12 months from the initial SRS treatment and the median dose was 18Gy. One-year local control was 79% and the risk of RN was 20% (7% were symptomatic). In one series of 84 patients with 108 tumors, the tumor control rate was only 53.5%, likely related to the predominance of larger tumors (mean 6 cc).⁹¹

One of the trends in SRS treatment has been the adoption of fSRS in the treatment of large metastases and those in eloquent areas. The rationale is that hypofractionation may improve the therapeutic window between tumor control and treatment toxicity (ie RN). A large metanalysis performed on 347 patients with 462 lesions demonstrated a 1-year LC of 69% with a RN rate of 16.1%. Each 1 Gy increase in dose resulted in a roughly 5% increase in LC.⁸³ Despite the encouraging results of single and fSRS, high quality studies are needed to guide treatment in this complex group of patients.

Future Directions

The contemporary management of brain metastasis requires human involvement to detect brain metastasis and delineate tumor and normal brain volumes.^5–8,102 Improvements in machine learning algorithms have steadily demonstrated improvements in detecting metastasis. Several studies have demonstrated high sensitivity and specificity for detecting brain metastasis which can outcompete human counterparts.^3,103 As this technology improves this will help improve the efficiency of the radiation oncologist. Software has also been developed to auto contour brain substructures which will improve the treatment planning workflow allowing for more quickly delineating structures and improving interobserver contouring differences.¹⁰³ Novel applications of machine learning are now able to plot connectomics maps utilizing standard MRI volumetric MRI sequences like T1 post contrast, T2 and DWI imaging sequences.^4,6 Neurosurgeons have begun using connectomics maps to plan for surgery in eloquent regions of the brain and there have been attempts to utilize connectomics maps to plan SRS.⁴ Future brain radiation treatment techniques will likely utilize these connectomics maps because the functional units within the brain are more complex than simply sparing any single structure, it is likely that the optimal cognitive and functional outcomes will come from sparing the precise functional connections of different brain tracts for an individual patient based on their anatomy rather than a single substructure like our current hippocampal avoidance WBRT. These innovations will help us to study cognition in a more personalized fashion. Additionally, machine learning and radiomics will likely utilize both MRI and PET based imaging to help best determine which lesions have responded to treatment and which have developed necrosis. Finally, future studies are underway to utilize machine learning to better understand the likely location of brain metastasis failure.^5,8 It is known that certain tumor histology’s have a predilection for certain brain structures like breast cancer and the cerebellum.² Future algorithms are needed to determine where a patient’s future brain metastases are likely to fail, this will aid both improve imaging surveillance of these regions but also inform our radiation fields to potentially treat these areas prophylactically. In the era of precision medicine and molecular characterization, radiation treatment strategies may change for patients with different histology’s, druggable targets, and with immune sensitivity. It is important to integrate targeted therapies or radiotheranostics to maximize tumor control rates for patients with these characteristics where brain directed radiation could be delayed or personalized.

Conclusions

Over the last decade, we have developed many novel radiation techniques and imaging technology to improve our detection and treatment of brain metastasis with radiotherapy. There is excellent tumor control with our available SRS, fSRS, and HA-WBRT techniques. The main limitations of our current treatments are the development of cognitive decline and radiation necrosis. Importantly, it is critical that patients be followed long term with MRI imaging for late toxicity and tumor failure. Integrating cognitive outcomes into clinical practice for long term follow-up is recommended. There are now novel radiation techniques to spare additional brain substructures, connectomics maps to spare important functional tracts and novel imaging to accurately determine radiation necrosis. We must integrate these innovative techniques and technology into prospective trials to continue to improve clinical and quality of life outcomes for patients with brain metastasis.

Funding

None

Supplement sponsorship

This article appears as part of the supplement “Pushing the Boundaries of Radiation Technology for the Central Nervous System,” sponsored by Varian Medical Systems.

Conflict of interest statement

J.D.P. discloses Honoraria from Huron Consulting Group and research support from Varian Medical Systems, GENENTECH and from Kroger outside the submitted work. All other authors declare no conflicts of interest in relation to the work presented in this review.