I think that there were a lot of very good presentations that were quite helpful to our understanding of treatment and variation in response, and some that were really thought-provoking.
Of the poster discussions, several should be highlighted: one study, TBCRC 030, which was presented by Erica Mayer from DFCI, looked at whether HRD [homologous recombination deficiency score] could predict benefit from weekly paclitaxel versus cisplatin as neoadjuvant therapy for triple-negative breast cancer.1 Interestingly, a high HRD score did not predict relative benefit from cisplatin versus paclitaxel. Very few patients had germline BRCA mutations (only 7 out of the study population), so this was really a study of patients with sporadic TNBC. I think this study tells us, maybe even more definitively than some other studies (like the TNT trial in the metastatic setting2), that HRD for a specific tumor doesn’t help us determine which patients with sporadic triple-negative breast cancer will benefit from platinum therapy.
We know that platinum by itself has some efficacy in the neoadjuvant setting for patients with germline BRCA mutations. Interestingly, although now multiple studies have shown that adding platinum to taxane as neoadjuvant therapy for triple-negative breast cancer (TNBC) improves pCR [pathological complete response], this effect is reduced in patients with BRCA mutations, likely due to increased sensitivity to chemotherapy and resulting higher pCR rates in these patients.3
Indeed, in this population of patients with largely sporadic TNBC, the pCR rate was almost identical between the 2 arms, which is interesting. The investigators tried to analyze various factors to determine whether anything predicted better efficacy of one treatment versus the other, but there was no particular feature that correlated with the likelihood of pCR. Additional correlative studies are ongoing.
Another interesting presentation was the updated data from the CLEOPATRA trial. CLEOPATRA is a phase 3 trial that evaluated the addition of pertuzumab to docetaxel and trastuzumab as first-line therapy for HER2-positive metastatic breast cancer, demonstrating marked improvement in progression-free survival (6.3 months), but an even more dramatic difference in overall survival (15.7 months), although patients on trastuzumab alone were eventually offered pertuzumab.
Very few patients had ever seen trastuzumab in the early-stage setting, so this study population was a relatively unique group of trastuzumab-naive patients.4 This trial led to the European/US approval of pertuzumab in addition to trastuzumab and docetaxel or taxane as first-line treatment for metastatic HER2-positive breast cancer, which has become a new standard of care.
Sandy Swain presented the 8-year follow-up from CLEOPATRA, confirming a continued marked difference in overall survival at 8 years. Docetaxel was continued for a median of 8 cycles, followed by continued antibody therapy until progression. The landmark overall survival at 8 years was 23% in the trastuzumab arm and 37% in the pertuzumab arm, and the median overall survival was 40.8 versus 57.1 months, representing a 16.3-month absolute improvement. The relatively larger impact on [overall] survival than PFS suggests an immune effect from trastuzumab that is enhanced with the addition of pertuzumab.
There were still a number of patients who were on first-line antibody therapy at the time of the 8-year analysis without progression, suggesting that the addition of pertuzumab to trastuzumab and taxane therapy resulted in more patients with long-term survival, despite a diagnosis of metastatic disease. The 50 patients who crossed over to receive pertuzumab were analyzed in the placebo arm.
There were no new toxicities; most of the additional toxicities from adding pertuzumab like diarrhea and neutropenia resolved after docetaxel was discontinued. Overall these are impressive results: both an 8-year landmark overall survival rate of almost 40%, with a lack of additional safety issues — amazing for metastatic incurable breast cancer.
There was another really interesting presentation on HER2 heterogeneity presented by Otto Metzger. Patients receiving T-DM1 and pertuzumab as neoadjuvant therapy had biopsies at baseline from different sites within the tumor to assess the impact of heterogeneity in intratumoral HER2 amplification on the rate of pCR. About 10% of patients met the study definition of HER2 heterogeneity and most of these had hormone receptor-positive disease.
Interestingly, the pCR rate was 0 in this population, compared with 55% in the patients with no heterogeneity.5 There were only 16 patients in the heterogeneous group, so this is a small group of patients. They also showed that there was a lower pCR rate in patients whose tumors were 2+ by IHC compared with those that were 3+, although all tumors had amplification of HER2. I don’t think we can do much with this clinically yet, but this data does possibly identify a group of patients that might benefit from additional or different therapy – perhaps more traditional treatment with chemotherapy plus trastuzumab and pertuzumab – and alternatively, these data identify a population (with HER2 homogeneity) that could be treated with the very well tolerated T-DM1 and pertuzumab combination.
Changing gears to talk about hormone receptor positive disease, we learned a lot about optimal therapy for younger women with hormone receptor positive metastatic disease, Sara Hurvitz presented incredibly important data from the MONALEESA-7 trial. MONALEESA-7 looked at adding ribociclib to either a nonsteroidal AI [aromatase inhibitor] or tamoxifen, versus placebo, as first-line treatment for premenopausal or perimenopausal women with untreated metastatic breast cancer, who were rendered postmenopausal with a GNRA agonist.6
Some patients received chemotherapy before starting study therapy as a sort-of induction therapy, which wasn’t true of the other first-line studies, but I don’t think this really impacted the results. The initial presentations of MONALEESA-7 showed a similar improvement in progression-free survival and a similar hazard ratio compared with all of the first-line studies in postmenopausal women. The presentation at ASCO, published the same day in the New England Journal of Medicine, showed a significant improvement in overall survival – the first time an improvement in survival has been shown in the first-line setting [when] adding a targeted agent to hormone therapy for metastatic hormone receptor-positive breast cancer. Overall survival in patients taking an AI was markedly improved – although the median had not yet been reached in the combination arm, and was about 41 months for the placebo arm, with a 30% relative improvement in overall survival – highly statistically significant.
What’s interesting is, the median follow-up for MONALEESA-7 is shorter than the last median follow-up for PALOMA-2, and MONALEESA-7 was already able to show a survival difference. Clearly in this younger population of patients, who were specifically excluded from the other registration trials, events occur earlier and outcome is overall worse. In the overall population, the hazard ratio was approximately 0.71, indicating a 29% relative improvement in overall survival, which was also statistically significant. A landmark analysis at 42 months showed a Kaplan-Meier estimate of overall survival of 46% in the patients who were on the control arm versus 70% in patients receiving ribociclib.
In MONALEESA-7 compared with the first-line studies in postmenopausal women, there were more women with poor prognostic features compared with their postmenopausal counterparts. About 40% of these women had de novo metastatic disease, and almost 58% had visceral organ involvement; features consistent with bad-risk disease. Perhaps this relative hormone-resistant phenotype was more likely to have a survival benefit seen with a shorter median follow-up compared with older women with more indolent hormone-sensitive disease.
Although the study allowed either an AI or tamoxifen to be the partner endocrine therapy, it is important to keep in mind that you cannot give ribociclib with tamoxifen, because it can be associated with prolongation of QTcF [corrected QT interval by Fridericia] — a unique toxicity with ribociclib that’s not seen with the other CDK4/6 inhibitors. Due to this effect, the investigators looked at the survival impact in the aromatase inhibitor group, and they saw essentially the same effect on survival.
The data from MONALEESA 7 really impact our approach to treating patients with hormone receptor-positive metastatic breast cancer — and they don’t apply to this population. We have seen a survival benefit in a patients treated with hormone therapy plus a targeted agent, and this tells us that we can actually power hormone receptor-positive trials to see a survival benefit. Perhaps in this setting, similar to early-stage disease, one of the keys is to pick a population with a higher event rate, in this case, premenopausal women with less endocrine-sensitive disease. Rather than reducing the effect of adding a CDK 4/6 inhibitor, this strategy rather unexpectedly amplified the survival benefit. Based on these data, as well as data from the other CDK4/6 inhibitor trials, CDK4/6 inhibitors are now an established standard of care in combination with endocrine therapy as first-line therapy for metastatic hormone receptor-positive breast cancer.
Interestingly, the only other survival data come from PALOMA-3, the second- or greater-line trial with palbociclib in post- and pre/perimenopausal women. In this trial there was no survival benefit in the overall population, but in the 79% of patients with hormone-sensitive disease, there was a significant improvement in overall survival, with a hazard ratio similar to that seen in MONALEESA-7 (HR, 0.72). For PFS the hazard ratios are also similar across trials in the same line of therapy, suggesting a potent class effect of CDK4/6 inhibitors.
These survival data clear up one area of controversy – the question was who really needs a CDK4/6 inhibitor, as perhaps patients with very endocrine-sensitive disease could do just as well with endocrine therapy alone. Based on these data, it appears that most patients do benefit from these agents, and based on MONALEESA-7, that first-line therapy is preferred.
Another important trial that impacts clinical practice for patients with hormone receptor-positive metastatic disease is the Young-PEARL study. This randomized phase 2 trial conducted in Korea randomized pre- and perimenopausal women with hormone receptor-positive metastatic breast cancer to receive ovarian suppression with exemestane and palbociclib versus the full FDA-approved dose of capecitabine. The concept behind this study was to address the bias that young women in particular would have a better outcome if they were treated with chemotherapy in the first-line setting. The women could have received one line of chemotherapy (similar to MONALEESA-7) and must have received prior tamoxifen; about half had visceral disease. Remarkably, the hormonal approach resulted in a longer median PFS of 20.1 months, versus 14.4 months in patients who were randomized to capecitabine, as assessed by investigators (HR, 0.659, P =.0469), although response rates were similar. A subgroup analysis showed similar benefit in those with no visceral metastases and those who had not received prior chemotherapy.
Toxicity was greater in patients who were receiving capecitabine. Although there was more [instances of] neutropenia and arthralgias in the hormone combination arm, there was more nausea, diarrhea, and of course, hand-foot syndrome in those receiving capecitabine – toxicities that have a greater impact on quality of life in general. I think this is a really intriguing study, which further validates the value of using hormone therapy combined with CDK4/6 inhibitors even in younger women before giving chemotherapy.
The other presentation of interest focusing on hormone receptor-positive disease, was an interesting study evaluating an AKT inhibitor called capivasertib. The FAKTION trial enrolled patients with metastatic disease progression on an aromatase inhibitor,8 who were then randomized to receive fulvestrant with either capivasertib or placebo. Patients were not required to have tumor mutations in the PI3K/AKT/PTEN pathway in order to be eligible.
In addition to progression on a prior AI, patients were allowed to have had one line of chemotherapy for metastatic breast cancer as well. These patients were not required to have had a CDK4/6 inhibitor, which always makes the results a little bit more difficult to interpret. Capivasertib is given on a 4 days on, 3 days off schedule, and like other AKT inhibitors, primarily cause gastrointestinal toxicity. Patients had received prior endocrine therapy; most of them had 1 line, some 2 or more lines. Activation of the PI3KCA pathway was seen in about 40% of patients, as expected. Treatment with capivasertib was associated with a marked and significant improvement in progression-free survival in this study including 140 patients — 4.8 months on fulvestrant plus placebo, 10.3 months with fulvestrant plus capivasertib — and the response rate was much higher. Improvement in PFS was seen regardless of activation of the PIK3CA/AKT/PTEN pathway, although the numbers become very small.
In the capivasertib arm, 39% of patients required dose reductions because of toxicity, including diarrhea, rash, hyperglycemia, and low-grade infection, and 12% discontinued due to toxicity, so this is an important issue to monitor with capivasertib. The authors did not report how many patients had prior treatment with CDK4/6 inhibitors; [it was] probably very few given the timeline of the study.
These are very interesting data; capivasertib will be tested in a larger phase 3 trial in hormone receptor-positive disease. Of interest, there is a hormone receptor-positive arm of a trial that combines the AKT inhibitor ipatasertib with docetaxel versus placebo plus docetaxel in the first-line chemotherapy setting, so this is clearly an emerging area of great interest, particularly given the activity regardless of activation of the targeted pathway.
In terms of targeting the PIK3CA pathway, alpelisib (an alpha subunit specific inhibitor) is the first approved inhibitor of PI3 kinase, given in combination with fulvestrant in patients with known mutations in PIK3CA in tumor or cell-free DNA, based on data from the SOLAR-1 trial. Due to the timing of approvals, very few patients in SOLAR-1 had previously received CDK4/6 inhibitors. For that reason, we started a study called BYLieve (which is still enrolling),10 where patients with known PI3KCA mutations, progressing after treatment with a CDK4/6 inhibitor, are treated with alpelisib with either letrozole or fulvestrant depending on their prior exposure to endocrine therapy.
We presented data from an interim analysis conducted 6 months after at least 20 patients had been enrolled into the fulvestrant cohort, so these are very early data. Of patients with measurable disease, 70.5% and 59.3% of patients in the fulvestrant and letrozole cohorts, respectively, had a decrease in best percentage change from baseline; the reduction was 30% or higher in 29.6% and 40.7% of patients in the fulvestrant and letrozole cohorts, respectively, despite the fact that many of these patients had already progressed on the partner endocrine therapy.11
Hyperglycemia was the most common reason for dose interruption/adjustment in both cohorts, leading to dose interruption/adjustment in 15 patients (23.4%) in the fulvestrant cohort and 11 patients (30.6%) in the letrozole cohort. Adverse events leading to treatment discontinuation was still a problem, occurring in 16 patients (25.0%) in the fulvestrant cohort and in 5 patients (13.9%) in the letrozole cohort. Rash doesn’t seem to be much of a problem when patients are treated upfront with antihistamines, but hyperglycemia management clearly requires education and experience – in our experience, we have not had patients discontinue once we knew how to monitor and manage glucose. Some studies are evaluating low-carbohydrate diets as a way to manage this class effect of PI3K inhibitors.
The interim data from BYLieve are certainly encouraging, providing data confirming continued efficacy after progression on CDK4/6 inhibitors, and even with relative endocrine resistance. Longer follow-up and evaluation of the full cohort of patients is of course critical.
BYLieve limits enrollment to the patient population who benefitted from alpelisib in SOLAR-1, [or] those with a PIK3CA mutation in tumor. The capivasertib data are intriguing, as the benefit appeared even in those without a mutation. However, the numbers are small, and it is likely that much of the tumor was archival – a larger study is necessary to understand this effect across subsets.
Peter Schmid presented the second interim analysis of the phase 3 IMpassion 130 study, which confirmed the survival benefit (an absolute improvement of 7 months) seen from the addition of the checkpoint inhibitor atezolizumab to nab-paclitaxel, compared with nab-paclitaxel alone in the first-line setting in patients with PD-L1-immune-cell–positive, triple-negative metastatic breast cancer. In this study, 41% of patients had at least 1% expression of PDL-1 on immune cells in the tumor bed, using the SD142 antibody.
There is ongoing discussion about the use of different antibodies for PD-L1 detection and scoring in immune cells versus a composite score including tumor cells. Clearly, additional analyses and results from studies without checkpoint inhibitors such as pembrolizumab will be needed to fully understand this subset of patients. Encouragingly, there were no new safety signals with longer follow-up.
These data are very encouraging, particularly now that we have approval in the US for this combination therapy. However, the treatment benefits less than half of patients with this poor-prognosis disease, and median survival in the atezolizumab group was still only just over 2 years. There are now many studies looking at how to improve or enhance the host immune response with combination therapy, and preliminary data are somewhat encouraging. There are also many studies in the neoadjuvant to adjuvant, and adjuvant settings. These trials, primarily evaluating atezolizumab and pembrolizumab, may have data as early as 2020.
In HER2-positive disease, there were 3 different trials that I think are important to discuss. The first is a phase 3 trial evaluating the efficacy of an Fc-gamma engineered HER2 antibody called margetuximab in patients with heavily pretreated HER2-positive metastatic breast cancer. Margetuximab has been engineered to have higher-affinity binding to the low-affinity FcϒRIIIA receptor CD16A F allele. Patients with the low-affinity F allele may have at least some degree of resistance to trastuzumab, due to poorer binding affinity to the antibody. In a phase 1 trial, 17% of patients with heavily pretreated HER2-positive metastatic breast cancer responded to margetuximab as a single agent, and 3 patients continue on therapy for 4 to 6 years.13 Patients also had evidence of enhanced HER2-specific T- and B-cell responses.
The phase 3 SOPHIA trial randomized patients with advanced HER2-positive disease progressing on trastuzumab and pertuzumab and 1 to 3 prior treatment lines in the metastatic setting to receive either margetuximab or trastuzumab in combination with chemotherapy of physician’s choice.14 Ninety percent of patients had also previously received T-DM1. The first interim analysis was presented, and showed a significant difference in progression-free survival favoring margetuximab, but the difference was only about 1 month, which isn’t clinically important. Clinical benefit rate was also improved with margetuximab.
What was really interesting is a subset analysis evaluating the differential impact of margetuximab in patients with low-affinity FcϒR CD16A F allele compared with those with the higher-affinity V allele. Patients with the F allele had an enhanced benefit in terms of progression-free survival. For overall survival, only 40% of the events needed for the survival analysis had occurred. Even at this early time point, although no difference in survival was seen in the overall group, the patients carrying FcϒR CD16A F alleles had a clinically big difference in overall survival. This is a very early analysis; we look forward to the more definitive second interim analysis, which we hope to have available by the end of the year.
The NALA trial was also presented, which randomized patients with HER2-positive metastatic breast cancer to receive neratinib or lapatinib with both agents combined with capecitabine. Approximately 40% of patients had received prior trastuzumab as their only HER2-targeted therapy, about 40% also received pertuzumab, and about 50% had received T-DM1. This study also showed a difference in progression-free survival favoring neratinib over lapatinib; using a restricted means analysis at 24 months, the absolute difference was 2.2 months, and there was no significant difference in overall survival. The absolute difference in PFS was not presented, but appears to have been smaller.
Of great interest, they showed a difference in time to intervention for brain metastases favoring neratinib. The overall cumulative incidence was 22.8% for neratinib versus 29.2% for lapatinib (P =.043). There was more diarrhea with neratinib than lapatinib with grade 3/4 diarrhea occurring in 24% versus 13%, respectively. In addition, there was moderately more low-grade nausea, vomiting, and decreased appetite.15 This finding is clinically important for our patients who have significant risk of disease in [the] brain, or are at risk for progressing in [the] brain. Understanding the impact of different agents improves treatment choices and effectiveness.
These studies are helping us understand the therapeutic options for HER2-targeted agents in the third- or greater-line setting. We know that we should give trastuzumab/pertuzumab followed by TDM-1, and now we have interesting data with margetuximab and neratinib. Hopefully we will see an update of survival data with margetuximab from the SOPHIA trial in the near future, and we are very much looking forward to data from the phase 3 HER2CLIMB study, which is evaluating a novel HER2-targeted oral tyrosine kinase inhibitor called tucatinib. A phase 1/2 trial generated great interest by showing benefit in CNS disease as well as generally, and this agent appears to have less risk of diarrhea. HER2CLIMB is fully accrued, and we hope to see data early next year, which I think will be very interesting.
The CONTROL study is evaluating the best way to control diarrhea in patients receiving extended adjuvant therapy with neratinib. Several approaches appear to be effective in reducing grade 3 diarrhea, including use of the bile acid binding salt colestipol, and a dose-escalation approach.17
Lastly, for HER2-positive disease, there was a study of pyrotinib, another oral tyrosine kinase inhibitor from China, in a very heterogeneous population of patients with metastatic HER2-positive disease. They randomized patients to receive pyrotinib and capecitabine or capecitabine alone – reflecting current practice in China, which does not include continuing trastuzumab following progression of disease.16
The PHENIX trial randomized women with metastatic HER2-positive breast cancer and a prior taxane to receive pyrotinib or placebo in combination with capecitabine. All patients had previously received trastuzumab, although for about 45%, trastuzumab was given only in the early-stage setting. The investigators showed a marked increase in PFS with pyrotinib versus placebo, at 11.1 versus 4 months, with a hazard ratio of 0.18 and a P value of <.001. Response rates were markedly higher in patients receiving pyrotinib with capecitabine versus capecitabine alone, and patients also seemed to benefit from single-agent pyrotinib after progression on capecitabine alone. There also seemed to be a central nervous system effect; fewer patients in the pyrotinib arm had new or progressive brain metastases. However, like many HER family TKIs, pyrotinib was associated with a significant increase in diarrhea – about 31% had grade 3/4 diarrhea with pyrotinib versus about 13% with capecitabine alone. Interestingly, the incidence of hand-foot syndrome was tripled in the combination arm.
This data demonstrate the effectiveness of pyrotinib in HER2-positive breast cancer, but there is no way to compare this agent to other TKIs, or to evaluate the effectiveness of the combination in patients with greater prior exposure to HER2-targeted agents. Although not available during this trial in China, currently patients in the US receive trastuzumab and pertuzumab with a taxane as first-line therapy, followed by T-DM1 as second-line therapy – both the combination antibodies and T-DM1 have been shown to improve survival over the prior standard of care, and have been incorporated into international guidelines for the treatment of HER2-positive metastatic breast cancer. A trial comparing this combination against a combination with an approved HER2-targeted agent seems a necessary next step, as does improved management of toxicity.
The most important data in the early-stage setting was another analysis of the TAILORx trial,18 which, similar to MONALEESA-7, was published the same day in the New England Journal of Medicine.
Joe Sparano presented an interesting but complex assessment of the impact of clinical risk on the benefit from chemotherapy in, and prognosis of, patients with node-negative intermediate risk recurrence score hormone receptor-positive breast cancer receiving adjuvant endocrine therapy. Prior data had suggested that younger-age women benefitted from chemotherapy at the higher range of intermediate scores, but that older women did not. Not unexpectedly, this analysis showed that clinical risk correlated with prognosis, but did not predict benefit from chemotherapy in older women.
In younger women, the authors hypothesized that ovarian suppression from chemotherapy might have contributed to the benefit from chemotherapy seen in high-clinical-risk, higher-intermediate-score disease. A further subset analysis that showed benefit from chemotherapy in older premenopausal women (more likely to go into menopause with chemotherapy) but no benefit in very young women (who are unlikely to have permanent ovarian suppression from chemotherapy), supported this hypothesis.
It’s important to keep in mind that this is a subset analysis, but the data do fit in well with recent data demonstrating the benefit of ovarian suppression in young women seen in the SOFT and TEXT trials, particularly in those with higher clinical risk. In addition, although clinical risk impacts prognosis, it is not well correlated with the benefit from adjuvant chemotherapy – but the gene-expression score is. Clearly there are patients with node-negative tumors and worse overall prognosis who need alternative therapy to improve outcome. Perhaps the adjuvant trials with cyclin dependent kinase 4/6 inhibitors will demonstrate just this effect. However, for younger women, clinical risk may determine benefit from chemotherapy or more intensive hormone therapy.