Welcome to the ACMG Genomics Case Conference, everyone. This series has been planned by the ACMG Education Committee. Each month, the conference will rotate among laboratories that perform genome-scale testing and the clinicians involved in these cases. Use of all the case materials that will be presented here have been formally consented for use with the patients of the guardians. This activity is supported by an unrestricted educational grant from QIAGEN Bioinformatics and the Ingenuity Clinical Decision Support Team. This activity is intended for educational purposes only. The presenters on this conference are Alicia Gomes, who is a Senior Board Certified Lab Genetic Counselor, and myself, Lejeune Messiaen. I'm the Director of the Medical Genomics Laboratory and a professor in genetics. We do not have any disclosures, and these are the Planning Committee members' disclosures. The session is approved for one continuing medical education credit, and for those who paid for CME, they will receive an email with evaluation questions and a certificate of completion at the end of the session. These are the overall learning objectives for the Genomics Case Conferences, but our specific learning objectives can be seen here. Indicate NF1 versus Lasius syndrome versus mosaic NF1 by applying appropriate testing and recognizing the current genotype-phenotype correlations. Review unique testing approaches to analyze mosaicism. Illustrate the ambiguity and overlap between NF2, mosaic NF2, and schwannomatosis. Utilize germline and somatic testing approaches to interpret and examine pathogenicity of variants associated with schwannomatosis predisposition disorders. If you want to ask a question, you can type your question into the Q&A box in the lower left corner of your control panel. Questions will be answered at the end of the presentation. So now we will get started. We will discuss three groups of disorders, NF1, NF2, and schwannomatosis with complementing clinical scenarios. NF1 was first described in 1881 by Friedrich von Recklinghausen. It is a fully-penetrant disorder caused by pathogenic variants in the NF1 gene that is located on chromosome 17q. The gene contains 60 consecutively transcribed exons, and there also exist multiple unprocessed tritogenes, which initially hampered the development of DNA-based approaches. To avoid these nonprocessed tritogenes, we developed an RNA-based testing approach and reported a 95% pathogenic variant detection rate in classic NF1 cases back in 2000. The spectrum of pathogenic variants in NF1 is very diverse and includes missense, frameshift, nonsense, total gene deletions, copy number changes, and most commonly, the splice variants. So NF1 is a tumor suppressor gene that is acting in the RAS pathway, where it is converting the active RAS GDP into the GDP form. NF1 is an autosomal dominant disorder that is affecting one in 3,000 individuals worldwide. And an individual will meet the clinical criteria for NF1 when he or she fulfills two of the following criteria. The presence of more than six cafiolespots in a pre-peripertal individual of more than 5 millimeters, in a post-peripertal individual more than 50 millimeters in diameter. Presence of skin fold frackling, or of two or more Leishner genes, which are benign iris hematomas, optic pathway gliomas, two or more cutaneous or subcutaneous neurofibromas, or one or more plexiform neurofibromas. Presence of one of these typic osseous lesions, and a first degree relative with NF1. Now these clinical features have different ages of onset, and hence they do not all appear at the same age in an individual. The cafiolespots are usually the very first sign heralding a potential diagnosis of NF1, and there are usually two or more at the age of six in more than 90% of children with NF1. Plexiform neurofibromas occur in about 20 to 25% of N individuals with NF1, and they are frequently congenital. Although these are benign tumors, they need careful monitoring, as they can become malignant and become malignant peripheral nerve sheet tumors that is associated with high morbidity and mortality. Skin fold fracking is appearing around the age of one to two years of age, and then we have the cutaneous and subcutaneous neurofibromas that develop later, usually around puberty, and these as well, they will occur in about 90% of all of the individuals with NF1, with their size and their numbers increasing with age. They are and they remain benign, but they can be a major cosmetic and psychological social burden for the patients. So as we said, about 50% of the patients with NF1 are sporadic individuals, sporadic cases. Now this means that either they have constitutional NF as a result of a mutation in the germ cell of a parent, and so consequently the mutation is present in each and every cell developing from that fertilized egg, or the patient may have actually mosaic NF1 with a first hit originating post-psychotically during development. Now NF1 mosaicism can present in different shapes, so it can present as pure gonadal mosaicism only. Now for NF1, this must be a very rare event because only two cases so far have been reported in the literature, and the third case was reported recently at the first global meeting for the neurofibromatosis in Paris. Then we have mosaic generalized NF1, which means that the features seen in the patient are not limited to a specific quadrant, and or they do not stop at the midline. It is often very difficult to distinguish mosaic generalized NF1 solely on clinical grounds from patients who have an inherited constitutional mutation. Finally, we have mosaic localized or segmental NF, where the patients are presenting with features that are limited to one or a few areas of the body. The features there can range from pigmentary changes only or neurofibromas only or a combination of features but localized to a certain area of the body. Now early case studies by our group and others demonstrated that in patients who had, for instance, a micro deletion detectable in blood only in a small fraction of the alleles, around 10 percent of their alleles, that in such mosaic cases, when we were looking at fibroblasts from the café au lait spots, there was absolutely no enrichment of that deletion in the fibroblasts to be seen, indicating that this cell type was not enriched for somatic mutations. Therefore, we set out to identify the cells of origin for the most prevalent and easily accessible NF1 features. These are the cutaneous neurofibromas and the café au lait spots, with the goal to obtain a better understanding of the genetics of these lesions, as well as with the goal to translate this into a useful diagnostic approach. These studies resulted in the discoveries that Schwann cells are the neurocrest-derived cells that are carrying two hits in the neurofibromas, and that it are the melanocytes but not keratinocytes or fibroblasts or epithelial cells that are carrying the two hits in the café au lait macules. To appreciate further the necessity of specific testing approaches, we need to mention the recent developments in our understanding of genotype-phenotype correlations. Given a number of factors, as described here, that are listed, the identification of specific genotype-phenotype correlations is very challenging, and it will never be the norm for the majority of the patients. Nevertheless, there are some recurrent pathogenic variants that are associated with genotype-phenotype correlations, and that we need to bring to attention. First of all, we have the microdeletions that are encompassing the NF1 gene and its flanking regions. They are the most recurrent pathogenic variants, affecting up to 8% of the NF1 population. Broadly, we recognize three types of microdeletions. We have the type 1 deletions, with a loss of 14 genes, that is due to meiotic non-analytic homologous recombination between the NF1 rep A and rep C. And they are the vast majority of the microdeletions found in 70% to 80% of patients with the microdeletions. Then we have the type 2 microdeletions, with a loss of only 13 genes, where now the non-analytic homologous recombination is intrachromosomal and is mitotic in origin, and it is due to a non-analytic homologous recombination between CIS-12 and its pseudogene CIS-12B. Now, the vast majority of the type 2 microdeletions, they are starting in an individual as a mosaic deletion, and thereby, these patients very frequently exhibit only a mild phenotype. Importantly, of course, upon inheritance, the presentation in the next generation is typically more severe. Type 3 deletions have a loss of a smaller number of genes, only three, also due to non-analytic homologous recombination, now between the LRRC37 containing low copy repeats. They are only a very small fraction of the microdeletions. Now, the patients with the constitutional microdeletions frequently have a more severe phenotype, as mentioned here. These patients, importantly, are also at an increased risk for malignant peripheral nerve sheet tumors, which are very aggressive tumors that are difficult to treat. The constitutional co-deletion of the CIS-12 gene, which is in itself a tumor suppressor, along with the NF1 gene, is thought to be a risk factor that is associated with the increased risk for these highly malignant neoplasms in these patients. So additional emerging clinically relevant genotype-phenotype correlations include a cluster of missense pathogenic variants at the codons 844 to 848 that are affecting almost 1% of all unrelated NF1 individuals in the UEB cohort. This study was recently published, beginning of this year, in the American Journal of Human Genetics. Besides the severe phenotypes, there are also a number of recurrent variants, pathogenic variants, that have been identified and recognized for the moment to be associated with a distinct more mild phenotype of cafeolespots with or without freckling, but no externally visible plexiform, cutaneous, subdermal neurofibromas, no symptomatic optic pathway gliomas, or other tumors. So this pertains to the deletion of 1-methyl-992 at position 992, so we published an update on this particular variant recently in Genetics in Medicine, and then we have all of the variants, the missense variants affecting arginine 1809. That brings us to SPREAD-1, which is responsible for about 2% of the individuals having a clinical diagnosis of NF1 have not a mutation in the NF1 gene, but actually do carry a mutation in the SPREAD-1 gene, so SPREAD-1 Leaky Syndrome is actually an important differential diagnosis in an individual with multiple cafeolespots with or without freckling, but no nervous system tumors as those that are seen in NF1, but learning disabilities and NUNA-like features can be present. Now SPREAD-1 is acting similarly to NF1 in the melanocytes, and besides the first hit, a second somatic hit is found in every different cafeolespot in patients with Leaky Syndrome. Now taking into account what has been learned over the last 20 years since the cloning of the NF1 gene, which was actually the first tracepathy gene, there is definitely a role for panel-based testing, especially for those individuals with an ambiguous phenotype. There are several genes in the RAS-MAP kinase pathway that have been identified over the years and that have been associated with several syndromes, including Costello-Noonan syndrome, with multiple antigenes, and so on, and they have an overlapping phenotype, although there are also unique phenotypic characteristics that are associated with each of these syndromes. Now some of these features overlap with NF1 and SPREAD-1, and given the phenotypic variability typically observed in each of these syndromes, we have integrated what has been learned from the mutational spectrum in NF1, also with next-generation sequencing-based assay, including other genes within the RAS pathway. And now we will discuss some of the cases on how we approach this testing. Sure. So we'll start with one case that presents with a six-year-old boy who comes to clinic with seven typical cafeolespots with smooth borders, left inguinal freckling, some bilateral freckling. His mother notes that he is having difficulty in the first grade, and he's currently under evaluation for a learning disability. The parents were present at the appointment, so we were able to do an evaluation of the dad, who was also identified to have four typical cows, a couple atypical cows as well, and some axillary freckling. There was a lump that was noted, but upon palpation, it felt more like a lipoma than a neurofibroma. And then there was notes from the family that there was a paternal grandmother that may have had some birthmarks that were present when she was younger and maybe a learning disability as well. So, of course, the first concern always arises of whether this may be NF1 versus Legius syndrome. Now, when we take a look at our cohort of patients that had been tested in a study published in 2009, we learned that the vast majority of these individuals who have a family history of cafe au lait spots and freckling that meet clinical criteria and have a family history, that the majority of them will actually still have an NF1 mutation, roughly a little over 70%. About 20% of these individuals have a spread 1 mutation. So while they are not the majority, it should still be taken into account when creating your differential diagnoses. In this particular family, this individual was identified to have an arginine 1809 mutation. And as Dr. Messin has already mentioned, this is a mutation that does have a genotype- phenotype correlation where we know that there are multiple cafe au lait spots with or without freckling that are present. You can see Lish nodules, but most importantly, there are no externally visible plexiform neurofibromas or clear cutaneous neurofibromas that have been seen with this mutation. Also, you can see a higher prevalence of Noonan-like features, and there is also a higher prevalence of pulmonic stenosis and short stature. This is important as a clinician because cardiac evaluation may not have been a part of your initial workup for an individual with NF1. So having this information up front may change your management of this patient. So we'll move to another very common clinical scenario, which is an individual who is a sporadic case with six typical cafe au lait spots and axillary freckling. In this particular family, we're unable to do a clinical evaluation in mom and dad. They also note that there's an older sister and a younger sister who also are asymptomatic, but they did not come to clinic as well. So in sporadic cases, your differential diagnoses do broaden quite a bit. In addition to any type of constitutional NF1 variant, you're also concerned about mosaicism for NF1. And in some cases, this mosaicism may be detectable in blood. However, in other cases, it may require a biopsy-based test. And then, of course, we have to think about overlapping conditions such as SPREAD-1. Now, there are important counseling implications for any of these differential diagnoses, and it can evolve the way you manage the patient as well based on these different diagnoses. So to give you a little bit of background about our testing approaches for these types of patients, we'll start with the original testing approach that was initiated in this lab in 2003, which started with an RNA-based testing approach that uses a short-term lymphocyte culture that then concludes with a long-range RT-PCR for cDNA for sequencing. Now, this approach enabled us to identify quite a number of variants, but the one I'd like to highlight especially are the splicing variants. So within the years of 2003 to 2016, we were able to find more than 3,000 different variants in more than 8,000 unrelated positive individuals. But what's important to note is that the majority of these cases have their own unique mutations. Only 31 variants have a frequency of higher than 0.4 percent. One of the biggest challenges that we noticed is the ability to disseminate between missense and splicing variants. But with this RNA-based approach, we were able to actually reclassify a number of these missense mutations into splicing defects, which would have otherwise gone out as a variant of uncertain significance without this approach. We were also able to find that there were more than 65 different deep intronic splicing variants that were in these regions that were more than 50 intronic base pairs flanking each of the exons. So in 2016, this is kind of where we stood with our cohort of patients, and what we realized as a limitation to this testing approach is that it is still Sanger sequencing-based. So our biggest limitation was the ability to identify mosaicism without involving a more invasive process like a biopsy. And so in 2016, we moved to a Haloplex next-generation sequencing approach that is specialized to include all of those 65 intronic regions that we know have those deep intronic mutations present. This testing approach still was able to identify all of the insertions, deletions, and duplications that ranged from 1 to 112 base pairs, and it was still complemented by copy number analysis by MLPA for larger deletions and duplications. But the biggest thing I'd like to highlight is that with this next-generation sequencing approach with deep coverage of an average of 1,800x, we were able to identify low-level mosaicism as low as 2 to 3 percent of the alleles. And using that, especially since it was a next-generation sequencing approach, we were also able to offer more panel-based testing options. So our testing categories ranged from just NF1 only to a full rathopathy panel with low-level coverage for looking for mosaicism. Of course, we can never devalue the needs for an RNA-based test. And in a lot of cases, especially from affected tissue, this is still going to be the gold standard for identifying mosaicism. So our RNA-based testing approach is still available for melanocyte cells and Schwann cells for mosaicism testing, as well as for those rare scenarios where next-generation sequencing may not have picked up a mutation, but there is concern for a rare complex mutation that would be outside of those mutations that would have been identified using our next-gen approach. In this particular family that we started with, the final diagnosis was a spread 1 mutation. And so with the next-generation sequencing approach, looking at both genes simultaneously, this diagnosis was able to be made. Of course, this is helpful to the family because they can now assess his risk of passing down a milder phenotype to another generation as well. In addition, especially in his younger sibling, there may be a need to evaluate these individuals for minimal signs that may not have been appreciated before. We'll move to another scenario, but it's still very similar to the scenario before, because when we're talking about a sporadic case, we always need to be concerned about mosaicism. And mosaicism can be present in quite a spectrum of phenotypes. It could be with cutaneous pigmentary signs only. It could be with tumor burden, maybe not enough café au lait spots. The possibilities are quite broad when it comes to this. And so with our next-generation sequencing approach, with the ability to identify low-level mutations, the Haloplex uses multiple amplicons, so it helps to eliminate our concerns that a low-level mutation may be an artifact when we see it present in multiple amplicons. In this particular patient, that was especially helpful because we were able to identify a mutation present at 4% in NF1. This is not the only time that this has been found. We've seen a number of other variants as well that have been seen at 4%. And especially for patients who meet clinical criteria, the physician may not have been considering mosaicism when they were doing their differential diagnoses as well. So having this as a part of your straightforward testing approach helps to confirm mosaicism and assess the patient's reproductive risk without the need for an invasive test. However, we have to keep in mind that there are going to be cases where it doesn't matter how deep the coverage, a next-generation sequencing approach will not identify mutations. And so this goes back to Dr. Messiaen's array of questions. And so this goes back to Dr. Messiaen's original conversation regarding our biopsy-based testing approaches. So for the CAFLA spots, it will be specifically those melanocyte cells and then the neurofibromas, specifically the Schwann cells, that will be necessary to identify a first hit associated with disease. So in this particular patient, three CAFLA spots were sent. And as you can see in the table, there was one common first hit that was identified and a unique second hit in each of the cows. This specific deep intronic mutation was one that was identified using our RNA-based approach. So you can see in the chromatograms to the left that from a gDNA-based approach, the mutation simply looks like a simple substitution. However, when you look at the cDNA, it's more obvious that there's been exonization of 80 base pairs from the intron into the spliced exonic material that's read by the mRNA. And so this is helpful because this may have been filtered out or not even considered with the next-generation sequencing approach or any simple gDNA-based approach that does not have the ability to look for deep intronic mutations. So while we're talking about splicing variants, it's important to note that splicing variants make up about 30% of all of the NF1 variants that you can see. 20% of these are not necessarily a problem. These are the ones that are the first two base pairs before and after each exon that are more easy to identify. However, there are some problematic splicing defects that can be missed by any basic panel-based testing or next-generation sequencing approach. 10% of these will be those deep intronic variants. So if the test has not been specific to look for these, those could be missed. 37% of these will be nonsense or missense mutations, so they may be misclassified as variants of uncertain significance or confirmed to be pathogenic but not necessarily as the right characterization. However, some of these are silent variants, and so depending on the filtering pipeline, they may go undetected completely. And then 35 percent of these are going to be variants that affect the intronic regions that are further flanking the exons past the first two base pairs. And so these may also be undetected or considered variants of uncertain significance. This is also topped with the fact that many splicing variants have an atypical presentation. So now we will continue with NF2 and schwannomatosis. So we have here back our first slide. So NF2 was originally described back in 1822 by Wieschaert, and the NF2 gene was cloned by two independent groups back in 1993. And the hallmark of NF2 are the bilateral vestibular schwannomas, but also meningiomas and ependymomas are seen in about half of the patients. Patients on average will present with signs in the early 20s with hearing loss, tinnitus, dermal schwannomas, and this disorder is fully penetrant. Schwannomatosis is the latest syndrome that was discovered, and we do not have the large cohorts of thousands of well-characterized patients, and that is also why there is still so much unknown and so much that needs to be learned. The hallmark of schwannomatosis is the presence of multiple non-dermal schwannomas, and the patients on average present in their 30s or 40s with pain and multiple lumps, bumps, schwannomas. Already back in the 90s, the Boston Group clearly established that the gene for schwannomatosis must be located on chromosome 22q through linkage-based analysis on multiple large families with schwannomatosis. And they were excluding the NF2 locus to be the responsible locus for the inherited schwannomatosis, but they found in all the different schwannomas of these individuals from these families a different NF2 pathogenic variant in each and every one of the schwannomas. As I said, the NF2 gene was cloned in 1993. This is a much smaller gene, and it is more simple with regard to its architecture, and it is easier to set up the genetic testing. Again, copy number variant analysis on DNA can identify a pathogenic variant in more than 90 percent of the familial cases. This percentage is slightly higher when applying RNA-based testing. Now, importantly, about 50 percent of all the NF2 cases are sporadic cases, but the prevalence of mosaicism in sporadic NF2 is really high. It's as high as around 30 percent. Therefore, a pathogenic variant is identified in blood by sour sequencing and CNV analysis only in about 70 percent of the cases. This percentage may be slightly higher when NGS is applied with deep coverage, but sometimes, no matter how deep you're sequencing, no variant will be identified in the blood, and there's a need to analyze the tumors in such cases. Phenotype-phenotype correlations in NF2 have been reported, and so a good reference would be Holliday et al. from the Journal of Medical Genetics of 2017, with the most severe presentations associated with truncating variants in the exons 2 to 13. Milder presentation is associated with missense variants, in-frame deletions or duplications, and deletions involving the promoter region with or without exon 1, and then truncating variants in exon 1. Then schwannomatosis. It is very hard to distinguish schwannomatosis from mosaic NF2, and that they are actually two genetically different disorders took a very long time to establish as we refer to the work from the Boston group. In schwannomatosis, the majority of the cases are sporadic, and non-penetrance is very common. The first gene predisposing to schwannomatosis was identified in the Netherlands by Hans Vos et al. in 2007. This was the SMARCB1 gene. Second gene, LZTR1, was identified seven years later in 2014 by our lab. Besides presence of schwannomas, the presence of pain is very typical in schwannomatosis. Now in familiar schwannomatosis families, a SMARCB1 variant is found in about 48% of the cases, LZTR1 variant in about 38%. There's still a fraction of about 14% with an unknown cause of their familiar schwannomatosis. Other genes must exist and need still to be discovered most likely. In sporadic schwannomatosis cases, a SMARCB1 pathogenic variant is identified in about 10% of the cases, but the LZTR1 in about 30%. It's still unknown in about 60%. As I just said, therefore we think that additional loci are probably still out there and need to be identified. Given the phenotypic overlap now that there is between NF1, NF2, mosaic NF2, and to a lesser extent also to mini-pneumatosis, molecular analysis of all of the known predisposing genes is the preferred diagnostic method. Diagnostic criteria are currently undergoing review by a panel of experts, and they will result in updated diagnostic criteria in the next year, taking into account all of the advances in the field over the last years. After the identification of the SMARCB1 gene, the group of Flora Pappi from Florence, Italy proposed a four-hit and three-events hypothesis to describe the molecular events occurring within the schwannomas. This has been shown to apply not only to SMARCB1, but also to LZTR1-associated schwannomatosis. Basically, as a first hit, either a SMARCB1 or an LZTR1 variant is identified in the blood. So we are presenting this here as gray or red. So red is for the SMARCB1 and gray for the first hit in LZTR1. In the tumor, the SMARCB1 or the LZTR1 variant is retained always, and there is loss of heterozygosity of the other wild-type allele, including the commonly both the LZTR1, SMARCB1, and the NF2 loci. So we call this event two. Furthermore, there is an NF2 pathogenic variant that is now different in every schwannoma tumor analyzed. So we are thinking back about the studies so long ago from the group from Mia McClellan, so a different NF2 variant in every schwannoma. And these NF2 variants are residing in cis of the first hit, and we call this event three. Now, ultimately, at the end, that means there is inactivation of both copies of either the LZTR1 or SMARCB1 gene, as well as of both of the NF2 genes. Now, molecular diagnosis is confirmed if a germline SMARCB1 or LZTR1 pathogenic variant is found in a patient with multiple pathology-proven schwannomas or other schwannomatosis-associated tumors, and there is LOH and a different somatic NF2 mutation in each schwannoma. This scenario is, in our view, the most robust assessment allowing to ensure the exclusion of mosaic NF2. So mosaic NF2 is excluded given the fact that you have two different NF2 mutation in every tumor. So as we mentioned, NF2 is more prevalent compared to schwannomatosis, and penetrance for NF2 is complete. Schwannomatosis is much more rare, and incomplete penetrance exists both for SMARCB1 as well as for LZTR1. Although the exact frequency of the non-penetrance in schwannomatosis is still not established, it is higher in LZTR1-related schwannomatosis than in the SMARCB1-related schwannomatosis. In addition, from the start, we knew that an allelic disorder was associated with pathogenic variants in SMARCB1 because we knew that lots of function variants of SMARCB1 are associated with rhabdoid tumor predisposition syndrome, whereas the variants in schwannomatosis typically observed are more likely hypomorphic. That means they are mesence or splice-at variants or truncating variants that are specifically located at the first or last exome of the SMARCB1 gene. In addition to the rhabdoid tumor predisposition syndrome, also cough insurance has been associated with mesence variants specifically at the three-prime end of the SMARCB1 gene. Now we also know that some of these variants have been found to predispose to both cough insurance as well as schwannomatosis in one and the same individual. And some of these variants have been found to predispose to rhabdoid tumor predisposition syndrome or schwannomatosis within the same family. Now for LTTR1, besides it is predisposing to schwannomatosis, the gene also has been associated now with autosomal dominant Noonan syndrome, and more recently has also been with autosomal recessive Noonan syndrome. So much has still to be learned about the spectrum of syndrome-associated variants, and therefore this field is in full flux. So maybe also for the LTTR1, we still need to mention that also the control databases do contain a number of variants in the LTTR1 gene. So whereas some of these variants might still be artifactual, other variants definitely will be real and may reflect the issue of non-penetrance that we are dealing with in this gene. So during panel testing, whole exome sequencing or whole genome sequencing, a SMART-B1 or an LTTR1 pathogenic or latically pathogenic variant is identified in an asymptomatic individual. We think that it does not suffice to make a diagnosis of schwannomatosis, so we strongly recommend referral to clinical experts for further clinical as well as molecular evaluation of these individuals. Now we will continue with a number of cases from the real world. So this is one particular scenario where two individuals have had the same phenotype, a left vestibular schwannoma, spinal schwannoma, and an ependymoma. With the ependymoma, of course, the primary concern is going to be NF2, but if you notice, they are of different ages as well. So one is 22 and the other is 55. So a lot of people start with, you know, where do I begin with testing? Should I start with tumor? Should I start with blood? And there are definite pros and cons for both scenarios. While tumor, we feel, is the best option for patients that you're concerned about mosaicism for, or if there's several differential diagnoses, if, for instance, the patient overlaps with schwannomatosis, in a lot of cases the tumor's not always available. Either the patient hasn't had surgery yet or the tumor was removed prior and is at a different facility. On top of that, sometimes when you have the tumor, tumor block, tumor material can have poor DNA quality and can be problematic during testing process. On top of that, of course, there is a higher cost associated with the tumor-based approach. However, with blood, it is more readily available and if the lymphocyte cells are involved, next-generation sequencing would be able to identify the mutation. Of course, we always have to keep in mind it is less costly, but if this testing is negative, your job is not done. There can still be a possibility of mosaicism and tumor-based testing may be needed. Alternatively, a mutation could be seen that will require additional studies to be done to confirm if it is indeed pathogenic through tumor-based testing. In this particular scenario, these two individuals did have NF2 mutations, but as you can see, one had a more constitutional, classic heterozygous NF2 mutation, while the other had a low-level NF2 mutation that was seen via next-generation sequencing affecting 2% of the alleles that required confirmation in tumor. In another scenario, and this is quite a number of scenarios actually, you can have individuals who have minimal signs of schwannomas, even unilateral vestibular schwannomas that have a general overlap between NF2, mosaic NF2, and schwannomatosis. We'll highlight the two scenarios that are marked in gray. For one of these individuals who was a 48-year-old with cranial schwannomas, a paternal great-aunt with spinal schwannomas, a first-degree relative with a meningioma and schwannoma, and then another, first cousin, excuse me, and a second-degree relative with a unilateral vestibular schwannoma. In this particular case, a tumor was sent in, and we were able to identify all three events that you would expect to see in a tumor associated with schwannomatosis, with the first hit being a star-82 C to T in SMARCB1. This is one of the most recurrent SMARCB1 variants and is associated with schwannomatosis, located in the 3' UTR region. And so we also were able to confirm that this SMARCB1 mutation segregated with the phenotype in the family. However, in a second scenario of an individual who was 24 with a unilateral vestibular schwannoma, and schwannomas that affected both the right and left foot, two tumors were submitted. And in this particular scenario, we did see all three events, but if you notice, the first hit in LZTR1 is actually a synonymous variant. Based on our history with RNA, we were able to do a free-of-charge RNA-based study that was able to confirm that this led to out-of-frame skipping of exon 15. So it required a complementary approach of multiple testing processes to gain a diagnosis for this individual, but it was essential to have the tumors in this case, because finding a synonymous variant in LZTR1 alone would not have given this individual their diagnosis. So in conclusion, for NF1, assessment of mosaicism allows more than one way to determine the more accurate genetic counseling of sporadic patients, and it allows to have good counseling when it comes to the recurrence risk for the parents, and also the transmission risk for the next generation. Low-level mosaicism for pathogenic NF1 variants with a variant allele frequency of 2 to 3 percent can be identified, and this will reduce the need for biopsy-based testing, but definitely will not eliminate it, because we know that there are quite a number of individuals, no matter how deep we will sequence, you will not identify their first hit mutation, and we know that because we have also gone back from our patients that we did by biopsy-based testing to look whether our approach would be able to identify it when it's present in the blood at a very low concentration. So although we are offering DNA-based NGS testing, the RNA-based testing remains very crucial in a number of scenarios, as Alicia nicely illustrated. The importance of genetic testing will probably increase as additional genotype-phenotype correlations become reported, and given the phenotypic overlap and genetic heterogeneity amongst all of the neurofibromatosis, it is increasingly recognized that an NGS panel-based testing may become useful, and especially if it is cost-effective. For NF2, year two, the assessment of mosaicism allows a more accurate genetic counting of the sporadic patient, similar to NF1. Low-level mosaicism for pathogenic NF2 variants have also been identified so far in quite a number of individuals with mosaic NF2, year two. It reduces, but it does not eliminate the need for tumor-based testing. A lot has been learned over the last 10 years regarding the genetics of schwannomatosis and the differences with NF2 and mosaic NF2, and it has become very clear that these disorders show so much phenotypic overlap. Now, at the medical genomics lab, we do recommend molecular analysis from two anatomically different tumors and simultaneous analysis for all three genes, NF2, SMARCB1, and LDTR1, both by sequencing and copy number variant analysis, with further verification in the blood for any variant that is found in the tumors. So, at the end, we would like to provide you with some references for further reading and acknowledge our current staff that you can see here illustrated by pictures, but we also want to thank our previous staff, previous students and fellows, Bruce Korf, who is the UAB chief genomics officer, the patients and the providers who have trusted us for their care, and our collaborators on genotype-phenotype correlations from around the world. Finally, also, I want to thank the support through the Children's Tumor Foundation by the Isaac and Sally Fuchs Genotype-Phenotype Study, and we will be happy to take questions. Thank you. So, we do have a couple questions that have already come in. The first was regarding whether or not the slots would be available after the talk. I think that is the ACMG that will address that. I thought that the theme was that the slides were available to those that have registered. Yes, and another question that has come in is regarding the list of 31 variants that have a frequency of higher than 0.4 percent. Is this published, or is it available? It's a list of 31 NF1 variants that have a frequency of... So, part of it is actually published. These 31 variants could be found in the chapter that we have authored back in 2009. It's a chapter that I wrote together with Professor Katarina Wimmer, a close collaborator from Innsbruck, Austria, and it is a chapter from the book Neurofibromatosis, so probably it is in my library. We can add this to the list of references. It is not a paper, it's a chapter. It is on my ResearchGate for those who want to go there. Another question has come in about... It's a very interesting connection between SUS-12 and NF1. Can you please point out the reference for further reading about this? Sure. So, we have your name and your affiliation, so we can send a number of papers on that. That is actually all very new still, but there are a couple of references that I can definitely send to you. Another question is regarding the percentage of missense variants of uncertain significance that were updated to splice site variants. I don't know that from the top of my head, though. I will have to pause on that. Yes. We can come back to that and respond privately. Let's see. What do you counsel a patient who does not share... A parent. Oh, excuse me. What do you counsel a parent who does not share a parent? What do you counsel a parent who does not share his child's NF1 mutation, but has a few undiagnosed sub-q masses removed and not available? Recurrence risk with gonadomoebiasis as a possibility? Well, actually, so in a case like that, so a few undiagnosed sub-q mass removed and not available. So you cannot even check whether this were neurofibromas, because that's what you suspect it to be if the child has a clear pathogenic NF1 mutation. Recurrence risk, if these are really subcutaneous masses, we are beyond gonadomoebiasis. And we may have to be thinking about gonosomal mosaicism, which is more, I mean, affecting obviously many more cells than gonadomoebiasis. So I think in a case like that, you at least need to mention that this is a possibility in a patient like that. Would you agree, Anisha? I would. We also have sperm-based analysis, if the parent is a male, that could also be done for targeted testing via next-generation sequencing to look for a low-level mutation that may or may not be present in the sperm. That can help provide reassurance to some families as well. So another question has come in. If a case is strongly suspected for mosaicism, but the mutation cannot be found in blood, tissue testing is not available, do you recommend testing the children? Well, we have done that in a few cases, haven't we, where the parents were very, very concerned that they didn't want to go through. So of course, if you're going to testing a child, we have the issue of normally a child who will have constitutional NF1 in this particular case. Because if the parent, of course, was a mosaic, if it's being passed on to the next generation, the next generation is a constitutional mutation. Normally, children will present with some signs. So if the children, none of them present with any signs related to NF1, this is very reassuring. Right? So in that respect, there is not an absolute definite advantage to go to testing. Some individuals are very concerned. You can do testing if that is something that they absolutely want. So then you take into account that the detection rate for germline mutations in NF1 hovers around 95%. And so you would still have a 5% post-negative rate. Would you consider it the same for NF2 with the later onset? So for the NF2, the question is, of course, different. And so for NF2, we have had this question more frequently. Because, of course, for NF2, if the patients do not present with the specific congenital cataract or with the dermal schwannomas early on, it can take until their 20s before they show any signs. So there is definitely value in knowing whether the child has inherited NF2. I think this concludes the questions that we have available right now. ACMG would also like us to remind everyone to please visit their website regarding upcoming meetings as well. And thank you to QIAGEN Bioinformatics and the Injunuity Clinical Decision Support Platform for the support of this program. And thank you, everyone, for attending. Thank you. Ladies and gentlemen, thank you for attending today. This concludes our program. You may now disconnect.

ACMG Genomics Case Conference

genomics education

genetic disorders

Neurofibromatosis type 1

Neurofibromatosis type 2

Schwannomatosis

diagnosis and testing

genotype-phenotype correlations

next-generation sequencing