Abstract
Background:
Brain radiotherapy is used in the management of melanoma brain metastases (MBM) and can result in radionecrosis. Anti-PD-1 is active in the brain and may increase the risk of radionecrosis when combined with radiotherapy. We studied the incidence, associated factors and management of radionecrosis in longer-term survivors with MBM treated with this combination.
Methods:
Patients with MBM treated with radiotherapy and anti-PD-1 who survived >1 year were identified to determine radionecrosis incidence (Cohort A, n = 135). Cohort A plus additional radionecrosis cases were examined for factors associated with radionecrosis and management (Cohort B, n = 148).
Results:
From Cohort A, 17% developed radionecrosis, with a cumulative incidence at 2 years of 18%. Using Cohort B, multivariable analysis confirmed an association between radionecrosis and elevated lactate dehydrogenase (p = 0.0496) and prior treatment with ipilimumab (p = 0.0319). Radionecrosis was diagnosed based on MRI (100%), symptoms (69%) and pathology (56%). Treatment included corticosteroids, bevacizumab and neurosurgery.
Conclusions:
Radionecrosis is a significant toxicity in longer-term melanoma survivors with MBM treated with anti-PD-1 and radiotherapy. Identification of those at risk of radionecrosis who may avoid radiotherapy is required.
Keywords: brain metastases, immunotherapy, melanoma, radionecrosis, radiotherapy
1 |. INTRODUCTION
Melanoma is one of the most frequent cancers to metastasize to the brain (Barnholtz-Sloan et al., 2004). Brain metastases develop in most metastatic melanoma patients at some point in the disease course and are a major cause of melanoma death (Davies et al., 2011). Until recently, treatment options for melanoma brain metastases were limited to radiotherapy and surgery (Andrews et al., 2004; Kocher et al., 2011). Chemotherapy had a low response rate and short duration of response, and it was rare for patients to survive more than 1 year (Agarwala et al., 2004).
Modern systemic therapies targeting the mutated BRAF protein or immune checkpoints have revolutionized melanoma treatment. BRAF inhibitors alone and in combination with MEK inhibitors have high response rates, including the brain, but the duration of response is short, such that most patients with brain metastases progress within 6 months and die within a year (Davies et al., 2017; Long et al., 2012). In contrast, the hallmark of immunotherapy is durable survival (Schadendorf et al., 2015). Anti-PD-1 antibodies, alone or in combination with ipilimumab, have changed the treatment landscape dramatically, such that these drugs are now used for most patients with metastatic melanoma (Larkin et al., 2015; Robert, Long et al., 2015; Robert, Schachter et al., 2015). Until recently, there were little data on the efficacy of immunotherapy for patients with brain metastases, such that many patients underwent upfront or concurrent cerebral radiotherapy with immunotherapy.
Radionecrosis is a well-known long-term complication of cerebral radiotherapy, with pathologic features of liquefactive necrosis and inflammation (Miyatake et al., 2015). Histologic confirmation is often not possible, and the diagnosis is frequently based on radiological criteria (Kohutek et al., 2015; Minniti et al., 2011), largely defined by studies including multiple cancer subtypes in the era prior to immunotherapy. The incidence of radionecrosis after whole brain radiotherapy (WBRT) and/or stereotactic radiosurgery (SRS) varies in the literature between 2% and 30% according to the diagnostic criteria, modality and dose of radiotherapy, and patient/disease characteristics (Kocher et al., 2011; Kohutek et al., 2015; Minniti et al., 2011; Shaw et al., 2000). Moreover, risk of radionecrosis increases over time and may vary according to the type of cancer (Kohutek et al., 2015; Shaw et al., 2000).
To our knowledge, the incidence of radionecrosis in longer-term survivor melanoma patients with brain metastases treated with radiotherapy and anti-PD-1-based immunotherapy has not been explored specifically. One could hypothesize that since survival is prolonged with immunotherapy, radionecrosis may be more prevalent. Furthermore, radiotherapy may interact with immunotherapy to improve T-cell activation and anti-tumour response (Bernstein, Krishnan, Hodge, & Chang, 2016; Liniker et al., 2016) and thus may exacerbate or trigger radionecrosis.
In this retrospective multicentre study, we sought to investigate the incidence, associated factors, presenting features and management of radionecrosis in patients with melanoma brain metastases treated with anti-PD-1 that had survived for more than 1 year from time of radiotherapy.
2 |. METHODS
The study was performed with institutional ethical review board approval. Eligible patients from twelve academic centres between July 2010 and July 2016 fitting in the following inclusion criteria were included as follows: (a) had a confirmed diagnosis of metastatic melanoma to the brain; (b) had received anti-PD-1 therapy; (c) had received WBRT/SRS during or within 1 year prior to anti-PD-1 treatment; and (d) must have survived for longer than 1 year from the date of starting radiotherapy to enable assessment of longer-term toxicity.
In order to assess radionecrosis incidence, consecutive patients who fulfilled inclusion criteria from nine centres were included (Cohort A). Radionecrosis was defined by the investigator based on pathology (reactive changes and necrosis, without viable tumour cells), and in those not undergoing surgery, conventional radiologic features (peripheral enhancement and central hypointensity; Kohutek et al., 2015; Mullins et al., 2005). Patients from Cohort A plus additional consecutive cases of radionecrosis from three additional centres were included to study factors associated with radionecrosis, clinical features and management (Cohort B).
Data examined included demographics (age and sex), prognostic factors at start of anti-PD-1 therapy (mutation status, LDH, ECOG PS, AJCC M-staging according to AJCC 7th edition), brain metastasis characteristics (number and size of the largest lesion), treatments (anti-PD-1 treatment and radiotherapy) and survival. Anti-PD-1 treatment included pembrolizumab or nivolumab. Radiotherapy included WBRT, SRS or both. The timing (prior to or concurrent with anti-PD-1 treatment), number and size of irradiated lesions, dose (Gy) and number of fractions were studied. The proportion of patients who had radionecrosis over the total follow-up period (Cohort A) was calculated by number of patients with radionecrosis/total number of patients. The cumulative incidence of radionecrosis (Cohort A) was estimated using a competing risk model assuming radionecrosis and death as two competing events, which takes into account the time of event occurrence and number of patients at risk. Univariable and multivariate regressions were performed using Cohort A together with additional cases of radionecrosis (Cohort B) to identify factors associated with radionecrosis. The Kaplan–Meier method was used to describe overall survival (Cohort B). All statistical analyses were conducted using the Statistical Analysis System (SAS version 9.4; SAS Institute, Cary, NC, USA).
3 |. RESULTS
3.1 |. Patient characteristics
A total of 135 patients met eligibility criteria for Cohort A; median age was 54.5 years, 61.5% were male, 46.7% had BRAF-mutant melanoma, LDH was elevated in 34.1%, and 61.5% had anatomical (extracranial) AJCC 7th edition M1c disease (Table 1). Almost half (48.1%) the patients were ECOG ≥1, two thirds (63%) were previously treated with ipilimumab, while approximately one third (39.3%, 84% of the BRAF-mutant population) had received prior BRAF/MEK inhibitors. The vast majority of patients (96%) received anti-PD-1 monotherapy (nivolumab or pembrolizumab), and only six patients (4%) received combination nivolumab plus ipilimumab. The median follow-up from starting immunotherapy was 24.2 months (95% CI, 20.1–26.8). Most patients (64%) received SRS, 15% had WBRT, and 21% had both SRS and WBRT in the same treatment course or sequentially, separated by a number of months. The median follow-up from starting radiotherapy was 26.3 months (95% CI, 22.3–29.8).
TABLE 1.
Patient characteristics and treatment from Cohort A (N = 135) and Cohort B (N = 148)
| Characteristics | Cohort A | Cohort B |
|---|---|---|
| Gender | ||
| Male | 83 (61.5%) | 91 (61.5%) |
| Age at anti-PD-1 commencement | ||
| Median (range) | 54.5 (16−89) | 55.5 (16–89) |
| Mutation status | ||
| BRAF | 63 (46.7%) | 70 (47.3%) |
| NRAS | 23 (17.0%) | 23 (15.5%) |
| WT | 49 (36.3%) | 55 (37.2%) |
| LDH level | ||
| Normal | 82 (60.7%) | 87 (58.8%) |
| Elevated | 46 (34.1%) | 54 (36.5%) |
| Missing data | 7 (5.2%) | 7 (4.7%) |
| ECOG PS | ||
| 0 | 63 (46.7%) | 67 (45.3%) |
| ≥1 | 65 (48.1%) | 74 (50%) |
| Missing data | 7 (5.2%) | 7 (4.7%) |
| M-staginga | ||
| M1c | 83 (61.5%) | 92 (62.2%) |
| Brain metastases characteristicsb | ||
| Number of metastases | 2 (1−20) | 2 (1−20) |
| Size of the largest metastasis (cm) | 1.4 (0.3–6.5) | 1.4 (0.3–6.5) |
| Prior systemic therapy | ||
| Ipilimumab | 85 (63.0%) | 93 (62.8%) |
| BRAF/MEK inhibitors | 53 (39.3%) | 59 (39.9%) |
| Anti-PD-1 treatment | ||
| Anti-PD-1 monotherapy | 129 (95.6%) | 142 (95.9%) |
| Combination therapyc | 6 (4.4%) | 6 (4.1%) |
| Radiotherapy | ||
| WBRT | 20 (14.8%) | 21 (14.2%) |
| SRS | 86 (63.7%) | 95 (64.2%) |
| WBRT + SRS | 29 (21.5%) | 32 (21.6%) |
Note. ECOG PS: Eastern Cooperative Oncology Group performance status; LDH: lactate dehydrogenase; SRS: stereotactic radiosurgery; WBRT: whole brain radiotherapy; W T: wild type.
A JCC v7 anatomic staging, excluding LDH and brain metastases.
Reported as median and range.
Anti-PD-1 + Ipilimumab.
3.2 |. Incidence of radionecrosis
With a median follow-up of 26.3 months (95% CI, 22.3–29.8), radionecrosis was diagnosed in 23 (17%) patients, with the cumulative incidence of 2%, 8%, 11% and 18% at 6, 12, 18 and 24 months, respectively (Figure 1). Radionecrosis was defined clinically and prospectively by the investigator based on conventional radiologic appearances in all cases and confirmed pathologically in 56% of cases when surgery was performed, with no discordance between radiology and pathology observed. In this study, features in three sequences on MRI (T1, T1 post-gadolinium and FLAIR) appear to discriminate between radionecrosis or melanoma metastases in patients treated with radiotherapy and anti-PD-1 therapy (Figure 2). Radionecrosis typically displayed the following: (a) isointense signal within the lesion on T1 sequence, (b) peripheral rim enhancement (the “Capsicum/Pepper sign”) on T1 post-gadolinium and (c) extensive oedema on FLAIR sequence. Melanoma metastases typically displayed the following: (a) hyperintense signal within the lesion in T1 sequence, (b) intrinsic high signal and more solid contrast enhancement in T1 post-gadolinium and (c) minimal oedema in FLAIR (Walker et al., 2014). There was no significant difference in the incidence of radionecrosis between those who received SRS (14/86, 16%) or WBRT (2/20, 10%), with a trend to increased incidence in those who received both WBRT and SRS (7/29, 24%).
FIGURE 1.
Cumulative incidence of radionecrosis. This plot shows the cumulative incidence of radionecrosis (RN) in melanoma patients with brain metastases treated with anti-PD-1 who received SRS/WBRT during or within 1 year of anti-PD-1 and who survived ≥1 year from time of radiotherapy. Follow-up is measured in months. The table below shows the number of patients at risk and the cumulative number of events (radionecrosis—RN) at different time points (0, 6, 12, 18, 24, 30 and 36 months)
FIGURE 2.
Typical MRI features of radionecrosis and melanoma metastases. This figure shows the features in three sequences on MRI (T1, left images; T1 post-gadolinium, middle images; and FLAIR, right images) that discriminate between radionecrosis (upper images) or melanoma metastasis (lower images)
3.3 |. Factors associated with radionecrosis
In order to identify factors potentially associated with radionecrosis, we examined Cohort A together with thirteen additional cases of radionecrosis (Cohort B, N = 148; Table 1). Univariable analysis was performed to study the association between different factors and the presence of radionecrosis, including baseline characteristics of patients (Table 2), systemic treatment and surgery (Table 3) and radiotherapy (Table 4). On univariable analysis, radionecrosis was associated with elevated LDH levels (OR 2.61, CI 1.20–5.66, p = 0.015) and prior treatment with ipilimumab (OR 3.11, CI 1.26–7.69, p = 0.014). Interestingly, there was a trend observed with radionecrosis and higher anatomical (extracranial) AJCC 7th edition M-stage (HR 2.17, 0.93–5.04, p = 0.07).
TABLE 2.
Association between baseline patient characteristics and the incidence of radionecrosis (Cohort B)
| Characteristics | No radionecrosis (n = 112) | Radionecrosis (n = 36) | Univariable analysis | |
|---|---|---|---|---|
| OR | p Value | |||
| Gender | ||||
| Female | 45/57 (79%) | 12/57 (21%) | 1 | 0.4637 |
| Male | 67/91 (74%) | 24/91 (26%) | 1.34 (0.61, 2.96) | |
| Age at PD-1 | ||||
| Median | 54.5 | 54.8 | 1.0 0 (0.98, 1.03) | 0.7629 |
| Mutation status | ||||
| WT | 38/55 (69%) | 17/55 (31%) | 1 | 0.4096 |
| BRAF | 53/70 (76%) | 17/70 (24%) | 0.72 (0.4 4, 1.19) | |
| NRASa | 21/23 (91%) | 2/23 (9%) | ||
| LDH levelb | ||||
| Normal | 71/87 (82%) | 16/87 (18%) | 1 | 0.0151 |
| Elevated | 34/54 (63%) | 20/54 (37%) | 2.61 (1.20, 5.66) | |
| ECOG PSb | ||||
| 0 | 53/67 (79%) | 14/67 (21%) | 1 | 0.7641 |
| ≥1 | 60/74 (81%) | 14/74 (19%) | 0.88 (0.39, 2.02) | |
| M-stagingc | ||||
| Other | 47/56 (84%) | 9/56 (16%) | 1 | 0.0717 |
| M1C | 65/92 (71%) | 27/92 (29%) | 2.17 (0.93, 5.04) | |
Note. ECOG PS: Eastern Cooperative Oncology Group performance status; LDH: lactate dehydrogenase; PD-1: anti-PD-1 therapy; W T: wild type.
Univariable analysis not per formed due to small numbers in the “Mutation status” variable (“NRAS” subgroup).
Missing data (n = 7 patients).
AJCC v7 anatomic staging, excluding LDH and brain metastases.
TABLE 3.
Association between systemic treatment and/or surgery and incidence of radionecrosis (Cohort B)
| Characteristics | No radionecrosis (n = 112) | Radionecrosis (n = 36) | Univariable analysis | |
|---|---|---|---|---|
| OR | p Value | |||
| Prior ipilimumab | ||||
| No | 48/55 (87%) | 7/55 (13%) | 1 | 0.0142 |
| Yes | 64/93 (69%) | 29/93 (31%) | 3.11 (1.26, 7.69) | |
| Prior BR AF/MEK inhibitors | ||||
| No | 65/89 (73%) | 24/89 (27%) | 1 | 0.3589 |
| Yes | 47/59 (80%) | 12/59 (20%) | 0.69 (0.31, 1.52) | |
| Prior surgery | ||||
| No | 76/95 (80%) | 19/95 (20%) | 1 | 0.1010 |
| Yes | 36/53 (68%) | 17/53 (32%) | 1.89 (0.88, 4.06) | |
| Anti-PD-1a | ||||
| PD-1 monotherapy | 106/142 (75%) | 36/142 (25%) | ||
| Combination therapyb | 6/6 (10 0%) | 0/6 (0%) | ||
Univariable analysis not per formed due to small numbers in the “Anti-PD-1” variable (“Combination therapy” subgroup).
PD-1 + Ipilimumab.
TABLE 4.
Association between radiotherapy (WBRT ± SRS, timing, number and size of irradiated lesions, dose, fractions) and incidence of radionecrosis (Cohort B)
| Characteristics | No radionecrosis (n = 112) | Radionecrosis (n = 36) | Univariable analysis | |
|---|---|---|---|---|
| OR | p Value | |||
| Radiotherapy | ||||
| SRS only | 72/95 (76%) | 23/95 (24%) | 1 | 0.3824 |
| WBRT only | 18/21 (86%) | 3/21 (14%) | 0.52 (0.14, 1.93) | |
| SRS + WBRT | 22/32 (69%) | 10/32 (31%) | 1.42 (0.59, 3.4 4) | |
| SRS (n = 95) | 72 | 23 | ||
| Timing | ||||
| Prior | 40/55 (73%) | 15/55 (27%) | 1 | 0.4155 |
| Concurrent | 32/40 (80%) | 8/40 (20%) | 0.67 (0.25, 1.77) | |
| Number of lesions Median (range) | 1 (1, 17) | 1 (1–20) | 1.07 (0.96, 1.21) | 0.2220 |
| Size of the largest lesiona Median (range) | 1.2 (0.3, 6.45) | 1.5 (0.5, 3.9) | 1.17 (0.77, 1.78) | 0.4673 |
| Dose in Gy Median (range) | 20 (12, 30) | 21 (18, 40) | 0.99 (0.90, 1.09) | 0.8490 |
| Number of fractions (median) | 1 (1, 5) | 1 (1, 5) | 1.04 (0.85, 1.28) | 0.6883 |
| WBRT (n = 21) | 18 | 3 | ||
| Timing | ||||
| Prior | 12/14 (86%) | 2/14 (14%) | 1 | 1.000 |
| Concurrent | 6/7 (86%) | 1/7 (14%) | 1.0 0 (0.07, 13.37) | |
| Number of lesions Median (range) | 2 (1, 20) | 2 (1, 3) | 0.77 (0.43, 1.38) | 0.3859 |
| Size of the largest lesionb Median (range) | 1.8 (0.4, 5.9) | 1.3 (0.9, 4.0) | 0.93 (0.43, 2.01) | 0.8501 |
| Dose in Gy Median (range) | 30 (20, 30) | 30 (20, 30) | 0.95 (0.73, 1.23) | 0.6784 |
| Number of fractions Median (range) | 10 (5, 10) | 10 (10, 10) | 8.16 (0.0 0, 84E34) | 0.9593 |
| SRS + WBRT (n = 32) | 22 | 10 | ||
| Timing | ||||
| Both prior | 11 (65%) | 6 (35%) | 1 | 0.2594 |
| SRS or WBRT prior | 10 (83%) | 2 (17%) | 0.27 (0.02, 3.67) | |
| Both concurrent | 1 (33%) | 2 (67%) | 0.10 (0.01, 1.71) | |
| Number of lesions Median (range) | 2 (1, 20) | 2.5 (1, 20) | 1.01 (0.86, 1.17) | 0.9212 |
| Size of the largest lesionc Median (range) | 1.5 (0.9, 4.3) | 2.0 (0.4, 3.5) | 1.05 (0.47, 2.33) | 0.9108 |
| Dose in Gyd Median (range) | ||||
| SRS | 20 (13, 40) | 20 (12, 30) | 0.95 (0.79, 1.14) | 0.5623 |
| WBRT | 30 (20, 45) | 30 (20, 37.5) | 1.11 (0.95, 1.31) | 0.1929 |
| Number of fractions Median (range) | ||||
| SRS | 1 (1, 5) | 1 (1, 5) | 1.26 (0.66, 2.42) | 0.4802 |
| WBRT | 10 (5, 15) | 10 (8, 18) | 1.42 (0.98, 204) | 0.0632 |
Note. Gy: Gray; SRS: stereotactic radiosurgery; WBRT: whole brain radiotherapy
Size of the largest lesion: 27 missing values
Size of the largest lesion (WBRT only): two missing values
Size of the largest lesion (SRS + WBRT): eight missing values
Bivariate analysis was per formed for “Dose” and “Number of fractions” including SRS and WBRT as variables.
No association was found with other factors, including sex, age, mutational status, ECOG PS or prior treatment with BRAF/MEK inhibitors or surgery. Due to small numbers, univariable analysis was not performed to study the association between radionecrosis and treatment with anti-PD-1 alone versus the combination of anti-PD-1 and ipilimumab. No association was found with type of radiotherapy (WBRT, SRS or both), number and size of irradiated lesions, radiation dose or number of fractions. Sixty-two (42%) patients had concurrent anti-PD-1 and radiotherapy, and 86 (58%) had prior radiotherapy and then anti-PD-1 therapy. In the prior radiotherapy group, median time from radiotherapy to anti-PD-1 treatment was 3.8 months (0.03–12) in the SRS group and 3.7 months (0.02–12) in the WBRT group. Incidence of radionecrosis did not differ whether radiotherapy was given concurrently or sequentially with anti-PD-1 therapy.
Multivariable logistic regression confirmed the association between elevated LDH levels (OR 2.33, CI 1.00–5.44, p = 0.0496) and prior treatment with ipilimumab (OR 2.79, CI 1.09–7.10, p = 0.0319) with radionecrosis (Table 5), with the median time between ipilimumab and anti-PD-1 treatment or radiotherapy being similar at approximately 4 months.
TABLE 5.
Univariable and multivariable analyses of factors associated with radionecrosis. (Cohort B, N = 148)
| Univariable analysis | Multivariable analysis | |||
|---|---|---|---|---|
| Label | OR | p-Value | OR | p-Value |
| LDH level | ||||
| Normal | 1 | 0.0151 | 1 | 0.0496 |
| Elevated | 2.61 (1.20, 5.66) | 2.33 (1.0 0, 5.4 4) | ||
| M-staginga | ||||
| Other | 1 | 0.0717 | 1 | 0.6345 |
| M1c | 2.17 (0.93, 5.04) | 1.26 (0.49, 3.22) | ||
| Prior ipilimumab | ||||
| No | 1 | 0.0142 | 1 | 0.0319 |
| Yes | 3.11 (1.26, 7.69) | 2.79 (1.09, 7.10) | ||
Note. LDH: lactate dehydrogenase.
AJCC v7 anatomic staging, excluding LDH and brain metastases.
3.4 |. Features and management of radionecrosis
From Cohort B, 36 patients were diagnosed with radionecrosis based on radiographic findings (Supporting information Table S1), after a median 11.6 months (1–34 months). Twenty-five patients (69%) developed symptoms, with headaches, dizziness, confusion, weakness and seizures the most frequent among the wide spectrum of symptoms. Interestingly, in these 69% of patients, the median time to first radionecrosis-related symptom was 8.5 months (0.4–34 months), similar to the radiographic findings (approximately 9 months). The diagnosis of radionecrosis was confirmed pathologically in approximately half of the patients (56%), after a median 11.8 months (1.2–36 months).
Twenty-six patients (72%) received treatment for radionecrosis, including steroids (n = 20, 56%), bevacizumab (n = 10, 28%) and surgery (n = 20, 56%; Supporting information Table S2). Most patients were treated with steroids as first-line therapy (53%), and in all but one case, bevacizumab was given after steroids. Six patients (17%) were treated with all three strategies, twelve patients (33%) were treated with two, eight patients (22%) were treated with one, and ten (28%) patients did not require treatment. The best symptomatic and/or radiological outcomes were observed in patients who underwent surgery (19/20, 95%) or were treated with bevacizumab (9/10, 90%), rather than steroids alone (11/20, 55%). Dexamethasone was dosed between 4 and 16 mg per day, and the median number of doses of bevacizumab was two (range 1–5).
3.5 |. Survival
In this selected population of longer-term survivors with brain metastases, median overall survival was 39.1 months (95% IC, 32–NR), with no difference in overall survival between the group of patients with and without radionecrosis. Eleven patients (31%) with radionecrosis died from disease progression, with no reported deaths due to radionecrosis.
4 |. DISCUSSION
The optimal management of patients with brain metastases is a clinical dilemma. Anti-PD-1 therapy improves survival for patients with several cancer types, including melanoma, Hodgkin’s lymphoma, lung, kidney and head & neck cancers, and is increasingly becoming standard care. Many patients with brain metastases also receive cerebral radiotherapy, which has high local control rates, but can cause radionecrosis. The risk, features and management of radionecrosis in the era of anti-PD-1 immunotherapy are unknown. In this study, the first to specifically examine radionecrosis in long-term survivors treated with radiotherapy and anti-PD-1 therapy, radionecrosis was diagnosed in 17% of longer-term survivors, with a cumulative incidence of 18% at 2 years. We identified factors that appear to be associated with radionecrosis, including elevated LDH and prior treatment with ipilimumab, and reported the utility of various strategies to manage radionecrosis in this setting.
While radionecrosis is well documented after radiotherapy alone (SRS and/or WBRT; Kohutek et al., 2015), little data are available regarding the risk in longer-term survivors, as survival has generally been poor in the era prior to effective systemic therapy. Few small studies have suggested an increased incidence of radionecrosis with immunotherapy compared to radiotherapy alone (Choong et al., 2017; Colaco, Martin, Kluger, Yu, & Chiang, 2016; Four et al., 2018; Kaidar-Person et al., 2016; Martin et al., 2018; Rahman et al., 2018); however, most patients in these studies were treated with ipilimumab, and few patients survived more than 1 year. In this study, of longer-term surviving patients treated with the more active and more widely used anti-PD-1 therapy, the incidence of radionecrosis was a clinically relevant 18% at 2 years. Whether this risk is simply driven by longer survival, a drug-radiation interaction that boosts local autoimmunity, or a combination of factors remains unknown. Clinical trials examining the efficacy and safety of combination immunotherapy with/without stereotactic radiosurgery should definitively answer this question. Nevertheless, the fact that immunotherapy improves survival across many cancers suggests that this phenomenon will be observed in the clinic more frequently in the years ahead.
Elevated LDH and prior treatment with ipilimumab were associated with a higher incidence of radionecrosis in this study. LDH is a prognostic marker in melanoma, including the subset of patients with brain metastases (Eigentler et al., 2011). Tumours that produce higher or lower LDH may also react differently to radiotherapy and anti-PD-1 therapy as it has been shown that a high serum LDH level reflects an anaerobic oxidative metabolic state (Ho et al., 2012). Even though prior treatment with ipilimumab was associated with a higher likelihood of developing radionecrosis in this study, we were unable to show the same effect on patients treated with the combination ipilimumab plus nivolumab, possibly due to small numbers. Interestingly and unlike prior studies, there was no difference among different types of radiotherapy (WBRT, SRS or both), but this was likely influenced by the small proportion of patients receiving WBRT (14%) or both treatments (22%). Similarly, radiation dose and fractionation, as well as size and number of irradiated lesions, were not associated with radionecrosis. These results differ from previous studies where larger lesions have an increased risk of radionecrosis (Blonigen et al., 2010; Kaidar-Person et al., 2016; Kohutek et al., 2015), and could be explained by the limited number of patients in this study, or possibly the use of immunotherapy and longer-term survival. We did not have a control group of patients treated with radiotherapy alone; however, prior studies have reported rates from 2% to 30% (Kocher et al., 2011; Kohutek et al., 2015; Minniti et al., 2011; Shaw et al., 2000), albeit with very short follow-up, largely due to poor overall survival in an era before effective immunotherapy.
The diagnosis of radionecrosis, that is, the ability to differentiate between radionecrosis and tumour progression, remains a challenge (Walker et al., 2014). A limitation of this study is that all patients had the diagnosis of radionecrosis based on MRI imaging; however, almost prior studies have had this similar limitation, and the majority of the patients in this study had pathological confirmation. This histologic confirmation is much higher than most previous studies of radionecrosis (Kohutek et al., 2015; Leeman et al., 2013; Wang et al., 2018). Recent reports have shown that [F-18]fluoroethyl-L-tyrosine (FET) PET may differentiate brain metastases from radionecrosis with a higher accuracy, and this should be further evaluated, including in the immunotherapy-radiotherapy setting (Galldiks & Langen, 2016).
There are little data to suggest the best management of radionecrosis. Surgery, steroids and bevacizumab are used with varying success (Glitza et al., 2017; Levin et al., 2011; Patel, Patel, Cobb, Benkers, & Vermeulen, 2014). Results of this study, the first in the context of immunotherapy, suggest that radionecrosis management is difficult, with most patients needing two or more strategies to control symptoms. The ideal dosing and duration of treatment with steroids and/or bevacizumab, and the role of surgery remain to be determined, and management should be coordinated through a multidisciplinary team.
Two clinical trials of systemic immunotherapy have recently demonstrated impressive activity in the brain. The ABC and CheckMate204 trials of combination ipilimumab plus nivolumab in untreated patients with asymptomatic melanoma brain metastases had an approximate 50% intracranial objective response rate, and no cases of necrosis were reported (Long et al., 2018; Tawbi et al., 2018). Further follow-up is required, but if responses remain as durable as seen with extracranial disease, immunotherapy may be considered as first-line treatment for selected patients with asymptomatic brain metastases, thus avoiding the risk of radionecrosis. At this time, however, surgery and radiotherapy remain important local treatments for most patients with brain metastases, and the optimal management for patients with melanoma brain metastases should be defined by a multidisciplinary team composed by medical and radiation oncologists, a radiologist and a neurosurgeon.
In the modern era of anti-PD-1-based therapy for patients with metastatic melanoma, this study suggests that the incidence of radionecrosis in longer-term surviving patients treated with cerebral radiotherapy is clinically significant, and that the management of radionecrosis in this setting is challenging. Recent trials suggest that patients with melanoma brain metastases treated with combination immunotherapy have intracranial responses, such that some patients may not need radiotherapy and thus avoid the risk of radionecrosis. Further research is required to select patients who may safely avoid radiotherapy, and how to best deliver radiotherapy to those who truly need it.
Supplementary Material
Significance.
Many patients with melanoma brain metastases receive radiotherapy and anti-PD-1 therapy, and this study demonstrates that approximately one fifth of those who survive beyond 1 year develop radionecrosis, which causes symptoms that can be difficult to manage. Recent data demonstrate high activity of anti-PD-1-based therapy in patients with brain metastases, such that some patients may be spared radiotherapy and the risk of radionecrosis.
ACKNOWLEDGEMENTS
GDG is pending patent on radiosensitivity and the immune microenvironment. SMG is supported by the University of Zurich and medAlumni. MS is supported by Fellowships from Pfizer Australia, the Australian National Health and Medical Research Council veski and the Victorian Cancer Agency. ZE is supported by Moffitt Cancer Center NCI Skin SPORE (5P50CA168536). DBJ is supported by NCI/NIH K23 CA204726. AMM is supported by a Cancer Institute NSW Fellowship. We thank all the patients who participated in this study and their families.
Funding information
Moffitt Cancer Center NCI Skin SPORE, Grant/Award Number: 5P50CA168536; Australian National Health and Medical Research Council veski; Cancer Institute NSW; NCI/NIH K23, Grant/Award Number: CA204726; University of Zurich and medAlumni; Victorian Cancer Agency
Dr. Goldinger has received travel support from MSD, BMS, Roche and Novartis. She has received compensation as member of the scientific advisory board of MSD, BMS, Roche and Novartis. Dr. Roberts-Thomson has received travel support from MSD. She has received compensation as a speaker for BMS and Novartis. Dr. Atkinson has received travel support from BMS, MSD, Novartis, Pierre Fabre, Merck Serono. She has received compensation as member of the scientific advisory board of BMS, MSD, Novartis, Roche and as a speaker for BMS and MSD. Dr. Guminski has received travel support from BMS. He has received compensation as member of the scientific advisory board of Merck KgA, BMS, Roche, Sun Pharma and Regeneron. Dr. Carlino has received compensation as member of the scientific advisory board of MSD, Novartis, BMS, Amgen and Pierre Fabre. Dr. Long has received compensation as member of the scientific advisory board of Amgen, Array, BMS, MSD, Novartis, Pierre Fabre, Roche and Incyte. Dr. Johnson has received compensation as member of the scientific advisory board of BMS, Genoptix, Incyte, MSD and Novartis. Dr. Menzies has received compensation as member of the scientific advisory board of MSD, BMS, Novartis and Pierre Fabre.
Footnotes
CONFLICT OF INTEREST
All other authors declare no conflict of interest.
ETHICS APPROVAL AND CONSENT TO PARTICIPATE
This study was performed in accordance with the Declaration of Helsinki. Ethical approval from the Sydney Local Health District Human Research Ethics Committee, Protocol Number X15-0454 and HREC/11/RPAH/444.
SUPPORTING INFORMATION
Additional supporting information may be found online in the Supporting Information section at the end of the article.
REFERENCES
- Agarwala SS, Kirkwood JM, Gore M, Dreno B, Thatcher N, Czarnetski B, … Rankin EM (2004). Temozolomide for the treatment of brain metastases associated with metastatic melanoma: A phase II study. Journal of Clinical Oncology, 22(11), 2101–2107. 10.1200/JCO.2004.11.044 [DOI] [PubMed] [Google Scholar]
- Andrews DW, Scott CB, Sperduto PW, Flanders AE, Gaspar LE, Schell MC, … Curran WJ (2004). Whole brain radiation therapy with or without stereotactic radiosurgery boost for patients with one to three brain metastases: Phase III results of the RTOG 9508 randomised trial. Lancet, 363(9422), 1665–1672. 10.1016/S0140-6736(04)16250-8 [DOI] [PubMed] [Google Scholar]
- Barnholtz-Sloan JS, Sloan AE, Davis FG, Vigneau FD, Lai P, & Sawaya RE (2004). Incidence proportions of brain metastases in patients diagnosed (1973 to 2001) in the Metropolitan Detroit Cancer Surveillance System. Journal of Clinical Oncology, 22(14), 2865–2872. 10.1200/JCO.2004.12.149 [DOI] [PubMed] [Google Scholar]
- Bernstein MB, Krishnan S, Hodge JW, & Chang JY (2016). Immunotherapy and stereotactic ablative radiotherapy (ISABR): A curative approach? Nature Reviews Clinical Oncology, 13(8), 516–524. 10.1038/nrclinonc.2016.30 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Blonigen BJ, Steinmetz RD, Levin L, Lamba MA, Warnick RE, & Breneman JC (2010). Irradiated volume as a predictor of brain radionecrosis after linear accelerator stereotactic radiosurgery. International Journal of Radiation Oncology Biology Physics, 77(4), 996–1001. 10.1016/j.ijrobp.2009.06.006 [DOI] [PubMed] [Google Scholar]
- Choong ES, Lo S, Drummond M, Fogarty GB, Menzies AM, Guminski A, … Hong AM (2017). Survival of patients with melanoma brain metastasis treated with stereotactic radiosurgery and active systemic drug therapies. European Journal of Cancer, 75, 169–178. 10.1016/j.ejca.2017.01.007 [DOI] [PubMed] [Google Scholar]
- Colaco R, Martin P, Kluger HM, Yu JB, & Chiang VL (2016). Does immunotherapy increase the rate of radiation necrosis after radiosurgical treatment of brain metastases? Journal of Neurosurgery, 125(1), 17–23. [DOI] [PubMed] [Google Scholar]
- Davies MA, Liu P, McIntyre S, Kim KB, Papadopoulos N, Hwu W-J, … Bedikian A. (2011). Prognostic factors for survival in melanoma patients with brain metastases. Cancer, 117(8), 1687–1696. 10.1002/cncr.25634 [DOI] [PubMed] [Google Scholar]
- Davies MA, Saiag P, Rober t C, Grob J-J, Flaher t y KT, Arance A, … Long GV (2017). Dabrafenib plus trametinib in patient s with BRAFV600-mutant melanoma brain metastases (COMBI-MB): A multicentre, multicohort, open-label, phase 2 trial. The Lancet Oncology, 18(7), 863–873. 10.1016/S1470-2045(17)30429-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Eigentler TK, Figl A, Krex D, Mohr P, Mauch C, Rass K, … Schadendorf D. (2011). Number of metastases, serum lactate dehydrogenase level, and type of treatment are prognostic factors in patients with brain metastases of malignant melanoma. Cancer, 117(8), 1697–1703. 10.1002/cncr.25631 [DOI] [PubMed] [Google Scholar]
- Four SD, Janssen Y, Michotte A, Van Binst A-M, Duerinck J, Neyns B, … Brussels B. (2018). Focal radiation necrosis of the brain in patients with melanoma brain metastases treated with pembrolizumab. Cancer Medicine, 1–10, 10.1002/cam4.1726 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Galldiks N, & Langen KJ (2016). Amino acid PET – An imaging option to identify treatment response, posttherapeutic effects, and tumor recurrence? Frontiers in Neurology, 7, 120. 10.3389/fneur.2016.00120 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Glitza IC, Guha-Thakurta N, D’Souza NM, Amaria RN, McGovern SL, Rao G, & Li J. (2017). Bevacizumab as an effective treatment for radiation necrosis after radiotherapy for melanoma brain metastases. Melanoma Research, 27(6), 580–584. 10.1097/CMR.0000000000000389 [DOI] [PubMed] [Google Scholar]
- Ho J, de Moura M, Lin Y, Vincent G, Thorne S, Duncan LM, … Moschos SJ (2012). Importance of glycolysis and oxidative phosphorylation in advanced melanoma. Molecular Cancer, 11(1), 76. 10.1186/1476-4598-11-76 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kaidar-Person O, Zagar TM, Deal A, Moschos SJ, Ewend MG, Sasaki-Adams D, …Chera BS (2016). The incidence of radiation necrosis following stereotactic radiotherapy for melanoma brain metastases: The potential impact of immunotherapy. Anti-Cancer Drugs, 28, 669–675. [DOI] [PubMed] [Google Scholar]
- Kocher M, Soffietti R, Abacioglu U, Villà S, Fauchon F, Baumer t BG, … Mueller R-P (2011). Adjuvant whole-brain radiotherapy versus observation after radiosurgery or surgical resection of one to three cerebral metastases: Results of the EORTC22952–26001 study. Journal of Clinical Oncology, 29(2), 134–141. 10.1200/JCO.2010.30.1655 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kohutek ZA, Yamada Y, Chan TA, Brennan CW, Tabar V, Gutin PH, … Beal K. (2015). Long-term risk of radionecrosis and imaging changes after stereotactic radiosurgery for brain metastases. Journal of Neuro-Oncology, 125(1), 149–156. 10.1007/s11060-015-1881-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Larkin J, Chiarion-Sileni V, Gonzalez R, Grob JJ, Cowey CL, Lao CD, … Wolchok JD (2015). Combined nivolumab and ipilimumab or monotherapy in untreated melanoma. The New England Journal of Medicine, 373(1), 23–34. 10.1056/NEJMoa1504030 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Leeman JE, Clump DA, Flickinger JC, Mintz AH, Burton SA, & Heron DE (2013). Extent of perilesional edema differentiates radionecrosis from tumor recurrence following stereotactic radiosurgery for brain metastases. Neuro-Oncology, 15(12), 1732–1738. 10.1093/neuonc/not130 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Levin VA, Bidaut L, Hou P, Kumar AJ, Wefel JS, Bekele BN, … Jackson EF (2011). Randomized double-blind placebo-controlled trial of bevacizumab therapy for radiation necrosis of the central nervous system. International Journal of Radiation Oncology Biology Physics, 79(5), 1487–1495. 10.1016/j.ijrobp.2009.12.061 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Liniker E, Menzies A. m., Kong B. y., Cooper A, Ramanujam S, Lo S, … Long G. v. (2016). Activity and safety of radiotherapy with anti-PD-1 drug therapy in patients with metastatic melanoma. OncoImmunology, 5(9), e1214788. 10.1080/2162402X.2016.1214788 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Long GV, Atkinson V, Lo S, Sandhu S, Guminski AD, Brown MP, … McArthur GA (2018). Combination nivolumab and ipilimumab or nivolumab alone in melanoma brain metastases: A multicentre randomised phase 2 study. The Lancet Oncology, 19(5), 672–681. 10.1016/S1470-2045(18)30139-6 [DOI] [PubMed] [Google Scholar]
- Long GV, Trefzer U, Davies MA, Kefford RF, Ascierto PA, Chapman PB, … Schadendorf D. (2012). Dabrafenib in patients with Val600Glu or Val600Lys BRAF-mutant melanoma metastatic to the brain (BREAKMB): A multicentre, open-label, phase 2 trial. The Lancet Oncology, 13(11), 1087–1095. 10.1016/S1470-2045(12)70431-X [DOI] [PubMed] [Google Scholar]
- Martin AM, Cagney DN, Catalano PJ, Alexander BM, Redig AJ, Schoenfeld JD, & Aizer AA (2018). Immunotherapy and symptomatic radiation necrosis in patients with brain metastases treated with stereotactic radiation. JAMA Oncol, 4(8), 1123–1124. 10.1001/jamaoncol.2017.3993 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Minniti G, Clarke E, Lanzetta G, Osti MF, Trasimeni G, Bozzao A, … Enrici RM (2011). Stereotactic radiosurgery for brain metastases: Analysis of outcome and risk of brain radionecrosis. Radiation Oncology (London, England), 6(1), 48. 10.1186/1748-717X-6-48 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Miyatake S, Nonoguchi N, Furuse M, Yoritsune E, Miyata T, Kawabata S, & Kuroiwa T. (2015). Pathophysiology, diagnosis, and treatment of radiation necrosis in the brain. Neurologia Medico-Chirurgica (Tokyo), 55(1), 50–59. 10.2176/nmc.ra.2014-0188 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mullins ME, Barest GD, Schaefer PW, Hochberg FH, Gonzalez RG, & Lev MH (2005). Radiation necrosis versus glioma recurrence: Conventional MR imaging clues to diagnosis. American Journal of Neuroradiology, 26(8), 1967–1972. [PMC free article] [PubMed] [Google Scholar]
- Patel U, Patel A, Cobb C, Benkers T, & Vermeulen S. (2014). The management of brain necrosis as a result of SRS treatment for intracranial tumors. Translational Cancer Research, 3(4), 373–382. 10.21037/2950. [DOI] [Google Scholar]
- Rahman R, Niemierko A, Cortes A, Oh KS, Flaherty KT, Lawrence DP, … Shih HA (2018). The use of Anti-PD1 Therapy in Melanoma Patients with Known Brain Metastases: Survival, Durable Intracranial Progression Free Survival and Radionecrosis. International Journal of Radiation Oncology, Biology, Physics, 102(3), e275. [Google Scholar]
- Robert C, Long GV, Brady B, Dutriaux C, Maio M, Mortier L, … Ascier to PA (2015). Nivolumab in previously untreated melanoma without BRAF mutation. New England Journal of Medicine, 372(4), 320–330. 10.1056/NEJMoa1412082 [DOI] [PubMed] [Google Scholar]
- Robert C, Schachter J, Long GV, Arance A, Grob JJ, Mortier L, … Ribas A. (2015). Pembrolizumab versus ipilimumab in advanced melanoma. New England Journal of Medicine, 372(26), 2521–2532. 10.1056/NEJMoa1503093 [DOI] [PubMed] [Google Scholar]
- Schadendorf D, Hodi FS, Robert C, Weber JS, Margolin K, Hamid O, … Wolchok JD (2015). Pooled analysis of long-term survival data from phase II and phase III trials of ipilimumab in unresectable or metastatic melanoma. Journal of Clinical Oncology, 33(17), 1889–1894. 10.1200/JCO.2014.56.2736 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Shaw E, Scott C, Souhami L, Dinapoli R, Kline R, Loeffler J, & Farnan N. (2000). Single dose radiosurgical treatment of recurrent previously irradiated primary brain tumors and brain metastases: Final report of RTOG protocol 90–05. International Journal of Radiation Oncology Biology Physics, 47(2), 291–298. 10.1016/S0360-3016(99)00507-6 [DOI] [PubMed] [Google Scholar]
- Tawbi HA, Forsyth PA, Algazi A, Hamid O, Hodi FS, Moschos SJ, … Margolin K. (2018). Combined nivolumab and ipilimumab in melanoma metastatic to the brain. New England Journal of Medicine, 379(8), 722–730. 10.1056/NEJMoa1805453 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Walker AJ, Ruzevick J, Malayeri AA, Rigamonti D, Lim M, Redmond KJ, & Kleinberg L. (2014). Postradiation imaging changes in the CNS: How can we differentiate between treatment effect and disease progression? Future Oncology, 10(7), 1277–1297. 10.2217/fon.13.271 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wang B, Zhang Y, Zhao B, Zhao P, Ge M, Gao M, Liu Y. (2018). Postcontrast T1 mapping for differential diagnosis of recurrence and radionecrosis after gamma knife radiosurgery for brain metastasis. American Journal of Neuroradiology, 39(6), 1025–1031. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.