Dear colleagues,

I was delighted to see you all on Thursday 15 April 2021, for the 17th EADO congress, conjugated this year with the 10th World Congress of Melanoma. The conference was supposed to have been held in Rome’s new La Nuvola congress hall, which was inaugurated in 2016, but today serves as a Covid-19 vaccination centre. This is the second 100% digital EADO congress since the pandemic started, and I noticed that the speakers appeared less frustrated and more accepting of the fact that organising a physical event was still impossible. At times like these, we realise the importance of social contact during previous congresses (getting together with colleagues, discussing ideas informally over coffee, etc.), as well as the purely informational advantages of such events. Luckily, vaccination campaigns allow us to hope that our next event will be a physical one, enabling us to appreciate all the more the contact that we miss so much today.

Patient Advocates
According to the oral communications of Samantha Guild (Richmond, USA), Chiara Puri Purini (Rome, Italiy) and Gillian Nuttall (Manchester, UK)

This morning’s first session was about “patient advocates”.

Although there is no specific term in French, we could perhaps describe them as “expert patients”, or “militant patients”. Overall, I understood that a “patient advocate” is a person considered to be an expert on an illness (often having experienced it themselves or having a loved one affected), and who spends time informing and supporting other patients, often as a member of a non-profit association. So, these certainly exist in France too, one example being the members of the “Vaincre le Mélanome” association. I have my own neologism for this kind of patient, who is highly active and seeks to understand all aspects of the illness. I call them “actients”, to compensate for the unfortunate absence of a medical term for an “active” patient. As Samantha Guild, president of the American association Aim at Melanoma, and her European peers have said, “knowledge is power”. Representatives of “patient advocate” associations now attend international cancerology congresses, such as EADO, ESMO, ASCO, etc. One such representative, Chiara Puri Purni, founder of the Italian association “Melanoma Italia Onlus”, described some emotional experiences, like when she was not permitted to enter an ESMO session because she was “too sensitive to what would be said”, or a decade ago when some Italian patients had to travel to the USA for targeted therapies that were not yet refunded at home... They are aware that, sometimes, the consequences of starting treatment really can change in a matter of days. This is also what I feel in my oncodermatology practice, although no scientific studies have proved that the consequences of delaying the start of treatment can sometimes be as harmful as providing ineffective treatment. These associations are therefore all the more determined to accelerate the accessibility of new treatments, without suffering the delays that are often observed in Europe compared with the USA. “Patient advocates” aim to be involved in all stages of therapeutic innovation for melanomas, from development to clinical trials and the distribution of new treatments.

Their means of progress include websites, with up-to-date inventories of clinical trials in progress in different places, discussions via Facebook or Twitter, poster campaigns in public transport showing “shock images”, notably of serious sunburn, which are an effective means of attracting attention…

Gillian Nuttal spoke about her application to help patients manage their side effects (“My melanoma app”, launched in 2016). 648 people have already registered for the application, 60% of whom are patients with stage III-IV melanomas. Interestingly, patients have reported side effects that doctors had apparently not anticipated, such as “myalgia”, “oedema”, “insomnia”, “anxiety”, “sadness”, “despondency”, etc. Obviously, it is difficult to separate the side effects of the treatment from those of the illness, but these sources of distress certainly do exist and must be taken seriously.
The associations of several countries have grouped together to form “MI-PAC”, the Melanoma International Patient Advocates Coalition, to pool efforts to provide psychological, social and economic help for patients, mediation between patients and healthcare providers, information on patient rights, legal approaches that affect treatments and research, public awareness, notably via sun prevention actions in primary schools, which have the advantage of reaching parents and extended family through the children…

One unexpected benefit of the pandemic has been the organisation of digital meetings, which make it easier for the different member associations to discuss these issues.

Liquid biopsy
According to the oral communications of Helen Rizos (Sydney, Australia), Teofila Caplanusi Seremet (Lausanne, Switzerland) and Renata Varaljai (Essen, Germany)

The morning’s second session was about progress in the field of liquid biopsies, and the dosage of BRAF mutation circulating tumour DNA in particular. This revolution makes the DNA analysis of a tumour much easier. It has long been believed that circulating tumour DNA was only released by apoptotic or necrotic cells, but it appears that it is also actively secreted by the metabolically active tumour, in replication, which is why it is rapidly reduced in response to treatment.

Helen Rizos explained the operation of the digital PCR, which is faster and more sensitive than the high speed sequencing technique (known as “NGS”). For patients with a low tumour load, one way of improving sensitivity to prevent false negatives is to increase the number of mutations sought at the same time by the digital PCR: TERT and CDKN2A as well as BRAF. However, it is important to remember that some mutations of these genes can of course be germinal. The specificity of the BRAFV600E mutation is 100%.

The methods of the future will be INVAR, whose sensitivity is 0.001% (Integration of Variants Reads pipeline, Wan et al., Sci Transl Medicine 2020), and IDES-enhanced CAPP-Seq, with even better sensitivity of 0.0025% (Cancer Personalized Profiling by deep Sequencing, Newman et al., Nat Biotechnology 2016). If rapid results are more important than sensitivity, Lausanne’s Teofila Caplanusi Seremet, used to validate the Idylla platform already widely used for somatic tests, provides the BRAF result in circulating tumour DNA in just 1ml plasma within 2 hours.

Renata Varaljai presented the results of her article published in JCO PO in 2019. In her German cohort, circulating tumour DNA dosage (combining BRAF, NRAS and TERT) proved superior to the single dosage of LDH and PS100. Its variations preceded radiological responses by 1.5 months and radiological progression by an average of 3 months. However, circulating tumour DNA is often negative in blood in the case of metastases limited to the central nervous system (advantage of dosage in CSF) or strictly cutaneous (advantage of dosage directly in the ulcerated tumour rinsing liquid).

Circulating tumour DNA also offers a prognostic advantage, both before and after lymph node dissection.

For the 20% of patients with no BRAF, NRAS or TERT mutation, she proposes simply dosing the circulating free DNA (cfDNA), as described in her article published in JEADV in 2020.

Then came the plenary sessions.

Systemic versus intralesional immunotherapy – different biology?

How immune cells kill tumor cells
According to the oral communication of Thomas Gajewski (Chicago, USA)

Thomas Gajewski spoke about the mechanisms of resistance to immunotherapy in melanoma, explaining the differences between an inflammatory micro-environment (“T-cell inflamed”) and an immune evasion micro-environment (“non-T-cell inflamed”). A micro-environment favourable to the T immune response is characterised by a strong CD8+ marker, the presence of CXCL9 and CXCL10 type cytokines, a type 1 interferon signature and, most importantly, the presence of Batf3 dendritic cells. These dendritic cells are involved in both the initiator phase and the effector phase of the anti- tumour response. They activate the T cells, causing the synergy effect with the “brake release” of anti PD-1 treatments.

The tumour cell itself may also affect this micro-environment through beta-catenin activation or the double loss of PTEN. Thus, when the tumour activates its beta-catenin pathway, the Batf3 dendritic cells are not recruited in the tumour: in animal models, this causes resistance to all immmunotherapy, vaccination, and even cell therapy based on TILs (tumour infiltrating lymphocytes). He described a case of secondary resistance to immunotherapy in humans with acquired activation of the beta- catenin pathway.

Using a CRISPR screening method (elective gene extinction), he also identified DECR2 as an immunotherapy resistance gene: if this gene is extinguished, the tumour becomes resistant. This gene codes for a protein involved in the peroxidation of the membrane lipids required for ferroptosis, a recently described kind of cell death, which itself is induced in response to CD8+ lymphocytes. (Weimin Wang et al., Nature 2019).

Finally, the patient’s constitutional genetics may also be involved in the immune response, with the recent discovery of the impact of variants of the PKCdelta gene (Cron et al., article in preparation).

Checkpoint blockade therapy: focusing on combinations to improve outcomes
According to the oral communication of Jedd Wolchok (New York, USA)

Jedd Wolchok joking described the world-wide practice of “zoom” meetings as a mechanism of human resistance to SARS CoV2. More seriously, he reviewed the history of immunotherapy, from anti CTLA-4 to anti PD-1, to present day therapies with the triple combination atezolizumab + anti BRAF + anti MEK, which has been proposed in the USA since 2020. For the future, he believes in the potential of OX40 agonists as a means of overcoming the heterogeneity of antigen expression (particularly Daniel Hirschhorn’s work on murine models).

Microbiome transplantations versus diet
According to the oral communication of Jennifer Wargo (Houston, USA)

Finally, Jennifer Wargo spoke about extrinsic genome factors, such as internal and external exposome, which affect health in general and the response to immunotherapy in particular. In agreement with her accidental observation a few years previously that infection of cells in vitro by mycoplasma could make them resistant to chemotherapy, it appears that microbes are often detected in melanomas (however, I would like to know if these were cutaneous samples, in which case this is not surprising, or internal samples).

It has often been demonstrated that a diverse microbiome provides the best prognosis, with better tumour responses observed in melanoma patients. To diversify the microbiome, clinical tests involving faecal transplantation from complete response donors to patients in progression showed spectacular results in a small population (Baruch et al., Science 2020; Davar et al., Science 2021: 15 in progression under anti PD-1, 7 donors: 6 patients showed clinical benefit). Even immunotherapy- related autoimmune colitis appears to be treatable by faecal transplantation! (Wang et al, Nature Medicine 2018). However, care must be taken regarding the proven risk of transmitting multi-resistant bacteria during faecal transplantation.

Although there is no consensus on the damaging effects of antibiotics, Dr. Wargo insisted on the benefits of a high-fibre diet to diversify the intestinal flora and significantly increase the chance of responding to anti PD-1. Having conducted encouraging observational studies, Dr. Wargo has started interventional studies on diet modification in stable patients under anti PD-1.

And although it may sound like a joke, it appears that smart toilets, which perform a bacterial analysis of our faeces, sending real-time information on our flora to our smartphones, will soon be available! It is never too late to start eating more fibre…

At the end of the morning, there was a session on the effects of the pandemic for melanoma patients.

Managing advanced BRAF-mutant melanoma during unprecedented circumstances

Patient perspective
According to the oral communication of Violeta Astratinei (Romania)

It was another “patient advocate”, Violeta Astratinei, who spoke on this topic.
I think that congresses are moving in the right direction with this openness of mind, allowing more room for patients to express themselves. This patient was asked to speak on behalf of her association about how she felt about the disease during the current pandemic. She regrets the reduced access to clinical trials, delayed surgery, postponed follow-up examinations, etc.

Some patients have attempted to hide viral symptoms (temperature, cough, etc.) for fear of having their cancer treatments cancelled.

Eastern and Central Europe is less equipped (both legally and technically) for remote consultations.

Exchange of views: COVID-19 pandemic experience – an EU and US perspective
According to the oral communication of Hussein A. Tawbi (Houston, USA) & Paolo Ascierto (Naples, Italy)

In hospitals: Paolo Ascierto from Milan was deeply affected: his priority was to protect both patients and carers. As the months passed, he learned, as we all did, that patients under immunotherapy or targeted therapy were not exposed to any particular additional risk related to coronavirus. He participated in the multi-disciplinary expert consensus to draw up European recommendations, accessible to everyone, concerning cancer treatment during the pandemic (Curigliano et al., Annals of Oncology 2020).

Supported by the Society for Immunotherapy of Cancer, he believes that targeting IL6 could be an interesting approach to treating serious forms of Covid-19.

Hussein Tawbi also described his experience of the pandemic in the USA, which was somewhat different from Europe: over there, remote medicine and telephone consultations prevailed, depending on the laws applicable in the different states. Delayed surgery, notably for lymph node dissection, sometimes resulted in neoadjuvant therapy. When surgery took place, adjuvant treatments were cancelled if the risk of relapse was <50%. The pandemic also resulted in the authorisation to leave 6 weeks between pembrolizumab injections being accelerated in the US. Finally, for the metastatic phase, they preferred anti PD-1 over double immunotherapy, except in the case of non-BRAF- mutated brain metastases.

For patient vaccination, it is recommended to leave at least 3 days between immunotherapy injections and the vaccination. The discussion revealed real differences in terms of access to vaccinations, even within the same country, for the same illness.

Caroline Robert reported that her institute (Gustave Roussy) proposed the vaccine to all its patients, regardless of the stage of illness, which is certainly not the case throughout France. Regarding vaccination for healthcare workers, Dr. Tawbi acknowledged that in his centre (MD Anderson Cancer Center), he was lucky enough to be able to vaccinate all healthcare workers and their partners at the same time.

Next came a symposium on the new MelaGenix prognostic test:

MelaGenix prognostic test

A gene expression profiling test created and clinically validated (on stage I to III melanomas) by the German company, Neracare, in partnership with academic research labs. This statistically significant test to predict melanoma specific survival will soon be available commercially.

The test is based on the expression of 8 prognostic genes among the 11 correlated with survival according to the publication by Brunner et al., 2008: KRT9, DCD, PIP, SCGB1D2, SCGB2A2, COL6A6, GBP4, KLHL41. After a few mathematical adjustments, the expression of each gene enables classification of the patients from high to low risk of relapse, independently of the sentinel lymph node. Although adjuvant treatment has proved its benefits for stage III melanomas (except for stage IIIa<0.3mm, for which the prognosis is very good anyway), some patients receive this treatment unnecessarily (most likely stage IIIa patients) and all of them suffer adverse effects.

Conversely, adjuvant treatment is not proposed for stage II melanomas, even though the prognosis for IIc stage patients is far more pessimistic than for stage IIIa. Melagenix thus aims to improve the personalised dimension of the benefit-risk ratio for each individual with respect to adjuvant treatments.

Dirk Schadendorf has thus launched a multicentre clinical trial in Germany to test Melagenix on stage II melanomas, with 2:1 randomisation for adjuvant nivolumab treatment for patients classed as high risk by Melagenix. 130 patients have already been included.

Finally, as for any self-respecting congress in 2021, the topic of artificial intelligence had to be covered.

Artificial intelligence in medical oncology, pathology and radiology
According to the oral communications of Friedrich Overkamp (Berlin, Germany), Alexander Radbruch (Bonn, Germany), Titus Brinker (Heidelberg, Germany), Viktor H. Kölzer (Zurich, Switzerland) and Olivier Michielin (Lausanne, Switzerland)

Overall, artificial intelligence, along with deep and machine learning, has made a strong début in the fields of radiology and anatomic pathology. It is also entering oncodermatology, and all the speakers agreed that there is no point in resisting, we must accept and adapt to it.
The examples of new applications presented today were:

• Possible differentiation of benign or malignant polyps based on coloscopy images, with 98% precision (F. Overkamp)
• Automatic generation of MRI images with gadolinium, with no prior injection of gadolinium (A. Radbruch, neuroradiologist)
• Histological distinction between melanoma and naevus, where artificial intelligence beats anatomic pathologists (T. Brinker et al., JAAD 2021), and the capacity to predict the BRAF or NRAS status of melanomas based on their anatomic-pathology images (Kim et al., preprint)
• Classification of melanoma tumours according to their density of CD8+ T-cell infiltrate (hot vs. cold tumour, V. Kölzer)

And that was it for today! I hope that you found these topics as interesting as I did. I found this first day rich in terms of both societal and biological information. My fellow reporters, Oriol YELAMOS PENA and CONFORTI will take it from here.
Take care of yourselves!

Dermoscopy in melanoma management

Does dermoscopy save lives?
According to the oral communication of Harald Kittler (Vienna, Austria)

Dermoscopy is a non-invasive tool, useful for the early diagnosis of melanoma.
Early diagnosis of melanoma is based on the use of dermoscopy but also on the adequate dermatopathologic study of suspicious lesions and on mass screening.
The interesting question is whether dermoscopy and screening save lives: diagnosing melanoma at a very early stage means to solve the problem almost 100% in some cases, in others, it implies to perform a close follow-up. Late diagnosis, on the other hand, would imply for the patient the need for more invasive treatments and therapies. However, only future studies will show what is the risk of metastasis between early and late excised melanomas, although the metastasis rate is higher for late diagnosed stages.

Clinical clues for melanoma diagnosis
According to the oral communication of Iris Zalaudek (Trieste, Italy)

Several approaches can be used to diagnose melanoma early: the analytical approach (i.e., observing the lesion and understanding whether it is atypical) and the comparative approach (i.e., comparing lesions with each other to understand which is different from the others). The comparative approach can be summarized in the ''ugly duckling'' rule: the lesion that is dermatoscopically different from the others should be excised. In other cases, the dermoscopy is less important than the clinical evaluation because there are cases in which the dermoscopy examination is not diriment but the lesion is single, that is present in a patient without moles but who has recently noted a growing mole; in this case it must be removed in the suspicion of melanoma.

Another important rule is that flat (''junctional'') moles on the face in the elderly do not exist, so any flat pigmented lesion in the face of the elderly should be analyzed for strong suspicion of lentigo melanoma maligna.

Regarding nail melanoma, atypical criteria are the presence of gray and black color, granularity, and a band involving 2/3 of the nail plate.

In addition, to increase the ability to diagnose melanoma, it is always important to combine clinical and dermatoscopic informations.

Role of dermoscopy before and after treatment of melanoma and non–melanoma skin cancers
According to the oral communication of Zoe Apalla (Thessaloniki, Greece)

Dermoscopy can be useful in deciding where to perform skin biopsy in certain conditions for example in lentigo melanoma maligna or in lesions called ''collision lesions''.

In addition, dermoscopy is useful in assessing the extent of the tumor or even in post-surgical monitoring. In fact, the use of dermoscopy allows to distinguish a recurrent nevus from a recurrence of melanoma. In basal cell carcinoma, dermoscopy allows to distinguish superficial from deep forms; this is important to understand whether to use topical therapies (cryotherapy, imiquimod) or to resort immediately to surgery. Moreover, if the basal cell carcinoma is pigmented, it does not respond to photodynamic therapy; the rule is that if blue ovoid nests are observed, it will be necessary to surgically remove it.

The use of dermatoscopic examination is also useful in assessing the response to topical therapies for basal cell carcinoma and the risk of its recurrence.

The last application is for squamous cell carcinoma: the presence of white circles and keratin is indicative of well-differentiated forms while poorly differentiated forms show ulcerated areas with very intense red color but without pathognomonic features.

e-Health in the oncologic patients care
According to the oral communication of Josep Malvehy (Barcelona, Spain)

The term E-Health refers to the practice of medicine supported by mobile devices. In the field of skin cancer, they can be used for different purposes including sharing with other colleagues (oncologists, surgeons) the images of diagnosed tumors or to perform a triage of lesions that require a specialist evaluation. In addition, there are many Apps for the evaluation of moles in which to upload on a platform the clinical photo and the location of a mole where, through an algorithm, is expressed the
risk of that lesion although there is still a specificity and sensitivity not very high; instead perform televisits for the monitoring of some lesions or for some therapies can be useful to avoid overcrowding of hospitals and in fact much teledermatology has been performed during the lockdown caused by Covid.

In conclusion, with technology further evolving, new and updated versions of apps, will inevitably come to play in skin cancer care in the near future but, before integrating them into daily practice, safety and efficacy need to be proven. Therefore, regulations integrating scientific research and validation in real-life situations are urgently needed to provide consumers with reliable information. 

New druggable targets in …

… keratinocyte skin cancer
According to the oral communication of Paolo Ascierto (Naples, Italy)

Cutaneous squamous cell carcinoma is frequent, it accounts for 20% of non-melanoma skin cancers, but advanced SCC is not common → it is an uncommon presentation of a common disease.

Advanced SCC includes metastatic SCC and irresectable SCC and their treatment has changed recently.

In the past, only radiotherapy and chemotherapy were used and did not show much clinical improvement. Later, for a select group of patients EGFR inhibitors were used with response rates (RR) around 25-35%. The revolution has been the introduction of antiPD1 in SCC.

Why can we use antiPD1 in SCC? Because SCC has a high mutational burden which it is due to high UV exposure which induces multiple mutations. In fact, cutaneous SCC has a higher mutation burden than melanoma or lung cancer, meaning that SCC may respond to immunotherapy.

In addition, it has been seen that response to antiPD1 is higher in advanced SCC than localized SCC, probably because they have a higher mutational burden.

Currently there is one approved immunotherapy treatment in SCC: cemiplimab. The objective RR with cemiplimab in advanced SCC is 46% (complete response 16.1%, partial response 30.1%), with a disease control rate of 72.5%. Similar toxicity as other immunotherapies.

However, other antiPD1 has also been studied in SCC:
Pembrolizumab has also been tested in metastatic SCC → RR around 25%
Nivolumab has also been tried in advanced SCC (preliminary work) → 50% partial response Which is the future? There are some trials using immunotherapy in the adjuvant and neoadjuvant setting in advanced SCC.
To sum up, antiPD1 is currently the first line option for advanced SCC in most of the SCC guidelines (NCCN, EADO).

… immunotherapy
According to the oral communication of Dirk Schadendorf (Essen, Germany)

We have multiple immunotherapy drugs that regulate PD1/PD-L1 and CTLA4. However, there are several other immunologic targetable pathways.
One of them imply the metabolic hallmarks of tumors → blockade of IDO → Keynote-252: pembro + placebo vs placebo + IDO (epacadostat) → this study was negative showing no clinical improvement (in vitro it did show improvement but not in the clinical setting).

Other molecules:
HDAC inhibitor (entinostat): can be efficacious when in combination with immunotherapy antiLAG3 (relatilimab): LAG3 is also expressed in the immune synapsis, so blocking it is similar to block PD1 → the idea is to use it in anti-PD1 refractory patients à 18% RR in refractory patients

There are many other antiLAG3 trials using it as first line or also other molecules besides relatilimab, as LAG3 seems the new trendy pathway (the new PD1).
Many other molecules are being investigated in PD1 refractory patients involving cytokines, interleukins, tyrosine kinase inhibitors (TKIs), toll-like receptor agonists...
Lenvatinib: TKI used in PD1 refractory with median OS 13.9, and studies using it as first line but still pending to have its results.

antiCD122: it uses a pegilated IL2: bempegaldesleukin → good preliminary data in first line together with immunotherapy but there is a multicentric study together with antiPD1 that will show real clinical data.

TLR agonist Tolsotolimod: in a trial with ipilimumab has shown negative results (no differences) Oncolytic viruses:
T-VEC: in Feb 3, 2021 there was a press release showing the study did not reach the study primary endpoint so probably TVEC will probably go nowhere since there are other more effective immunotherapies.

Hence, many drugs that worked in vitro did not show any clinical benefit. The challenge is when to select a molecule that worked pre-clinically but fail in clinical studies.

… melanoma
According to the oral communication of Grant McArthur (Melbourne, Australia)\

What is druggable in melanoma?
1) Cell surface molecules
2) Enzymes
3) Small molecule ligand binding protein
4) Other not so druggable apparently: Protein:protein interactions
New technologies: RNA PROTAC
1) Cell surface molecules → these are the typical molecules targeted in immunotherapy such as PD1, CTLA4, LAG3...
2) Enzymes Protein kinases
Targeting wild type RAFs / RAF Dimers: novel RAF inhibitors that include lifirafenib and
belvarafenib → they may be indicated in NRAS mutated melanmas, melanoma refractory to first gen BRAFi , non-V600 mutant melanomas
CDK4/6 inhibitors: approved are palbociclib, ribocclib, abemaciclib → they increase immune memory and decrease immune suppression.
Epigenetic modifiers: this is a good approach to make cold tumors become hot tumors EZH2 inhibitor: tazemetostat → in increases the influx of immune cells in the tumor
HDAC inhibitor: entinostat → leads to downregulation of immunosuppressive cell types in the tumor microenvironment → synergy with antiPD1: entinostat quite efficacious when in combo with immunotherapy
Metabolic
Oxidative phosphorilation/mitochondrial complex 1: some trials in brain metastases using research molecules, not yet in the clinical setting
IDO inhibitor: epacadostat → disappointing results in a clinical trial showing no added benefit by using epacadostat
3) Small molecule ligand binding protein
RXR antagonist: there is a trial showing benefit when added to BRAF/MEK inhibition.
4) The not so druggable
Protein:protein interactions: BCL2-family members → venetoclax: still in proof of concept
Targeting RNA: targeting RNA splincing in spinal muscular atrophy → as a proof of concept it could be used in melanoma since one could target RNA
PROTAC: some attempts to drug the proteolysis targeting chimera (PROTAC)

Molecular tumor board for skin cancer

Genomics of acral melanoma: implications for therapy
According to the oral communication of Boris Bastian (San Francisco, USA)

Acral melanomas occur in glabrous skin which has some biologic differences. In addition, acral melanomas do not seem influenced by the sun exposure or the ethnic origin (similar distribution worldwide).

The mutational burden in cutaneous melanomas is very high, whereas for acral melanomas is much lower. In fact acral melanomas, and also mucosal melanomas, have chromosomic amplifications rather than genetic mutations.

Also, these amplifications are found in neighbouring melanocytes around an acral melanoma. These melanocytes are called field cells. This may suggest that these field cells are the actual origin of acral melanomas. This phenomenon can also be seen clinically, since many acral melanomas have multiple macules, but if one could does molecular testing between the macules in apparently healthy skin one could find field cells showing the same genomic aberrations as the primary melanoma.

Genomic rearrangements in acral melanoma have unique characteristics: they have typhonas (numerous of mutations in the areas of chromosomical amplifications), and these typhonas are also present in other tumors such as sarcomas.

Acral melanomas may be due to genomic repair defects which can afects telomer reparation, among other DNA repair genes. In fact, in syndromes with defects in the DNA repair genes, such as Werner syndrome, there’s an increase in acral melanomas.

Also, there are some melanomas located in acral skin that don't have typhonas and may have kinase fusions (NTRK...), or mutations in KIT, NF1, or SPRED1 (SPRED1 mutations in Legius sd) → these are not bona fide acral melanomas, but they may be low-CSD (low chronic sun-damage) cutaneous melanomas located in acral skin. Therefore, a subset of acral melanomas are actually UV-driven.

Hence, there is a tremendous genetic heterogeneity in acral melanoma → this may complicate the therapeutic management of acral melanoma.

How to use high resolution translational data in daily clinic?
According to the oral communication of Reinhard Dummer (Zurich, Switzerland)

The TuPro (tumor profiler) is a group of different technologies developed in Zurich's University Hospital in order to provide multidimensional analysis of a patient's tumor. Multiple diagnostic and therapeutic techniques are used to provide relevant information for the patient management. Which are some of these technologies? Standard pathology, digital pathology, NGS, CyTOF, living tumor cells analysis, scDNA, scRNA, prototyping… Later, a multidisciplinar tumor board analyses the sample, later the results are shown to conventional tumor board and therapeutic decisions are made. To have sense, this process needs to be fast, and needs to be done in less than 4 weeks.

Some of the technologies used in TuPro:

• Digital pathology: the point of using artificial intelligence in pathology is to be objective and identify markers
• Imaging Mass Cytometry: can be used in path slides. It may allow prediction of drug targets.
• Single-cell transcriptomics and genomics integration: these are single cells and multiple studies can be performed → gene aberration profiling, drug target analysis, biomarker discovery...
• Deep Drug platform: document the activity of the different pathways in different cell populations.

Example on how TuPro is used: on a patient with brain mets they analyzed the tissue from a brain met and showed: high mutational burden (potentially good responder for immunotherapy), the tumor had BRAFV600E mut, EGFR amplifications, they tested different targeted therapies but they have found that cabozantinib (MET and VEGFR2 inhibition) → they used this drug and prolonged the survival of the patient.

Learning from early processes in melanoma development: molecules implicates in breaking of senescence
According to the oral communication of Anja Bosserhoff (Erlangen, Germany)

BRAF is present in 90% of all nevi → BRAF activates proliferation but proliferation stops: senescence. Hence, nevi occur because they enter senescence. However, if these melanocytes do not enter senescence they continue to proliferate and then they become a melanoma

What disregulates senescence?

MIA: it positively regulates metastases → potentially drugable. MIA may need additional mutations to become oncogenic.

Role of HuR: HuR is upregulated in melanoma. When one downregulates HuR it reduces melanoma proliferation and an important increase in senescence → HuR can also be a therapeutic target Impact of acidosis in the tumor microenvironment: if you induce an acid pH on the tumor the proliferation does not stop but it's reduced → acidosis induces senescence.

Future research lines:
3D models: many studies use cell colonies which are 2D when it does not really represents the 3D tumor structure → Schmidt et al have developed 3D models printed → 3D biofabrication

Can serum markers guide treatment decisions?
According to the oral communication of Bastian Schilling (Wuerzburg, Germany)

The guidance of these markers needs to be robust, need to have positive results, not having negative results and be reproducible.

Melanoma:
BRAF mutant patients can receive targeted therapies (TT) with BRAF/MEK inhibitors or also immune checkpoint inhibitors (ICI): what to use first? It was thought that LDH would be a good marker: it was thought that high LDH would guide to TT but studies on ICI have shown that ICI also works well in patients with high LDH, so now multiple studies have shown that maybe it's better to start directly with ICI even if there's BRAF mutations. So, LDH does not seem to be a good marker in this scenario.
In patients with BRAF mutant melanoma + brain metastases, maybe it's better TT because we need a fast response and ICI may also may cause important side effects. Or sometimes it’s simply better to do nothing since drugs have side effects and the patient already has a very poor prognosis, so better to do supportive care in patients with brain mets and high LDH (very poor prognosis). Here, LDH tells you not to do harm.

Conclusion: can current serum markers guide melanoma treatment decisions? No, because they are prognostic and/or there is no choice. But serum markers such as LDH can indicate best supportive care when assessing patients with poor prognosis → the therapies sometimes may do harm and if the patient has a very poor prognosis it's better to improve the quality of life.

Adjuvant treatment of melanoma

Adjuvant immunotherapy
According to the oral communication of Jeffrey Weber (New York, USA)

Adjuvant therapies started with ipilimumab at high doses 10mg/kg for 3 years in 2015 which showed clear benefit in the adjuvant setting with 65% overall survival at 5 years (and patients were IIIB and IIIC). However, 21% patients had significant toxicities. This led to approval of adjuvant ipilimumab in the US but not in Europe or Australia.
Subsequent studies showed that the results between 3mg vs 10mg/kg of ipi are similar, so if needed to be used in the adjuvant setting, it is better at lower doses, although it may not be approved at this dose in some countries.

In other studies ipi has been used together with nivolumab but for 1 year since patients may not tolerate 3 years of adjuvant treatment. Checkmate-915: Ipi + Nivo vs nivo study → negative study for RFS but final data needs to be shown, but so far ipi + nivo is not superior than nivo alone.

Pembrolizumab vs placebo for 1 year shows that 59% of relapse free survival at 42 months (less long term data).

Adjuvant targeted therapy (TT)
According to the oral communication of Georgina Long (Sydney, Australia)

Around 50% (46%) melanomas have mutations in BRAF, and among them the majority are V600E
(73%) or V600K (19%) which are targetable.
Adjuvant treatment is performed for 12 months and the only TT approved is dabrametinib 150mg BID + Trametinib 2mg QD → COMBI-AD study: 12 months of treatment results in 50% reduction in risk of recurrence vs placebo.

But the question is if we need to treat with adjuvant therapy or do we wait until the patient relapses and then we treat them? We still do not know.

Biomarkers: Dabra+trame works well in most patients even if they have low tumor burden. The ones that don't respond to dabra + trame are if they have high mutation burden + low IFN.

How do we compare the different adjuvant regimens (ipi, nivo, pembro, dabra+trame)? We don't know so far because the results are similar, meaning they are good with the different regimens.

Future: maybe we may use TT in the neoadjuvant setting since they induce hyalinized fibrosis → hyalinised fibroses correlates with less relapse → good overall survival with good pathological response.

Who needs adjuvant treatment when melanoma can be cured by neo-adjuvant or palliative treatments?
According to the oral communication of Christian Blank (Amsterdam, The Netherlands)

Neoadjuvant therapy may be better than adjuvant because there's going to be a stronger immune response if you don't remove the tumor → more resident T cell → more response to immunotherapy → potentially better survival
Also, another benefit is that in responders the surgery is less aggressive thus improving the patients’ quality of live.

Neoadjuvant may not respond well if low IFN and low tumor burden.

Personalized neoadjuvant immunotherapy: they do multiple analysis to identify the best responders. In the future maybe neoadjuvant therapy will replace adjuvant treatment because one can perform personalized choices of treatment.

Multiple trials ongoing and the results will be available in the next 2 years.

Fundamental questions in dermato-oncology

UV induced DNA-damage in skin cancer: mechanisms and significance
According to the oral communication of Richard Marais (Manchester, United Kingdom)

What drives melanoma? Classic melanoma classification based on UV considers that one can divide melanomas in UV-driven and not-UV driven (sheltered regions such as mucosal areas):

• Cutaneous melanoma on hairy skin is UV-driven.
• Acral melanoma 50% is driven by UV light and the 50% not.
• Non-UV driven melanomas: Mucosal melanoma, uveal, conjuctival melanomas

But the whole situation is more complex. Different types of melanomas harbor different genetic mutations (their background):

• Cutaneous melanoma: mainly BRAF
• Mucosal melanoma: mainly NRAS (large genome changes)
• Uveal melanoma: mainly GNAQ and GNA11
• Acral melanoma and conjuctival melanomas: we still we don't know well their background and we are learning from them now.

However, we know that not only the background is important, but also the environment such as the UV radiation. Hence, BRAF and UV light may cooperate to induce UV-induced melanoma (Viros et al. Nature 2014), and actually 85% of melanomas are induced by UV radiation.

Alternatively, in the mucosa typically there’s no sun exposure, although we will see that some mucosae can get sun exposure. What happens in the eye? Since the conjunctiva, which is a mucosa, is in direct contact with UV light, the hypothesis is that they may share some aspects with cutaneous melanomas (which are UV-driven):

• conjunctival melanomas share the same signatures as cutaneous melanomas (signature 7, which is related UV-B radiation), but they still have some aspects related to mucosal melanomas (since the conjuctiva is a mucosa)
• something similar happens with the uvea; when you look at the uveal melanomas located on the iris, which is also exposed to the sun, they also have signature 7 which is due to UV light.

What is relevant is the melanocyte microenvironment (underlying mechanism, background) + the UV effect which accelerates the melanomagenesis:

• Cutaneous melanoma: underlying mechanism → BRAF
• Conjunctival melanoma: underlying mechanism → chromosomal aberrations Iris melanoma: underlying mechanism → GNAQ/GNA11

Hence, the melanomas on the conjuctiva and iris are actually UV-driven, highlight the need to protect the eyes from UV radiation!

Cure of metastatic melanoma on the horizon?
According to the oral communication of Georgina Long (Sydney, Australia)

Today we have therapies that increase survival but still there is still a lot of work to cure melanoma. Initially patients in BRAFi/MEKi do very well but later ipi+nivo is the best (but more toxicity).

It seems that patients that do best are the ones that have:

 

• complete response
• low volume disease
• normal baseline LDH
• low number of metastatic sites

All these aspects may be useful to determine the risk of a given patient, and actually there will be risk calculator that will be available soon at www.melanomarisk.org.au
What do we know from the last decade? We tried to develop biomarkers but many don’t really correlate well. So far, only tumor mutation burden (TMB) + IFNgamma together seem to correlate with treatment response. Another aspect that has been studied in the last decade is the concept of making tumors that do not respond to immunotherapy more immunogenic. In other words, making cold tumors hot. But has this approach worked? So far the answer is no:

• this failed in attempts such as IDO blockade which was meant to make a tumor hot: Epacadostat did not work by making the tumor hot.
• It also happened with the T-VEC trials which had negative results.

BRAFi seem to also make tumors hot, so the new idea now is to use triple therapy (BRAFi + MEKi + antiPD1) → in the clinical trials there's a trend in improvement but many studies using triplets were negative, so we may need to wait for more data. So where could the role of triplets be? In cases who are BRAF mutant who progress to immunotherapy.

Ongoing lines of treatment include:

• LAG3 inhibitors: relatlimab → positive trial for PFS
• Multikinase inhibitor: lenvatinib → not really clear results

Current most effective treatment is still ipi + nivo; however, ipi + nivo does not work better than nivo alone in the adjuvant setting.

The problem in immunotherapy is we don't know which ones will respond → translational research is crucial to identify responders:

• many studies are ongoing but still these biomarkers need to be validated
• microbiome may also have an impact

Neoadjuvant therapy → one of the most exciting lines of research in melanoma nowadays. The main advantage of neoadjuvant treatment is that it requires 6 weeks of therapy and later the melanoma is resected. Therefore, research can be done much faster (6 weeks). That is why there are multiple treatments being studied as neoadjuvant agents, because administration of the drug is for 6 weeks, and therefore studies can be done much faster. Also, it’s possible to know if the treatment work by analysing the excised specimen and analysing the pathological response.
But is neoadjuvant better than adjuvant treatment? It is unknown yet but there are some trials ongoing (ie the NADINA trial) that will answer this question.
Summary with current treatments:

• With CPI 5-year overal survival 50%
• Beware of translational studies at identifying responders
• Future probably is neoadjuvant therapy

Covid and cancer
According to the oral communication of Olivier Michielin (Lausanne, Switzerland)

SARS-CoV2 interacts with the enzyme ACE2 through the S protein. The first concern in oncology is if ACE2 is involved in cancer. Not really but we know tobacco interacts with ACE2
Also, some patients do very well to SARS-Cov2 and some don't, so the issue in oncology is that patients are immune-suppressed. Is then COVID worse in oncologic patients? Immune checkpoint inhibitors make COVID worse?

Actually, cancer patients are strongly affected by COVID-19 → immune suppressive, less ICU admission because it's full of COVID patients, less surgical availability...
The CCC-19 consortium has determined poor outcome in cancer patients due to COVID: age >40, male, comorbidities ECOG, hematologic malignancies, progressing cancer.

Immune checkpoint inhibitors: conflicting results since some studies show a detrimental effect if COVID, but other studies cannot confirm this. This may be due to the fact that there are different COVID patient subsets, some with higher levels of cytokines and others without → overall CPI does not have a massive impact on the prognosis of COVID. What maybe one could think twice if needed to use ipi + nivo because then you may need steroids which do have a worse impact if COVID.
And what about vaccination? All cancer patients should be offered vaccination, but the question would be if cancer patients will be able to mount an efficient immune response against SARS-CoV2, but evidence from Influenza vaccines shows that cancer patients can mount a good immune response.

To sum up, we should be careful not to undertreat cancer patients because of the pandemic because patients will not die more from COVID but from their cancers.

Immune checkpoint inhibition in immunocompromised patients

Anti-PD-1/PD-L1 immunotherapy in transplant patients with advanced non melanoma skin cancer
According to the oral communication of Céleste Lebbé (Paris, France)

One of the biggest concerns when using immunotherapy in transplant patients who have cancer is that the immunotherapy may induce a rejection of the transplanted organ. However, in advanced non- melanoma skin cancers if we don’t do something also the patient may die from the cancer. Therefore, using antiPD1 in transplant patients may also result in longer survival:

• For example for SCC, the response rates are around 60%
• What happens with the rejection of the organ when using CPI? It depends on many aspects in which there is an influence on the transplanted organ, the immunosuppressive drug used, risk factors…

Organs:

 

• Kidney: 54.5% rejection when CPI used, less if mTOR inhibitors used as immunosuppresant

Other factors:

• Less rejection if organ transplanted >=8 years, or if at least one drug other than corticosteroids used
• More rejection if prior history of rejection

Also, when there is a graft rejection:

• it tends to occur shortly after initiating CPI (median 22 days), maybe earlier in liver transplant
• high frequency of vascular rejection → vascular rejection can result from both T-cell and antibody- mediated mechanisms

• pure acute T rejection in most cases: 61%
• mixed T-cell and antibody-mediated in 38%

• mortality associated with graft rejection
• worse survival with liver and heart recipients
• better survival in kidney recipients\
• However, the mortality due to the cancer is worse than the mortality due to the transplant rejection (51% vs 10%)

◌◌ Therefore, if the cancer is deadly, use CPI earlier to reduce the cancer mortality, regardless of the possibility to have a graft rejection.

Toxicity associated with CPI is similar to non-transplanted patients.
A suggested scheme (Dana Farber scheme) is the switch of calcineurin inhibitors to mTOR associating corticosteroids (needs prospective studies).
Also, there’s room to identify biomarkers of acute graft rejection: CXCL9, CXCL10…

How to manage anti-PD-1/anti-CTLA-4 immunotherapy in transplant patients with metastatic melanoma
According to the oral communication of Brigitte Dréno (Nantes, France)

Transplant patients have an increased risk for melanoma (2.4-fold) and typically thick melanomas → more melanoma mortality. Also more risk of SCC, 60-250 fold.

Surgery remains the best solution for transplant patients who develop isolated metastases when complete excision with margins is possible. However, the discussion is whether to use immunotherapy in the adjuvant setting because of the risk of organ rejection. In this sense, intralesional drugs may be the solution.

Talimogene laherparpepvec (T-VEC), an oncolytic virus, can potentially be effective and safe in transplanted patients. Also, using the addition of topical imiquimod may be an option.

What happens with other immunotherapies? IFN and IL12 are not recommended in transplant patients. CPI: they increase the tumor killing with risk of graft rejection, mostly vascular rejection. The problem with CPI is that the transplant patients have been excluded from clinical trials and we don’t have data. Thus, the efficacy of CPI in transplant patients comes from retrospective reviews, databases and literature.

One retrospective cohort (Eur J Cancer 2018 Nov) including 16 centers with patients undergoing CPI found that among the 6 patients with solid organ transplant there were 2 graft rejections and 2 partial responses, thus suggesting that the role co CPI needed to be carefully considered.

A review of the literature (Manohar et al. 2020) found 27 articles with 44 patients with cancer (30 melanomas) + CPI + solid organ transplant in which 18 had acute graft rejection and among them 15 (83%) had a complete failure and 8 died, with a median time to rejection of 24 days.

Another retrospective cohort (Abdel-Wahab et al 2029) showed similar results in metastatic melanoma patients but 36% responded to CPI.

So, one would think that it would be good to lower the immunosuppressant regimen before the PD1 inhibitor: a 2019 study in Int J Mol Sci found that actually lowering the immunosuppressants actually increases the risk graft failure in a cohort of renal transplant patients.

Some evidence seems to show that switching the immunosuppressant regiment may be a way to prevent graft rejection (mTOR inhibitors such as sirolimus and anti-VEGF better).

Biomarkers such as PD-L1, although promising, do not seem to be useful to predict response or rejection. However, one option would be to control the expression of PD-L1 in grafts (mainly kidney). Maybe antiCTLA4 may be a better choice for transplant patients, since organ rejection occurs more frequently with antiPD1.

Also, if the patient is BRAF mutant, targeted therapy is the first line treatment in advanced melanoma.

New treatment options in the prevention of skin tumors in organ transplant recipients
According to the oral communication of Eggert Stockfleth (Bochum, Germany)

Cancer is the number 1 cause of death among solid organ transplant (SOT) recipients, and skin cancer is the most common cancer among SOT, so important to patient attention to them → increased cutaneous SCC and melanoma, thus potentially being more aggressive.

Since many SCC arise from actinic keratoses (AKs) which are caused by the UV light, sun protection is crucial in SOT patients to prevent the onset of SCC. Besides sunscreen and physical protection, chemoprevention with nicotinamide or retinoids is a good strategy. Additionally, it is important to treat the cancerization field with 5FU, photodynamic therapy or others to prevent the onset of AK and ultimately SCC.

Another potential origin of SCC in SOT patients is HPV. Therefore, a good strategy to prevent HPV- related SCC in SOT patients is HPV vaccination, since these patients seem to build a good immune response against HPV.

Another potential way to prevent skin tumors in SOT is switching the immunosuppressants to mTOR inhibitors such as sirolimus, even if co-administered with cyclosporine.

Finally, the way to prevent the progression of skin tumors is early detection ideally using dermoscopy.

Promising agents for PD1-refractory patients in skin cancer

RP1, RP2 and RP3: a new oncolytic immunotherapy platform
According to the oral communication of Robert Coffin (Woburn, USA)

Oncolytic viruses are potentially good targets in melanoma because tumor cells often lose or have down-regulated antiviral defence pathways that are active in normal cells.

In addition oncolytic viruses may improve immunotherapy since tumors have immune-evasion mechanisms.

RP1, RP2 and RP3 are oncolytic human herpesvirus modified to carry GALV (a fusogenic protein) to enhance antitumoral immune response:
RP1: fusogenic protein, GM-CSF → provides potent systemic efficacy in large tumors in rats, even having abscopal effect, meaning the response happens not only in injected tumors but also in non- injected tumors → this doesn’t happen with T-VEC that doesn’t have response in non-injected tumors. RP1 in combination with nivolumab in non-melanoma skin cancers has a high response rate (CR/PR) of 72.8%, with many achieving CR (45.5%).

Also, RP1 could be used potentially in PD-1 refractory patients à the concept of making cold tumors hot → this especially interesting because the combo with CPI may benefit from the abscopal effect. RP2: fusogenic protein, GM-CSF, anti-CTLA-4
RP3: fusogenic protein, GM-CSF, CD40L, 4-1BBL

CMP-001, a CpG-A TLR9 agonist packaged in a virus-like particle
According to the oral communication of Art Krieg (Cambridge, USA)

CpG-A DNA potentially stimulates type 1 IFN secretion and induces P1-plasmacytoid dendritic cells (P1-pDC)

CMP-001: CpG-A is a toll-like receptor 9 agonist (TLR9) using as a vector a bacteriophage-derived virus-like particle (VLP), self-assembling nanoparticles.
CMP-001 has some differences from oncolytic viruses (OV):

 

• OV replication is inhibited by high type 1 IFN
• OV are inhibited by the antibody response

CMP-001, also named vidutolimob, has been used in trials with or without anti-PD1 (phase 1/2 trials):
CMP-001 + pembro in anti-PD1 refractory patients → overall response rate of 26% (which is very high for 2nd line treatment!)

Interestingly, CMP-001 also has abscopal effect (response in injected and non-injected lesions)

FixVac (BNT-111): a new melanoma vaccine
According to the oral communication of Patrick Brück (Mainz, Germany)

BNT-111 is a cancer vaccine with shared melanoma antigens which uses as a vector a lipoplex nanoparticle.

Clinical trial using BNT-111 +/- anti-PD1 (phase 1/2 trial)

With this treatment there’s an immune activation that can be visualized by increased metabolic activity in the spleen

Response rate
- BNT-111 alone: 23.8%
- BNT-111 + anti-PD1: 35.3%

Primary prevention of skin cancer

Primary melanoma prevention in Australia: an update
According to the oral communication of David Whiteman (Brisbane, Australia)

Australia has a big problem with melanoma since it causes 14260 new cases a year, being the 3rd most common cancer.

Multiple strategies have been developed in terms of preventing melanoma (primary melanomas). In fact, these strategies started in 1963 with Neville Davis. Multiple campaigns to raise awareness have been developed, but also the development of apps and the banning of tanning beds in 2015.

The longest running campaigns are education campaigns: SunSmart campaigns → they educate on how UV contributes to melanoma, also on how to apply sunscreen (Slip, slop, slap campaign), and also early detection campaign.

There have been regulations also, such as the banning of tanning beds, the regulatory guidelines on sunscreens, sun protective clothing...

Another instrument is legislation: for example, there are some laws that enforce the presence of shaded areas for example in playgrounds.

Also, if someone can prove their skin cancer is due to their occupation, people can claim. Finally, there are some fiscal advantages such as deductions for sunscreen.
But does all this work?

Using surveys, they have seen that after some of the SunSmart campaigns people used more sunscreen and got less sunburnt. So, education works, but this has an impact in the melanoma incidence? If one looks at the melanomas stratified by age, people who were born in the 80-90's seems to have a trend on having less melanomas, so there's hope. The challenge now is to repeat a large SunSmart campaign (which hasn't happened in 10 years) which is a very large and complex campaign.

The sunscreen and vitamin D debate: where are we in 2021?
According to the oral communication of Dagmar Whitaker (Cape Town, South Africa)

Do we need to change our recommendations regarding sun protection? The problem is that there are too many variables: how much sunscreen, how often, which factor, which SPF, which is the best? The problem is that there's no guidelines, therefore there's no consistency. The South African recommendation would be to apply SPF50 and if outdoor activities reapply every 4 hours. The problem is that we are probably reaching the right target: we don't reach the middle-aged people who work outdoors.
Also, maybe we are going too far with sun protection. Maybe we are only focusing on the deleterious effects of sun (skin cancer, eye pterygium...) but we tend to forget the benefits of sun light improves the circadian cycle, prevents the onset of depression, prevents the onset of myopia, increases vitamin D... Who is at risk of vitamin D deficiency? With the current levels we can all be at risk. So, is this really relevant? Do we all get symptoms of vitamin D deficiency? No. But "a little bit of sun is good for you". The problem is how much sun is a little bit of sun. How can we prescribe sun exposure without being deleterious? We should encourage outdoor activities but not in harsh sun.

From primary to secondary prevention: a holistic concept
According to the oral communication of Hans Peter Soyer (Brisbane, Australia)

There has been a rapid increase in the melanoma diagnosis. This is due to an increase of melanoma in situ which can be due to a lower threshold in excising lesions, a lower threshold in diagnosis (even pathologists), among other. This is controversial but it true that overall, the number of melanoma diagnosis increased.

In the future there will be a integrative risk stratification for melanoma, in which automated image collection + genotype information will provide a holistic risk score. So far, we are calculating the risk by using melanoma risk calculators from the Melanoma Institute of Australia.

Artificial intelligence (AI) will help identify changes in total body photography images and also integrate it with the genotypic information. In order to improve the diagnostic accuracy of these AI algorithm is to improve datasets, which are overpopulated with pictures with Caucasians, but they need to include images from other phototypes. Also, many of the AI algorithms use dermoscopic images which are not the real images that can be obtained easily, so it may need to use wide-field conventional images, and some attempts show it is possible and promising. More images coming from 3D full body imaging will help in this sense.

To sum up, primary prevention activities as well as genetic testing and imaging technologies underpinned by AR are rapidly advancing and will ultimately bring us a step closer to a world without melanoma.

Diagnostics

Why dermoscopy?
According to the oral communication of Giuseppe Argenziano (Naples, Italy)

Clinical and dermoscopic morphology go together. It is true that 70% of melanomas are diagnosed with the naked eye, but the 30% remaining no, and for these 30% we need dermoscopy. In fact, dermoscopy helps with 20% of melanomas, and for the remaining 10% is when we need to apply other strategies like rules.

Why using dermoscopy then? Because it has a high sensitivity for melanoma: it makes sense that dermoscopy is useful since melanomas do not start as large, irregular, multiple-colored lesions, they start with subtle changes which can be identified many times with dermoscopy.

Also, dermoscopy is also very useful in patients with multiple ugly melanomas, as well as with amelanotic lesions.

Another reason to use dermoscopy is because it has a high specificity and prevents the excision of many benign lesions. With the clinical analysis one needs to excise 30 benign lesions to identify one melanoma, whereas with dermoscopy it is half of this number.

From ugly duckling to artificial intelligence
According to the oral communication of Jean Jacques Grob (Marseille, France)

Multiple publications have tried to compare human intelligence and artificial intelligence.

Human intelligence (HI) performance in melanoma recognition is far from being perfect and is heterogeneous. Many aspects influence in the decision on whether a lesion is a melanoma, and are influenced by psychological factors such as self-confidence, lawyers...

How does HI work? We can use criteria to identify melanomas (7 point checklist, ABDCD...), also dermoscopic algorithms. But this has limitations like oversimplification + distortion of the process. But this is not really how our brain works, since we identify global patterns that classify images in a category already stored in our brain. Hence, pattern recognition needs training to set the foundations in our brain. An advantage of pattern recognition is that it is very fast because it has an evolutionary meaning: we had to identify dangers and identify them fast.

The disadvantages of pattern analysis is that they are useless if you don’t get trained, and also you cannot train someone if that person does not care about it. So, if someone is not interested in dermoscopy, this is going to be a problem. In addition, in order to recognize a pattern, we need to simplify to work (memory saving process). The problem is then oversimplification. We then could learn more features but is this useful? What kind of representation do we recognize when facing melanoma vs nevus on dermoscopy?

We can be taught features, but we do not really agree on what we see, except for disorganization/chaos. In addition, we tend to learn more from images than actual teaching of a feature. This is crucial for patients, hence brochures with images with melanomas are better than telling people how to identify a feature. Also, for dermatologists, since showing many images or seeing many patients increases the diagnostic accuracy through pattern recognition.

Another thing about brain does well is identifying the outliers, the ugly duckling. This is useful in melanoma detection since people have 2 to 4 nevi patterns. Therefore, the pigmented lesion that does not fit in these patterns is an outlier and may be a melanoma: the ugly duckling sign. In addition, the ugly duckling sign improves the performance of dermoscopy.

Another cognitive process that work to identify melanoma with HI is the ability to identify changes: chronological recognition.

On the other hand, how does artificial intelligence (AI) work?

We (humans) can create an algorithm to identify dermoscopic features, ugly duckling... But this has limitations in identifying things we have not taught the machine.

Another way is machine learning, in which the computer learns alone how to distinguish a dog from a cat, for example. This needs large datasets, and we do not know how the machine reached a conclusion since it goes through different layers.
How current AI algorithms perform in melanoma detection? They are performing at the same level or better than dermatologists. But is this a fair assessment? Probably not because the current algorithms have been trained in artificial conditions with standardized image datasets (most of the times with dermoscopic images which are pretty controlled) and included many frequent lesions. Hence, to be able to use these algorithms we need larger datasets with real-life images including also rare lesions, confounders (ink on the skin, hair…).

So, what could be the role of AI in melanoma detection? Can AI replace the doctor?
AI could replace the doctor if AI had access to all information available at clinical examination, AI was trained with large datasets with large samples for rare situations, AI was performed in non- standardized conditions, AI could replace social interaction... Also, if AI is wrong who is responsible for it? How is the information delivered to the patient? A percentage of risk of a diagnosis? So, for now AI is not going to replace us but it will be a complement to the diagnostic process. Also, AI may help us identify patients with aggressive melanomas, and could help patients in melanoma self- detection (although it needs people’s motivation).

Melanoma and aging

How age impacts on my treatment decisions in high risk and metastatic melanoma
According to the oral communication of Reinhard Dummer (Zurich, Switzerland)

In elderly patients with melanoma there are some general considerations to take into account: patient characteristics (comorbidities, medications, mobility, quality of life [QoL], motivation) and characteristics of the tumor (location, melanoma subtype, driver mutations, tumor mutation burden [TMB]).

Regarding quality of life, patients' QoL treated with dabra+trame actually improve and is linked to clinical efficacy. So, if the patient has BRAF mutation, dabra+trame can be used as adjuvant treatment in stage III patients.

A type of melanoma more common on the elderly is desmoplastic melanomas, and studies have shown that they respond very well to antiPD1 treatment.

But what do we consider elderly? The ESMO considers elderly >75. In these cases, one should discuss the need for SLNB. However, it depends on the status of the patient and the characteristics discussed above.

Does age influence on the outcome? Age does not really have a negative impact on the outcome. Hence, age is not a limitation for CPI or TT.

What do we do in metastatic disease? Typically, immunotherapy is the first line, but we need to assess: if the patient is symptomatic, the patient's spirit, level of LDH, organs involved...

Which treatments can be used? Ipi + nivo is the most powerful treatment the added benefit vs nivo alone is not that high at the 5-year survival rate. In addition, the added benefit with ipi+nivo is more pronounced in younger patients. Therefore, because of the toxicities and this efficacy data, in elderly it tends to be better to use antiPD1 alone. Maybe only exceptions if brain metastases or high LDH, but it is important to individualize the decision. Also, if BRAF mutated elderly people tend to respond equally well to younger patients.

Age against the machine: the impact of aging on melanoma progression
According to the oral communication of Ashani Weeraratna (Baltimore, USA)

Cancer is a disease of the elderly due to multiple factors such as accumulation of genetic alterations. But it seems that microenvironment has an impact.

Experiments in mice using the same melanoma cells injected in young and old mice, they saw that melanoma grows faster in old mice. It seems that fibroblasts in the microenvironment are altered because of decrease in the genes related to dormancy and have an increase in the proliferation genes.

Lipids are also important in the microenvironment being upregulated: FATP2 is increased in aged fibroblasts and melanomas in old animals (Alicea et al. Cancer Discovery) → hence, targeting FATP2 (which appears in elderly models) could be make a melanoma responsive to BRAF/MEKi

Another aspect they investigated is the sexual dimorphism in melanoma (the incidence and survival are difference in men and women). Female fibroblasts are more senescent than men, more with age. It is the same for the shape of fibroblasts: cell shape asymmetry increases with age. Also, mitochondrial ROS are different and higher in men than female.

Melanoma intratumor heterogeneity, growth and metastasis
According to the oral communication of Jean-Christophe Marine (Leuven, Belgium)

Tumors are complex and evolving ecosystems and melanomas grow according to hierarchical model. From their experiment they have found that:

• Stem-like melanoma cells are located in the perivascular niche in primary melanoma lesions.
• Productive clones are in proximity with blood vessels.

Also, it is important to know if in this niche there are cells that will induce metastatic disease. What they say was that growth and metastatic dissemination are driven by transcriptionally and spatially distinct melanoma subpopulations.

Management of melanoma in pregnancy

Survey of management practices across
According to the oral communication of Europe Simone Ribero (Turin, Italy)

Melanoma accounts for 31% of malignancies present in pregnant women. Whether melanoma during pregnancy is worse is still debated. So far, it is unclear whether pregnancy worsens melanoma progression.

Also, what do physicians do in these cases was evaluated in a survey. In that survey they presented different case scenarios presenting different melanoma scenarios in women.

The risk of relapse of melanomas stage IA and IB are almost constant in the first 10 years, so if thin melanomas in pregnancy it may not necessarily be a problem.
With thick melanomas it is unknown but maybe it may be good to wait 2 years since the risk of relapse is higher in the first years. But the problem is that there are no consensus papers.

Another scenario would be a post-menopausal woman in hormonal replacement therapy: there's no evidence hormonal therapy increases the risk of melanoma.

Another scenario is whether SLNB was advised or not in pregnancy: the risk of the procedure is very low (anesthesia and the technetium are not contraindications during pregnancy). Also, we can use MRI and ultrasonography to do staging.

Oestrogen progesterone receptors in melanoma and nevi
According to the oral communication of Emi Dika (Bologna, Italy)

Sex hormones can be oncogenic, for example in breast cancer. Also, prognosis is better in men than women. So, there is the hypothesis that estrogens may have a role in melanomagenesis. This has been studied in the last 20 years with the onset of tamoxifen. Typically, two receptors were studied, estrogen receptor alfa and beta. However, there is a new pathway potentially implicated in melanoma: GPER → G protein-coupled receptor.

Melanoma cells lack significant classical ER but express GPER.

Binding of estrogen or a selective GPER agonist (G-1) signals through adenylate cyclase to activate PKA. PKA phosphorylation of CREB drives expression of MITF, thus increasing pigmentation and differentiation. PKA further deplets c-MYC leading to decreased cell proliferation.

Double receptor expression of estrogen receptor (ER) beta and GPER was detected in 73% of pregnancy-associated melanomas (PAM) and was associated with positive prognostic markers such as lower Breslow thickness, less mitoses and presence of peritumoral lymphocyte infiltrate.

ER alfa and progesterone receptor (PR) positivity in cases of melanoma occurring in women after in vitro fertilization. These melanomas were also very pigmented clinically and dermoscopically.

Other studies have shown conflicting results: Giorgi et al. shown that ER beta supports the proliferation of melanocytic cells, whereas ER alfa has an inhibitory and regulatory effect on the invasive capacity of melanoma.

In 2020 they did another study and in breast cancer patients, aromatase inhibitors may contribute to the nuclear localization of ERbeta, which is a known estrogen responsive receptor with a repressor activity, thus playing a protective effect → aromatase inhibitor may have a protective melanoma effect. Regarding women having ovarian stimulation, they did not find any nuclear estrogen receptors due to ovarian stimulation.

To sum up, the role of ER in melanoma is still debated and further studies are needed.

Does pregnancy affect melanoma prognosis?
According to the oral communication of Catarina Longo (Reggio, Italy)

Melanoma is the most common cancer during pregnancy (31%, incidence 5-10 cases per 100000 pregnancies)

Pregnancy is characterized by a physiologic immunosuppression to increase tolerance not to reject the fetus. However, there is no evidence this immunosuppression increases the incidence of melanoma. Another question is whether if the prognosis of melanoma during pregnancy is worse.
Multiple studies tried to analyze this issue, but the problem is that there are no clear definition of what is considered a pregnancy-associated melanoma (PAM) → some studies include melanomas in the post-partum, others not...

Kygidis et al. performed a meta-analysis assessing the PAMs, and showed PAM was associated with 17% higher mortality compared with melanomas diagnosed in non-pregnant women. Also, PAM was associated with a 50% higher recurrence rate. So, they concluded that PAM is associated with a worse prognosis.

Byrom et al. also performed another systematic review and confirmed an increased mortality associated with PAM. However, further studies are needed since the topic is still controversial.

Melanoma pathology – genomics and morphology

Spitz tumors, towards a unified definition
According to the oral communication of Iwei Yeh (San Francisco, USA)

Spitz tumors are the IV pathway in the new WHO classification for melanocyitic proliferations. Spitz nevi (SN) are typically more common in children and can present as pink papules or also as heavily pigmented macules (Reed nevus).

Histologically: epidermal hyperplasia, Kamino bodies, large melanocytes with enlarged nuclei. Mutations are diverse in Spitz nevi, and many are activating kinase fusions. Mutations may be HRAS mutations, BRAF fusions, ROS fusions, ALK fusions, NTRK1 fusions among others (MET...).
Also, these fusions can be arranged in multiple ways, and the expression of the fusion protein depends on the gene initiator. Interestingly, different mutations and fusions have different histological presentations:

• HRAS mutated SN: Histologically, enlarged melanocytes, prominent nucleoli. Sometimes they have gain of 11p
• NTRK3 fusions: may have spindled cell morphology
• NTRK1 fusions: may have rossettes histologically.

It is important to use ancillary testing to identify the initiating mutation:

• Immunohistochemistry: potentially very good tool and easy --> can stain for ALK, NTRK and ROS.
• FISH
• CGH

Spitzoid tumors vs Spitz tumors: Spitz tumors are the classic atypical Spitz tumors or now called Spitz melanocytomas.

What happens with spitzoid melanomas? They analyzed the spitzoid melanomas and saw the purely Spitz melanomas which have HRAS mutations and fusions, whereas others did not have these patterns but had BRAF point mutations, mutations in the MAP kinase pathway (maybe these melanomas are UV-driven), and then MAPK wild type melanomas which also lacked HRAS and fusions.

TERT promoter mutations are predictive of aggressive clinical behavior in patients with spitzoid neoplasms.

Spatial pathology to predict the response of melanoma to immunotherapies
According to the oral communication of James Wilmott (Sydney, Australia)

Not all patients respond to immunotherapy, and also these therapies have toxicity. It is crucial to determine who will respond and who will not improve response and avoid toxicity.

There are multiple data used to triage the patients and probably the future will be an integration of clinical, serological and pathologic data.ю

Their group performed whole genome sequencing, RNAseq and immunohistochemistry for PD- L1/TILS to assess the response. They have seen that patients who respond have higher TILS and higer PD-L1 expression (although other studies have been controversial). To perform these studies, they used multiplex immunofluorescne and quantitative pathology. Later they assessed this information using spatial pathology which calculates distances and interactions between cells. Then, this information was used as input for response models to see if the tumors may respond. Hence, they could prove that spatial pathology can translated into clinical practice to predict treatment response.

Later, this information was integrated with clinical information such as LDH, prior treatments and so on, used machine learning algorithm to provide models of treatment prediction.

So spatial pathology may be a promising tool to integrate pathology into determining treatment response in melama.

How do you recognise BAP1 inactivated melanocytic tumors and what is their significance?
According to the oral communication of Arnaud de la Fourchadière (Lyon, France)

BAP1 is normally expressed in cells. When lost, this can induce the onset of BAP1-inactivated melanocytic tumors (BIMT) --> single sporadic lesions, or multiple lesions associated with BAP1 germline mutations and associated with a cancer syndrome.
When there's a BAP1 germline mutation this can be associated with multiple BIMTs but also uveal melanoma, cutaneous melanoma, leptomeningeal melanoma, mesothelioma, clear-cell renal cancer, meningiomas, multiple basal cell carcinomas, intrahepatic cholangiocarcinomas, among others probably yet to be described.
Clinically BIMT present as polypoid papulonodules, with raised areas that loose pigment. Histologically, they present an intradermal expansive clonal population of atypical spitzoid melanocytes that are large, unpigmented, and epithelioid, surrounded by a more banal nevus. Also, one can see "kissing figures": melanocytes are close to lymphocytes suggesting an interaction between melanocytes and lymphocytes.

BAP immunohistochemistry show BAP1 is lost in the nuclei of the atypical cells. Additionally, they are positive for BRAF V600E.

If one single BIMT is found, the risk for a BAP1 germline mutation is between 12 to 16%. Hence, there's no guidelines on what to do: for now, only removal with 2mm margins, maybe discuss if needed to have genetic testing and having an ophthalmologic check-up.

It is also possible that some melanomas arise within a BAP1 inactivated nevi, which although they are rare, they can be very aggressive.

UV-radiation and melanocytic neoplasia
According to the oral communication of Amaya Viros (Manchester, United Kingdom)

UV exposure drives melanoma incidence: UV induces inflammation and drives DNA damage. However, UV exposure influences melanoma survival? Some studies show a protective effect of UV light in some melanomas, whereas some other studies showed the opposite. What Viros' lab did was analyze the effect of UV on fibroblasts and collage. They have found that signature 7 is linked to UV damage. They say that UV degrades collagen, and they could also see the same in collagen around melanomas. In addition, they could see that the collagen around the melanoma cells have different invasion capacity. Hence, collagen integrity modulates melanoma invasion. Later they evaluated these results in a cohort of melanoma samples to study melanoma survival.

They saw that chronic sun damage (CSD) can have different effects on melanoma depending on the microenvironment related to the collagen: if the collagen is degraded the invasion capacity decreases. However, if the collagen is reversed then these same melanomas have more invasion and do worse. Hence, collagen can be a predictive marker in melanoma response.

To sum up, in the presence of severe CSD this can induce collagen degradation, less invasion and better outcomes. But if the collagen is restored then the invasion increases and there is a worse prognosis.

Similarities and differences between blue nevi and related tumors and uveal melanomas – are they twins or distant relatives?
According to the oral communication of Klaus Griewank (Mainz, Germany)

Blue nevi and uveal melanoma harbor GNAQ and GNA11 mutations. So, are they related? Later, more mutations such as BAP1, SF3b1 EIF1AX were also found.

BAP1 is important in uveal melanoma; uveal melanoma has two subsets of patients, one that does very well and one that does very bad. The bad ones typically have loss of BAP1 which is linked to losses in chromosome 3p. Tebentafusp is a new drug that may work well for uveal melanoma. Ideally, we would like to target BAP1 and restore its function.

A different tumor is malignant blue nevus (also known as blue-nevus like melanoma): probably often not designated as such histologically. Lymphatic metastases seem to be common and should be monitored. In terms of treatment, BRAFi are not useful since they do not have BRAF mutations.

So, are these "blue" lesions related? The answer is probably yes because they share genetic abnormalities.

Immunotherapies beyond checkpoint inhibition

IL–2 like drugs including NKTR-214 and ALKS 4230
According to the oral communication of Adi Diab (Houston, USA)

Nemvaleukin:

• there is a clear activity as antitumor response but very limited if used as monotherapy
• the adverse events are mainly flu-like symptoms, and does not get worse when associated with antiPD1
• can be administered subcutaneously or intravenously

THOR 707:

• pegilated IL2 R-alpha
• the adverse events are mainly flu-like symptoms, and does not get worse when associated with antiPD1

NKTR-214 (nektar) Bempegaldesleukin:

 

• another modified IL2
• there is a clear activity as antitumor response but very limited if used as monotherapy
• the adverse events are mainly flu-like symptoms, and doesn't get worse when associated with antiPD1
• together with nivo: CR34%, OR 53%

Summary: similar efficacy and safety profile. They work best with antiPD1, most have flu-like symptoms that are grade I and transient. Need for more combination trials beyond PD1, such as combo with antiLAG3 among others.

Adoptive cell therapy with TIL and other effector cells
According to the oral communication of Inge-Marie Svane (Copenhagen, Denmark)

Adoptive T cell transfer (ACT) approaches involve tumor TILS (tumor infiltrating lymphocytes) or genetically modified T cells (with TCR or CAR).
Although immunotherapy works, still 50% of patients do not respond, so additional treatments are needed.

Lifileucel is a group of TILS that achieves 36.4% response in anti-PD1 resistant patients. There is an ongoing trial comparing these TILS vs ipi in PD1 resistant patients.

Which melanoma patients are candidates to TILs as potential standard of cancer?
Age < 70
Good ECOG 0 or 1
No brain metastases
sufficient organ function
No active infections

To sum up, TIL therapy is still an experimental therapy but has potential since if patients respond they may have complete responses even in metastatic patients. However, we still do not know on how to predict response, define the TIL potency and develop better TIL therapies.

Evil cytokines in cancer immunology and immunotherapy
According to the oral communication of Ignacio Melero (Pamplona, Spain)

Some cytokines are evil for cancer: TGFbeta, VEGF, GDF-15, LIF. Others maybe, but unknown. TGFbeta has a complex immunobiology, with different subunits: 1, 2 and 3.
Fresolimumab: blocks TGFbeta1, 2 and 3 Bintrafusp-alfa: anti-PD-L1 moiety + TGFbeta trap moiety → failed in phase 2 trial against pembro Other cytokines that are interesting in melanoma treatment toxicities are TNFalpha and IL8:
What happens with TNFalpha: in cases of CPI toxicities there's an increase of TNFalfa → for example in colitis induced secondary to ipi+nivo. Hence, these side effects can be improved with antiTNFalpha What happens with IL-8: it is neoangiogenic and IL8 is elicited via TNFalpha. Some studies found that higher IL8, worse prognosis → IL8 induces a release of neutrophils which is known to worsen melanoma prognosis. Thus, IL8 is potentially a target. In fact, they have just started a trial to block IL8.

Epigenetic modifiers
According to the oral communication of Ryan J. Sullivan (Boston, USA)

Although CPI have revolutionized melanoma, still many patients do not respond. Why? Many explanations:

• alternative immune checkpoints: TIM3, LAG3 insufficient priming
• insufficient TMB/neoantigens
• presence of regulatory T cells and/or tumor associated macrophages inadequate T cell recruitment/infiltration
• loss of tumor antigen presentation machinery/loss of IFN signaling

Also, epigenetics can have an important role for oncogenic pathways and immune cancer regulations. Therefore, epigenetics regulators are potential therapeutic targets in cancer.

G9a suppresses DKK1 suppresses WNT pathway → this is conserved across cancers associated with "cold" immune microenvironment.

G9a is oncogenic through copy gains or activating mutations. Functions through canonical Wnt activation. Promotes cold environment.

G9a suppression synergizes with immune checkpoint blockade, and these inhibitors are in active development in the pharma industry.

Thus, targeting these epigenetic regulators may reverse immunosuppressive pathways (ie Wnt).

Another example of importance of epigenetics is Beta2-microglobulin (B2M):

 

• B2m aberrations are associated with acquired resistance to CPI.
• B2M loss of heterozygosity (LOH) in pretreatment samples is associated with poorer responses rates, PFS and OS with ipi and antiPD1.
• Potential strategies to target the resistance mechanisms include agents that increase/enhance NK cells numbers/activation and agents that overcome epigenetic silencing of B2M antigen presentation machinery.

Another epigenetic modification that may enhance immune checkpoint inhibition:

• HDAC inhibition → entinostat (ENT) shows synergy with anti-PD1 in clinical models. In clinical practice, ENT + pembro showed more durable responses in some patients.

New approaches with neoantingen vaccines/ T-cell therapies
According to the oral communication of Samra Turajlic (London, United Kingdom)

Neoantigens are key targets of tumor-specific immune responses.The goal of neoantigen-directed vaccines is to prolong response of immunotherapy. However, there are some challenges for neoantigen-targeting immunotherapies:
1. Neoantigen (NA) identificaiton
2. Adequate tissue procurement (TILs)
3. Delivery platforms (vaccines)
4. Adjunct therapies to protect tumor-specific T-cells immunosuppressive mechanisms
5. Intratumor heterogeneity