Prevention of brain metastases in human epidermal growth factor receptor 2-positive breast cancer

INTRODUCTION

Metastatic breast cancer (MBC) is the second most frequent provider of brain metastases, immediately after lung cancer [1], and their incidence is of partic- ular importance for human epidermal growth factor receptor 2 (HER2)-positive MBC. In the Epidemiolog- ical Strategy and Medical Economics (ESME) study, a French real-life cohort of MBC patients, brain metas- tases were detected at diagnosis for 9% of patients with estrogen receptor-positive estrogen receptor- positive/HER2-positive MBC and 17% of patients with estrogen receptor-negative/HER2-positive MBC [2]. Furthermore, 34 and 49% of patient with estrogen receptor-positive/HER2-positive and estro- gen receptor-negative/HER2-positive MBC experi- enced a brain metastasis progression during their metastatic disease, respectively. After a 30-month follow-up, median overall survival after brain metas- tasis diagnosis was 19 months for patients with estro- gen receptor-positive/HER2-positive and 13 months for patient with estrogen receptor-negative/ HER2-positive MBC. Brain-metastasis evolution is associated with increased neurologic symptoms, more aggressive treatments, and decreased survival [3]. As a consequence, prevention of brain metastases is an important clinical issue, and could potentially change these patients’ outcome. The aim of this reviewisto describe the current knowledge onscreen- ing and prevention strategy, with a special focus on recent drugs, neratinib and tucatinib.

EARLY DETECTION OF BRAIN METASTASES

The identification of patients at risk of developing brain metastases could be crucial, as their early treatment. Better outcomes may come from hippo- campal avoidance WBRT as it has proven to better preserve cognitive function in a recent large ran- domized phase III trial [11].

‘PREVENTIVE’ WHOLE-BRAIN RADIOTHERAPY AFTER FOCALIZED TREATMENT

Following surgical resection or SRS of a unique or limited number of brain metastases (up to four), whole-brain radiotherapy (WBRT) can still be con- sidered as a therapeutic option in HER2-positive breast cancer brain metastases [4]. The aim of this ‘preventive’ (adjuvant) WBRT is to eradicate micro- metastases that could be present in the brain and avoid neurologic progression. There have been numerous large randomized phase III clinical trials testing the added value of WBRT after surgical resec- tion or SRS [7–10]. Overall, these trials all demon- strated an improvement of the intracranial disease control, but without benefit in overall survival, and with more neurocognitive toxicities. Of note, none of these trials has been specifically conducted in a population of breast cancer patients. These results do not support WBRT as a standard preventive diagnosis may potentially provide easier access to focalized treatments (i.e., surgery or stereotactic radiosurgery [SRS] [4]). So far, trials investigating the effect of early detection of brain metastases via systematic imaging have failed to demonstrate an effect on survival [5,6], which explains that cur- rent guidelines do not recommend systematic imag- ing to diagnose brain metastases [4]. However, clinicians should have a low threshold for perform- ing diagnostic brain magnetic resonance imaging testing in the setting of any neurologic symptoms suggestive of brain involvement [4]. Of note, these studies have been published more than ten years ago. New trials are currently underway and could call into question current practices.

SYSTEMIC TREATMENTS IN THE FIRST- LINE METASTATIC SETTING

Lapatinib

Lapatinib is a first-generation tyrosine kinase inhib- itor (TKI) that inhibits both EGFR and HER2. Owing to its small size (compared with monoclonal anti- bodies) and supposed better ability to cross the blood– brain barrier (BBB), it has been suggested that it might have a specific effect on brain metastasis prevention and central nervous system (CNS) pro- gression [12,13]. In combination with capecitabine in a population of HER2-positive MBC with previ- ously untreated brain metastases the lapatinib– capecitabine combination was associated with an objective CNS response rate of 66% (LANDSCAPE trial) [14]. Importantly, the median time to CNS progression was six months and the median time to brain radiotherapy was eight months. Despite these results, it has never been shown that the lapatinib-capecitabine combination was superior to trastuzumab-based treatments for brain metasta- sis control. The phase III CEREBEL trial was specifi- cally conducted to address the question of the best CNS protective treatment [15]. Five hundred and forty patients without brain metastases at enrol- ment received either lapatinib–capecitabine, or tras- tuzumab– capecitabine. The aim of the trial was to compare the incidence of CNS metastases as site of first relapse. No difference was reported regarding the primary endpoint (3% of CNS first site of relapse in the lapatinib– capecitabine arm, compared with 5% in the trastuzumab– capecitabine arm). How- ever, patients receiving trastuzumab fared better and had a prolonged progression-free survival (PFS) and overall survival. This result is consistent with the hypothesis that the BBB is altered in case of brain metastases, and, as a consequence, might be permeable to all systemic treatments, even mono- clonal antibodies [16].

Pertuzumab and trastuzumab

Pertuzumab is a monoclonal antibody that inhibits dimerization of HER2, which increases treatment efficacy when combined to trastuzumab and a taxane [17]. This combination is approved in association with docetaxel in the first-line treatment of HER2- positive MBC, following the results of the random- ized phase III placebo-controlled CLEOPATRA trial [17]. In this study, all patients had mandatory brain imaging before inclusion, because brain metastasis (even asymptomatic) was an exclusion criterion. However, 13% of the patients in both arms experi- enced brain metastases as site of first relapse. This population has been described separately in an exploratory analysis, with interesting findings [18]. First, patients receiving pertuzumab had a delayed diagnosis of CNS recurrence, compared with patients with trastuzumab alone (15 months compared with 12 months, hazard ratio ¼ 0.58; P ¼ 0.0049). Second, median overallsurvival was 26 months in the placebo arm, compared with 34 months in the pertuzumab arm, although not statistically significant. Overall, the better outcome of patients with the addition of pertuzumab seems to have a preventive effect on the occurrence of brain metastases. As patients in this substudy did not have extracranial disease progres- sion at the time of brain metastasis diagnosis, the ‘BM prevention’ effect of pertuzumab is most likely because of a direct activity of this molecule on micro- metastatic brain localization that were not diagnosed at inclusion. It is, in fact, striking that the hazard ratio for death in the subgroup of patients developing brain metastases is exactly the same (hazard ratio ¼ 0.66; 95% confidence interval [CI] 0.39– 1.11) that the one for the whole population (hazard ratio ¼ 0.66; 95% CI 0.52–0.84; P ¼ 0008) [18,19].

SYSTEMIC TREATMENTS IN THE SECOND- LINE METASTATIC SETTING

Trastuzumab–emtansine

Trastuzumab–emtansine (T-DM1) is an antibody– drug conjugate (ADC) that has been approved in the second-line treatment of HER2-positive MBC, after failure of trastuzumab and pertuzumab, follow- ing the results of the randomized phase III EMILIA trial [20], where this treatment was compared with lapatinib and capecitabine combination. T-DM1 was associated with a statistical and clinical meaningful improvement of PFS (+3 months, hazard ratio ¼ 0.65; P < 0.001) and overall survival (+6 months, hazard ratio ¼ 0.68; P < 0.001). The presence of brain metas- tases at inclusion was allowed provided they were asymptomatic or treated, and patients experiencing CNS at inclusion have been described in a specific exploratory analysis [21]. For patients without brain metastases at inclusion, CNS progression occurred in 2% of patients treated with T-DM1, compared with 1% ofpatients treated withlapatinib andcapecitabine (not statistically significant). For patients with brain metastases at inclusion, PFS was not different in the two arms (5 months). However, patients inthe T-DM1 arm had a doubling of median overallsurvival (hazard ratio ¼ 0.38; P ¼ 0.008; median overall survival 27 versus 13 months). These results show that T-DM1 has activity on brain metastases comparable to the lapatinib–capecitabine combination and CNS disease evolution should not be taken into account when choosing between these two treatment options. A recent subgroup analysis of the 398 patients with brain metastases at inclusion in the KAMILLA sin- gle-arm open-label, phase IIIb study (NCT01702571) confirms the efficacy and safety of T-DM1 in this situation, with a median PFS of six months, and a median overall survival of 19 months [22].

Trastuzumab–deruxtecan

Another ADC, trastuzumab– deruxtecan (also known as DS-8201a or T-DXd) is currently under investigation in MBC. The payload, deruxtecan, is a topoisomerase 1 inhibitor with a powerful cytotoxic effect, and a short half-life [23,24]. The drug– anti- body ratio is high, and this drug has proven to have a bystander effect [23– 25]. These interesting pharma- cologic properties seem to translate into a clinical benefit, as the DESTINY-Breast 01 trial has reported an impressive overall response rate (ORR) of 61%, and a median duration of response of 15 months, in a population of heavily pretreated patients (all had received T-DM1 and 66% had received pertuzumab) [26]. An exploratory subgroup analysis of this study, describing 24 patients that presented with CNS metastases at baseline showed that this ADC can be active on brain metastases [27]. The pivotal ran- domized phase III study comparing T-DXd and T- DM1 is currently recruiting (DESTINY-Breast03, NCT03529110). Patients with pretreated or asymp- tomatic brain metastases can be enrolled. The DEB- BRAH study, which tests the drug specifically in the HER2-positive brain metastasis population, is about to begin (NCT04420598).

NEW-GENERATION TYROSINE KINASE INHIBITORS

Neratinib

Neratinib is second generation, irreversible, pan- HER TKI [28,29]. In the first-line metastatic setting, the randomized phase III NEfERT-T trial has com- pared the neratinib– paclitaxel combination to the trastuzumab–paclitaxel combination [30]. The pri- mary endpoint was PFS and did not differ between the two arms (13 months each, hazard ratio ¼ 1.02; 95% CI 0.81– 1.27). However, the incidence of CNS recurrence was lower in the neratinib arm (relative risk ¼ 0.48) and time to CNS metastases was delayed (hazard ratio ¼ 0.45), suggestive of a preventive effect on development of brain metastases. With regard to specific activity on brain metastases, the drug was first explored as monotherapy in a phase II single-arm trial of 40 patients with documented CNS progression, after one or more line of CNS- directed therapy [31]. This study was negative for the primary endpoint, with only three patients achieving a partial response. Neratinib was then tested in combination with capecitabine, in a phase II single-arm study [32] for patients with previously treated progressive CNS disease. Neratinib was administered at a dosing of 240 mg orally once per day, in combination with capecitabine 750 mg/m2 twice per day for 14 days, then seven days off. CNS response was assessed with response assessment in neuro-oncology (RANO) criteria [33]. This trial was positive, and the ORR in the CNS was 49% in lapatinib-na¨ıve patients, and 33% in lapati- nib-experienced patients. In the phase III NALA trial [34&], patients who had already received two lines or more of prior HER2-directed systemic regimens for MBC were included. Among 621 patients, 130 (21%) had progressive brain metastases at inclusion. They were randomized to receive lapatinib and capecita- bine or neratinib and capecitabine. PFS was signifi- cantly improved with neratinib and capecitabine, with a hazard ratio of 0.76 (P ¼ 0.006). Of impor- tance, there was a significant decrease in overall cumulative incidence of intervention for symptom- atic CNS disease in the neratinib and capecitabine arm (23% only compared with 29% in the lapatinib and capecitabine arm, mostly salvage WBRT). Toler- ance was poor, with a high frequency of grade 3 diarrhea despite loperamide coprescription (24% in the neratinib arm, compared with 13% in the lapa- tinib arm). Data of the NALA trial are summarized in Table 1. Whether neratinib has a real preventive effect on the occurrence of brain metastases remains on an open question. The brain metastasis analysis of NEfERT-T was an unplanned subgroup analysis, but is in favor of a higher efficacy of neratinib in the brain over trastuzumab, whereas these two mole- cules are equivalent on systemic diseases. On the other hand, the NALA results are consistent with a higher overall efficacy of neratinib over lapatinib, both on brain metastases and systemic disease.

Tucatinib

Tucatinib (formerly ONT-380 or ARRY-380) is a selective oral HER2 inhibitor. This drug has shown promising results as a single agent as soon as in
phase 1, and has a better toxicity profile than dual epidermal growth factor receptor (EGFR)/HER2 TKIs, with lower incidence and severity of diarrhea [35]. Tucatinib has been evaluated in a phase 1b study in combination with capecitabine, or trastu- zumab, or both, in patients with HER2-positive MBC previously treated with trastuzumab, pertuzumab, and T-DM1 [36]. Sixty-one percent of patients (14 among 23) achieved an ORR with the triplet, which is quite exceptional in a phase 1 trial. The triplet was well tolerated: treatment-related toxicities of grade 3 and worse included fatigue in 8% of patients, diar- rhea, and palmar-plantar erythrodysesthesia in 7% of patients each. The combination has been rapidly tested in the phase III HER2CLIMB study [37&]. Patients with HER2-positive MBC were eligible if they had previously received trastuzumab, pertuzu- mab, and T-DM1. Six hundred and twelve patients have been enrolled in the study and received tras- tuzumab (6 mg/kg after a loading dose of 8 mg/kg every three weeks), capecitabine (1000 mg/m2 twice daily, 14 days on, seven days off), and either tuca- tinib (300 mg twice daily) or placebo. Randomiza- tion was 2 to 1. The presence of brain metastases was stratified. 48% of patients in the tucatinib arm had brain metastases at enrolment, and 46% in the placebo arm. The study was positive with an impres- sive benefit in PFS as the hazard ratio was 0.54 (P < 0.001). PFS at one year was 33 in the experimen- tal arm versus 12% in the placebo arm. The study was also positive for survival, with a median overall survival of 22 and 17 months in the tucatinib and placebo arm respectively (hazard ratio for death 0.66; P ¼ 0.005). Patients with brain metastases at inclusion had a worse outcome, but benefited from tucatinib: 25% of them had not progressed at one year in the tucatinib-combination group, whereas all had progressed in the placebo-combination group (hazard ratio ¼ 0.48; P < 0.001). The median duration of PFS was eight months and five months, respectively. Notable safety events included diar- rhea, which was managed with short courses of antidiarrheal agents, and transient, reversible eleva- tions in liver enzyme levels. Data regarding the HER2CLIMB trial are summarized in Table 1. More details regarding the subgroup of 291 patients with brain metastases at inclusion have been published recently, and confirm the impressive activity of tucatinib in the brain, with a 68% reduction in risk of CNS-PFS in the tucatinib arm (hazard ratio ¼ 0.32;
95% CI 0.22– 0.48; P < 0.0001) [38&&]. Median duration of intracranial response is significantly improved with tucatinib (increased from 3 to 7 months). In addition, in this subgroup of patients harboring brain metastases, the risk of death was reduced by 42% in the tucatinib arm versus the
control arm (hazard ratio ¼ 0.58; 95% CI 0.40– 0.85; P ¼ 0.005).

Results of HER2CLIMB are of high interest as it is the largest randomized study with a specific preplaned analysis of medical treatment for brain metastases. Nevertheless, as for neratinib, it mostly shows that improved efficacy of tucatinib in CNS is not different from improvement for systemic dis- ease, with comparable hazard ratios. An interna- tional study, recruiting patients with residual disease after neoadjuvant treatment, will compare standard T-DM1 to the T-DM1 and tucatinib combination (CompassHER2, NCT04457596). It may definitively show that these small TKI are indeed more efficient than ADC to prevent brain metastases.

CONCLUSION

The occurrence of brain metastases is a frequent event in HER2-positive breast cancer. Despite improvement of these patients’ outcome, a majority of them ultimately will die from CNS progression. WBRT after SRS or surgery should no more be rec- ommended as a prevention of subsequent CNS pro- gression but rather used as a salvage treatment in case of neurologic symptoms. The maximum con- trol of metastatic disease remains the best strategy to prevent brain metastases. Indeed, all targeted thera- pies that have demonstrated an improvement in PFS and/or overall survival since the commercialization of trastuzumab, (pertuzumab, trastuzumab– emtansine, neratinib, and tucatinib) have all dem- onstrated an activity on brain metastases, which ultimately delays CNS progression. In fact, these data are in favor of an altered BBB in brain metasta- ses, which leaves few pharmacodynamics differen- ces between brain metastases and systemic disease. If this is accurate for all ADC, important results should come with the ongoing CNS directed trials of new molecule such as trastuzumab–deruxtecan. In the adjuvant setting, new strategies such as the addition of tucatinib to T-DM1 for high-risk patients will definitively address the question of the theoret- ical advantage of TKI over ADC for brain metastases prevention.