Prof. Dummer in Gothenburg
In January 2023, Pierre Fabre arranged a national symposium in Gothenburg, Sweden, focusing on the latest insights into the treatment of BRAF-mutant melanoma. The symposium, moderated by Professor Lars Ny, gathered oncologists, surgeons and pathologists involved in diagnosing and treating malignant melanoma. This report summarises the presentations given.
Professor Lars Ny, Sahlgrenska University Hospital, Gothenburg
Professor Reinhard Dummer, Department of Dermatology, University Hospital of Zürich, Switzerland
Targeted therapy in BRAF-mutated metastatic melanoma
BRAF is one of the most frequently mutated oncogenes in melanoma and a driver oncogene in the MAP kinase pathway. Reinhard Dummer discussed how this pathway can be inhibited.
The Brain Rapidly Accelerated Fibrosarcoma (BRAF) gene, located on chromosome 7q34, encodes serine/threonine protein kinases that play a role in regulating the MAP kinase (MAPK)/ERK signalling pathway. This is one of the principal pathways in cancer, with many outcomes relevant for cancer biology, including proliferation, resistance to apoptosis and survival in the extracellular matrix (Becker JC et al. Cancer 2006; 107: 2317–27). BRAF mutations are the most frequently identified mutations in melanocytic lesions (Hoshino R et al. Oncogene 1999; 18: 813–22, Davies H et al. Nature 2002; 417: 949–54).
Most oncogenic BRAF mutations are found in just very few locations. Indeed, 98.65% consist of a missense substitution at codon 600 (predominantly V600E) leading to constitutive activation of the MAPK signalling pathway (Castellani G et al. Cancers (Basel) 2023; 15: 4026).
BRAF can also carry germline mutations that lead to so-called RASopathies, a group of genetic syndromes with thin, curly hair and hyperkeratosis on hands and feet, features that may occur as side effects of BRAF inhibitors (Niihori T et al. Nat Genet 2006; 38: 294–6, Siegel DH et al. Br J Dermatol 2011; 164: 521–9). It may seem paradoxical that BRAF inhibitors can cause symptoms that are typical for an activated RAS syndrome, but whether BRAF mutations are good or bad depends on the context. In amelanotic melanoma, tumour growth is highly dependent on BRAF activating the MAPK pathway. In some melanocytic lesions and nevi, on the other hand, mutated BRAF can do the opposite: initially expanding the melanocytic cell pool by speeding up proliferation and later driving senescence (Krauthammer M et al. Nat Genet 2015; 47: 996–1002).
In mutated melanoma cells, BRAF is able to signal as a monomer, and BRAF inhibitors suppress the MAPK pathway by blocking the catalytic domain of this monomer. In wildtype cells, however, BRAF inhibitors favour the dimerization of BRAF, thereby paradoxically activating the pathway and causing various adverse events during monotherapy (Rinderknecht JD et al. PLoS One 2013; 8: e58721).
All cutaneous toxicities of BRAF inhibitors are signs of paradoxical MAPK pathway activation. To avoid this, BRAF inhibitors are no longer used in monotherapy; combined BRAF and MEK inhibition has superior effect and is better tolerated (Larkin J et al. N Engl J Med 2014; 371: 1867–76, Long GV et al. N Engl J Med 2014; 371: 1877–88). While BRAF inhibitors cause paradoxical proliferation or activation of the MAPK pathway, MEK inhibitors suppress the pathway in both mutated and normal cells (Dummer R et al. Br J Dermatol 2012; 167: 965–7).
Vemurafenib, the first specific BRAF inhibitor developed for melanoma demonstrated improved survival in patients with BRAF V600E mutation compared to chemotherapy
(Chapman PB et al. N Engl J Med 2011; 364: 2507–16). Shortly after, so did dabrafenib (Hauschild A et al. Lancet 2012; 380: 358–65).
Development of the second-generation BRAF inhibitor encorafenib aimed to improve efficacy and tolerability through modified pharmacological properties. The major difference compared to first-generation inhibitors is its higher affinity for the BRAF protein (Koelblinger P et al. Curr Opin Oncol 2018; 30: 125–33). This is reflected in once-daily dosing (Sullivan RJ et al. Clin Cancer Res 2020; 26: 5102–12).
The COLUMBUS phase III trial demonstrated superior efficacy and tolerability for encorafenib plus the MEK inhibitor binimetinib compared to vemurafenib or encorafenib as monotherapy in patients with BRAFV600E/K-mutant metastatic melanoma (Dummer R et al. Lancet Oncol 2018; 19: 603–15). The difference in progression-free survival (PFS) in COLUMBUS Part 1 and 2 is shown in Figure 1.
The DREAMseq trial demonstrated that combination immunotherapy with nivolumab/ipilimumab, if necessary followed by BRAF/MEK inhibitor targeted therapy, should be the preferred treatment sequence for a large majority of patients (Atkins MB et al. J Clin Oncol 2023; 41: 186–97).
According to ESMO clinical practice guidelines for inoperable stage III–IV BRAF-mutated melanoma, immunotherapy is the first-hand choice in most cases, while combination targeted therapy can be an alternative in certain situations (Michielin O et al. Ann Oncol 2019; 30: 1884–901).
Figure 1. Progression-free survival in the COLUMBUS trial (Ascierto PA et al. J Clin Oncol 2023; 41: 4621–31).
What are the effects of BRAF/MEK inhibitors on brain metastases?
Reinhard Dummer: There is not much data available, but patients with a difficult brain metastasis load, represented by the necessity to use dexamethasone, have huge problems with monotherapy. The response rate to targeted therapy for brain metastases is very high, only slightly lower than for other metastases, but both response duration and PFS are shorter, generally in the range of six months. Based on the current knowledge, we aim to use targeted therapy in order to reduce the corticosteroid load for these patients and then initiate combination immunotherapy. (Davies. Lancet Oncol. 2017 July ; 18(7): 863–873. doi:10.101)
What precautions do you take against cardiac toxicity before starting targeted therapy?
Reinhard Dummer: We do a cardiac assessment every three months and reduce the dose if the ejection rate is below 50%. But as the efficacy is clearly dose dependent, we usually keep the same dose of the MEK inhibitor.
How is the cardiac assessment done?
Reinhard Dummer: By echocardiography every three months and ECG every two weeks in the beginning and later every four weeks. For patients on targeted therapy for more than six months, we try to increase the intervals, typically to six weeks. If the QT interval is prolonged, we check the medication; co-medications can also play a role. We also measure pro-BNP but its reliability is unclear.
Lars Ny: We use proBNP as a screening tool before initiating therapy and only use echocardiography in patients with current or a history of cardiac disease.
If you identify a patient with impaired cardiac function, how do you manage it?
Reinhard Dummer: We start by interrupting the treatment to see if the function improves; if not, there must be another reason for the impairment and the cardiologist takes over.
How about eye examinations?
We do an ophthalmologic assessment at initiation and in case of symptoms.
Dr Hildur Helgadottir, Karolinska University Hospital, Stockholm
Treatment of stage II and III melanoma – BRAF-mutated and wildtype
Hildur Helgadottir reviewed the literature for treating stage II and III melanoma and described how this is done in Sweden.
For stage III melanoma in the adjuvant setting, three trials have been imperative in demonstrating significantly improved relapse-free survival with PD-1 inhibitors: CheckMate 238 (Larkin J et al. Clin Cancer Res 2023; 29: 3352–61), SWOG (Grossmann KF et al. Cancer Discov 2022; 12: 644–53) and KEYNOTE-054 (Eggermont AMM et al. Lancet Oncol 2021; 22: 643–54). Regarding overall survival (OS), however, neither CheckMate 238 nor SWOG showed any significant differences, but part of the explanation for this may be that patients were allowed treatment with immune therapy or targeted therapy at relapse. OS data from KEYNOTE-054 are still eagerly awaited.
In Sweden, PD-1 inhibitors nivolumab and pembrolizumab were approved for adjuvant treatment in stage III melanoma in 2018, and BRAF/MEK inhibitors dabrafenib and trametinib in 2019 (fass.se).
A study designed to assess survival before and after implementation of adjuvant treatment in Sweden was based on the Swedish Melanoma Registry (SweMR). This registry has a near-complete and detailed coverage of primary cutaneous melanomas, including nodal status, but does not cover relapses or oncologic treatments. The study included 1,371 patients diagnosed with stage III primary melanoma between 2016 and 2020, divided into a pre- and a post-cohort depending on the time of diagnosis: before or from July 2018.
In the first report, patients had been followed up until the end of 2021, corresponding to just over two years in the post-cohort (Helgadottir H et al. J Natl Cancer Inst 2023; 115: 1077–84). Results showed no significant differences in neither melanoma-specific survival nor OS. One possible explanation may be that the follow-up time is too short, another that the analysis does not take into account which patients were actually given adjuvant therapy, only that adjuvant therapy was used on a broad scale in the post-introduction cohort. Further analyses will hopefully give the answer.
According to current Swedish recommendations, stage III patients with clinically detected, operable metastases are offered neoadjuvant treatment with a PD-1 inhibitor. Sentinel node-positive, BRAF V600-mutated patients are increasingly treated with BRAF/MEK inhibitors, while a PD-1 inhibitor is still recommended for those with BRAF V600 wildtype (cancercentrum.se).
In the stage II setting, the KEYNOTE-716 trial investigated the effects of pembrolizumab as adjuvant therapy in resected stage IIB/C melanoma. After 27 months, a significant improvement in distant metastasis-free survival and continued reduction in the risk of recurrence was seen compared to placebo, with the best results obtained in the T4a subgroup (Long GV et al. Lancet Oncol 2022; 23: 1378–88). The CheckMate 76K trial compared adjuvant nivolumab to placebo in resected stage IIB/C patients. At a median follow-up time of 16 months, the treatment significantly improved recurrence-free survival across different subgroups (Kirkwood JM et al. Nat Med 2023; 29: 2835–43). Based on these results, PD-1 inhibitors were approved by EMA for treatment of resected stage IIB/C patients and reimbursed in many European countries, although not in Sweden.
In the COLUMBUS-AD trial, BRAF/MEK inhibitor adjuvant therapy with encorafenib/binimetinib in patients with fully resected IIB/C melanoma was compared to placebo. In Sweden, recruitment started in June 2022, shortly after EMA’s approval of PD-1 inhibitors for treatment of resected stage IIB/C patients. Because of this approval, many centres in Europe considered it unethical to allocate patients to placebo, and recruitment diminished. An additional problem was a higher frequency of BRAF-negative patients than expected. For example, only four of the 14 patients screened in Stockholm were BRAF-positive. As a result of these difficulties, the company decided to close the study.
In Sweden, neither BRAF/MEK inhibitors nor PD-1 inhibitors are as yet implemented in stage II disease. OS data for both stage III and II are keenly awaited. There are also ongoing discussions regarding whether these therapies should be implemented based on prognostic tools.
Is it necessary to treat all patients in the adjuvant setting, or can we sometimes wait for the first local relapse?
Reinhard Dummer: As oncologists, we generally aim to improve overall survival, while patients generally aim to stay relapse-free; for patients, every relapse is a serious event that can turn their life completely upside down. We can wait, but relapse-free survival is also an important goal, and some patients progress in a way that renders them ineligible for the neoadjuvant approach.
When will overall survival data from the COMBI-AD trial be presented?
Reinhard Dummer: It will be presented at ASCO this year. The analysis has been delayed due to fewer survival events than expected. This is of course positive, but not when evaluating the study. We don’t know whether this is due to better management or to the fact that all patients who relapse get second-line treatment.
Professor Reinhard Dummer, Department of Dermatology, University Hospital of Zürich, Switzerland
Biomarkers in melanoma
By integrating molecular findings with the clinical picture, we can improve diagnostic accuracy and avoid giving the wrong treatment. Reinhard Dummer discussed how this can be achieved.
Biomarkers can be of value throughout the entire patient journey. They can be used for risk assessment and screening before diagnosis and then contribute to subtype definition, staging, grading and therapy selection. Later, they can be used to select additional therapy and monitor for recurrent disease.
COMBI-AD identifies two immunologically important features to identify patients who might benefit from targeted therapy: tumour mutational burden and IFNγ gene expression as a sign of inflammation (Dummer R et al. Lancet Oncol 2020; 21: 358–72). When these are combined, four different profiles emerge that can be used to predict response (Cristescu R et al. Science 2018; 362: eaar3593). Their impact on treatment outcome is shown in Figure 2.
In patients with a strong IFNγ response and low tumour mutational burden, targeted therapy can disturb the tumour microenvironment so that the pre-existing tumour response will take over when treatment is discontinued, leading to cure or at least long-term control. In patients with a high tumour mutational burden and low IFNγ, on the other hand, the targeted therapy completely controls proliferation and if treatment is discontinued, this control is immediately lost. The hypothesis based on these finding is that the latter group needs long-term treatment.
Figure 2. Subgroup stratification by baseline tumour mutational burden (TMB) and IFNγ gene signature reveals distinct patterns of clinical benefit with targeted therapy (Dummer R et al. Lancet Oncol 2020; 21: 358–72).
The COLUMBUS trial investigated the effects of encorafenib plus vemurafenib compared to vemurafenib monotherapy in patients with BRAF-mutant melanoma. Interestingly, the effect of combination therapy on both PFS and OS was greater in patients with high tumour mutational burden (Dummer R et al. Lancet Oncol 2018; 19: 603–615). The same held true for PD-L1 expression and IFNγ: high levels rendered better results when a MEK inhibitor was added. The conclusion was that MEK inhibitors exert a much better effect in microenvironments that favour immune responses.
A study investigating the transcriptional states that allow tumour cells to survive in the presence of BRAF/MEK inhibition identified eight transcriptional subtypes of melanoma, three of which are associated with resistance states (Rambow F et al. Cell 2018;174: 843–55). One of these is associated with what in vitro may be called a mesenchymal phenotype. The metabolic alterations elevate baseline reactive oxygen species (ROS) levels. When exposed to a ROS inducer, the cells regain their sensitivity to MEK inhibition (Eichhoff OM et al. Cancer Res 2023; 83: 1128–46).
The markers predictive for BRAF/MEK inhibitors are also predictive for immunotherapy, so how can they help us choose?
Reinhard Dummer: We have no specific markers to discriminate between the two. But it’s intriguing that patients in the adjuvant setting with low IFNγ and high tumour mutational burden seem to have a time-restricted benefit, as long as they are exposed to the treatment. Whereas for the other group, you could argue that the targeted therapy helps the immune system to establish protective T-cell responses. So there may be factors that will help us distinguish between the two, but we need to understand more about tumour heterogeneity and how different cell populations contribute to the outcome depending on the treatment interactions applied.
Have you introduced ctDNA as a tool for cutaneous melanoma?
Reinhard Dummer: Not for measuring treatment response, but we have a clinical trial on follow-up of melanoma patients that aims to compare imaging techniques to ctDNA. The goal is to only do blood tests, with CT scans triggered by changes in ctDNA.
Lars Ny: We have a similar study in patients with stage III disease.
Why did the translational analysis from COLUMBUS only present data for encorafenib with and without the addition of binimetinib? How can you know that the difference between combination therapy and monotherapy is due to the MEK inhibitor and not to the fact that vemurafenib is a different agent?
Reinhard Dummer: Data regarding vemurafenib is not yet available, and the study is based on the assumption that BRAF inhibitors in principle have the same target. Time will tell if this is true. But the major difference between the arms is obviously the MEK inhibitor. This makes sense from an immunological standpoint, because the MEK inhibitor is more relevant for T cells. In the T-cell activating cycle, MEK inhibitors are inhibitory during the priming phase, but in the effector phase they also suppress exhaustion, so that T cells are kept active. This could explain the observation that when there is a preactivated T-cell response in the tumour, the MEK inhibitor keeps it alive.
Dr Ana Carneiro, Skåne University Hospital, Lund, Sweden | Assoc. Professor Hanna Eriksson, Karolinska Institutet/Karolinska University Hospital, Stockholm | Professor Gustav Ullenhag, Uppsala University Hospital | Professor Lars Ny, Sahlgrenska University Hospital, Gothenburg
Challenges in BRAF/MEK inhibition treatment including patient cases
Ana Carneiro presented a young patient with BRAF-mutated melanoma treated with a PD-1 inhibitor. The first pathological evaluation showed non-response, with more than 50% viable tumour. One node had pathological response, one did not. The multidisciplinary conference debated changing to adjuvant treatment with BRAF/MEK inhibition or continuing with the PD-1 inhibitor.
Reinhard Dummer: Monotherapy with a PD-1 inhibitor is a good option in the near-adjuvant setting in patients with a high IFNγ gene signature. If IFNγ is low or not present, PD-1 inhibition is not enough.
Hanna Eriksson presented a patient with BRAF-positive M1d melanoma who had metastases in several organs, including the brain. The patient was unsuccessfully treated with so-called sandwich therapy: encorafenib/binimetinib, followed by ipilimumab/nivolumab, and then dabrafenib/trametinib plus bethamethasone, both at progression. There is uncertainty regarding patient selection, for how long the first targeted therapy should continue, and if high-dose immunotherapy should be considered.
Reinhard Dummer: This patient has very bad chances, so there should be no compromising with the dosage. This sandwich approach is a bit unusual, as dabrafenib after immunotherapy can increase the risk of pyrexia in cytokine release syndrome.
Gustav Ullenhag presented a young melanoma patient who relapsed with lymph node lesions after surgery. The patient was considered inoperable and the tumour was BRAF-mutated. BRAF/MEK inhibitor therapy was given to reduce tumour size but after four months, the patient had achieved complete response and only axillary surgery could be performed. A few months later, bone and lymph node lesions were discovered. BRAF/MEK inhibitor treatment was reintroduced, with almost complete response remaining more than one year after treatment discontinuation. Questions to consider are whether the first treatment period was too long and if treatment should have continued directly after the second surgery.
Reinhard Dummer: This strategy is very risky. The goal for this patient should be cure and although there is extensive disease, medical treatment rather than surgery is the critical issue. Immunotherapy is also an option for continued treatment in this situation.
Lars Ny presented a case with a young, previously healthy patient with V600E mutation-positive melanoma who achieved complete remission on dabrafenib/trametinib and remained so for the six years that the treatment continued but who quickly relapsed when it was stopped. Dabrafenib/trametinib was reintroduced, but the patient relapsed again. The tumour was judged unsuitable for radiotherapy and therefore surgically removed, but the patient soon relapsed. Nivolumab/ipilimumab was started, complete response was achieved after four cycles, and treatment then continued with nivolumab monotherapy. Two years later, the patient has continued complete response and no late-stage toxicity. The question is if and when we can stop treatment for patients progressing on targeted therapy who are switched to nivolumab/ipilimumab and have durable responses. There are reports suggesting such a strategy (Warburton L et al. Sci Rep 2020; 10: 18878, Lee J et al. Eur J Cancer 2023; 179: 87–97).
Reinhard Dummer: Five years is the minimum before considering stopping targeted therapy, after evaluating the site or sites of metastasis and the initial tumour burden. The major enemy in cancer biology is heterogeneity, which is a function of tumour burden.