End of a diagnostic odyssey: the added value of multi-tissue analysis in the identification of mosaicism in tumour predisposition syndromes Authors: L Damen, MF Broekema, MJ Vogel, E Korpershoek, MJ Koudijs, BR Boersma, KM Dreijerink, PJ Zondervan, HJ Dubbink, LP van Hest Published: 2026 ABSTRACT SummaryMosaicism refers to the presence of multiple cell clones with distinct genotypes arising from a single zygote. The phenotype of mosaic individuals depends on the extent of mosaicism, ranging from localised to almost generalised. We report three diagnostically challenging cases with previously unrecognised mosaicism in the tumour predisposition syndromes (TPS) von Hippel-Lindau syndrome, neurofibromatosis type 1 and neurofibromatosis type 2. In all three patients there was a strong clinical suspicion for a TPS; however, no pathogenic variant (PV) was detected in these patients in DNA extracted from blood. With next generation sequencing (NGS) analysis of multiple affected tissue samples of these patients we were able to detect a recurrent PV at varying variant allele fractions. In all three patients, low-grade mosaicism was in retrospect confirmed in normal tissue and blood (variant allele frequency: 1%–7%). Therefore, we conclude that NGS on multiple affected tissues is an effective strategy to detect low-level mosaicism. Identification of previously unrecognised low-grade mosaic cases not only allows for more precise diagnosis and management advice for the index patient, but also accurate genetic counselling for family members.Keywords: Mosaicism, Genetic Predisposition to Disease, Neoplasms Introduction Mosaicism refers to the presence of multiple cell lines with distinct genotypes in an individual,1 arising from postzygotic mitotic mutations at any developmental stage and in any cell type. Mosaicism is likely the rule rather than the exception in multicellular organisms.2 The vast majority of these mutational events do not have an apparent effect on the individual, because they occur in the non-coding regions or in tissue where expression of the mutant gene is not relevant.3 When mutations occur in coding regions, they can cause a de novo pathogenic phenotype. The phenotype of mosaic individuals depends on the extent of mosaicism, ranging from localised to almost generalised. The earlier the mutations occur in development, the more likely they are to affect the phenotype. Individuals carrying germline heterozygous loss-of-function pathogenic variants (PV) in tumour suppressor genes have an increased risk of certain tumours and sometimes non-tumour features. Most tumour predisposition syndromes (TPS) follow an autosomal dominant inheritance pattern. The mechanism for tumourigenesis is explained by Knudson’s two-hit model.4,6 Examples of TPS are von Hippel-Lindau syndrome (VHL), neurofibromatosis type 1 (NF1) and neurofibromatosis type 2 (NF2). VHL (OMIM 193300) is caused by PV in VHL and is characterised by haemangioblastomas, occurring in the brain, spinal cord and retina. Other features are renal cysts, clear cell renal carcinoma, pancreatic cysts, neuroendocrine tumours, pheochromocytomas and paragangliomas.7 8 NF1 is caused by PV in NF1 (OMIM 162200). NF1 is characterised by cutaneous and plexiform neurofibromas, multiple café au lait (CAL) macules and freckling in the axillary or inguinal regions. Other major features include Lisch noduli and distinctive osseous lesions as anterolateral bowing of the tibia and sphenoid wing dysplasia.9 PV in NF2 causes NF2 (OMIM 101000). Bilateral vestibular schwannomas are the most common feature. Other benign tumours can occur, including schwannomas of peripheral nerves or other cranial nerves, ependymomas and meningiomas.10 11 Mosaicism has been previously described in various TPS, including VHL, NF1 and NF2.12,19 Especially in NF1 and NF2, mosaicism is well known and has a high prevalence.12 13 In general, when a TPS is suspected, step one in the diagnostic process is molecular genetic analysis of a blood sample to confirm the diagnosis. However, in mosaic cases, the causative PV is often present with a variant allele frequency (VAF) at less than heterozygous levels (<50%) in blood and can even be completely absent in blood. In addition, the phenotype might be more subtle or localised. As a result, mosaic patients might not meet the clinical diagnostic criteria for the disease. Furthermore, family history is often negative because the PV is de novo. The combination of a subtle phenotype, negative family history and difficulties in detecting a PV in blood can lead to a diagnostic odyssey. In this report, we report three diagnostically challenging cases with previously unrecognised mosaicism in VHL, NF1 and NF2. Methods Index patients were counselled at the Department of Clinical Genetics from the Amsterdam University Medical Centers. Index patients gave consent for publication of non-identifiable details. DNA sequence analysis of peripheral blood samples In all cases, genetic analysis on peripheral blood samples was performed in a diagnostic accredited laboratory. Genomic DNA was extracted from peripheral-blood leucocytes in venous blood samples. In 2020, analysis of FH, MAX, MDH2, RET, SDHA, SDHAF2, SDHB, SDHC, SDHD, TMEM127 and VHL was performed in case 1 by the Illumina next generation sequencing (NGS) system. For case 2, NF1 and SPRED1 were investigated by Sanger sequencing and Multiplex Ligation-dependent Probe Amplification (MLPA). For case 3, analysis of LZTR1, NF1, NF2, SMARCA4, SMARCB1, SMARCE1, SUFU, TSC1 and TSC2 was performed on the Illumina platform. DNA sequence analysis of tissue samples DNA was extracted from manually-macrodissected formalin-fixed paraffin-embedded normal or tumour tissues using proteinase K digestion and 5% Chelex resin, as previously described.14 NGS analysis was performed using the Ion Torrent sequencing system (Thermo Fisher Scientific, USA) with different diagnostic NGS panels; for case 1, a panel encompassing the entire coding regions of HIF2A, KIF1B, MAX, NF1, PDH2, RET, SDHA, SDHAF2, SDHB, SDHC, SDHD, TMEM127 and VHL, and for cases 2 and 3, a panel encompassing the entire coding regions of ATRX, CDKN2A, CDKN2B, CIC, DDX3X, FUBP1, MLH1, MSH2, MSH6, NF1, NF2, PDGFRA, PIK3R1, PTCH1, PTEN, RB1, SETD2, SMO, SUFU, TP53 and TRAF7 was used. Loss of heterozygosity (LOH) was assessed by analysis of SNPs distributed over several chromosomes, as previously described.15 Results Case 1 – VHL mosaicism A 52-year-old male, with a negative family history, initially presented at 23 years with a pheochromocytoma in the left adrenal gland (figure 1A). In the following years, multiple surgeries were required for new and recurrent diagnosed pheochromocytomas and paragangliomas. A TPS was suspected. Despite extensive genetic testing in blood in 2000 and reanalysis in 2020, no underlying genetic cause was identified. As a next step, NGS was performed on DNA retrieved from two pheochromocytomas (left adrenal gland (2005) and para-aortic lesion (2012)). This approach revealed a PV in VHL (NM_000551.4: c.499C>T p.(Arg167Trp)) at VAFs of 47% and 39%, respectively (figure 1A). In the DNA from the pheochromocytoma of the adrenal gland also LOH of the VHL locus was observed, this could not be determined for the other tumour. NGS analysis of DNA from normal adjacent tissue, that is, large blood vessels of the adrenal gland and lymph nodes near the para-aortic lesion, showed the same VHL PV at VAFs of 6.2% and 11%, respectively. Re-examination of the NGS data from blood DNA identified this PV in a VAF of 7.8%, confirming VHL mosaicism. After his genetic diagnoses, the patient was diagnosed with clear cell renal cancer and spinal haemangioblastoma. His son opted for genetic testing. He did not inherit the PV. Diagnostic process of the patients. The timeline of the patients indicating the clinical phenotype (green), treatment (red), genetic consultation (orange) and other consultations (blue). (A) Case 1: IGV displays the pathogenic VHL variant c.499C>T p.(Arg167Trp) in pheochromocytoma (left side, left adrenal gland; right side, para-aortic), adjacent normal tissues and blood. The percentage of the sequencing reads (grey bars) depicting the variant is indicated for different tissues. The percentage of tumour cells was estimated by a pathologist as 80% and 70% for the lesion in the adrenal gland and the para-aortic lesion, respectively. (B) Case 2: IGV displays the pathogenic NF1 variant c.3826C>T p.(Arg1276Ter) in the four neurofibromas, blood and buccal mucosa. The percentage of the sequencing reads (grey bars) depicting the variant is indicated for different tissues. The percentage of tumour cells was estimated by a pathologist as 90% in all neurofibromas. (C) Case 3: IGV displays the pathogenic NF2 variant c.592C>T, p.(Arg198Ter) in meningioma, schwannoma and blood. The percentage of the sequencing reads (grey bars) depicting the variant is indicated for different tissues. The percentage of tumour cells was estimated by a pathologist as 90% for the meningiomas and 60% for the schwannoma. IGV, Integrative Genomics Viewer; CALM, café au lait cafe macules.MEN2, Multiple Endocrine Neoplasia type 2. VHL, von Hippel-Lindau syndrome. Case 2 – NF1 mosaicism A 35-year-old female was referred by her dermatologist because of a suspicion of NF1 (figure 1B). She had >10 CAL macules in varying sizes (0.5 to >7 cm). The patient had very subtle freckling in the axillae, which was more evident on the right than the left. On the trunk, multiple papules suspected to be neurofibromas were seen. Family history was negative for NF1 or CAL maculae. NF1 and SPRED1 were investigated, but no PV was found. At 36 years, ophthalmological examination revealed Lisch noduli in the right eye and four papules from trunk and breast were histologically confirmed to be neurofibromas. As a result, this patient now met the clinical criteria for NF19. Subsequently, NF1 analysis by NGS on the four neurofibromas identified an NF1 PV (NM_000267.3: c.3826C>T p.(Arg1276Ter)) at varying VAFs, ranging from 22% to 34%. NF1 mosaicism was confirmed by detection of this PV in 5.8% and 5.4% of the reads in DNA from buccal mucosa and blood, respectively. Following the identification of the PV, the patient’s children were tested and did not inherit the PV. Case 3 – NF2 mosaicism The third patient was diagnosed with bilateral vestibular schwannoma, multiple intracranial and spinal meningiomas and a schwannoma at the age of 38 years, thereby fulfilling the criteria for a clinical diagnosis of NF2 (figure 1C).10 Despite genetic testing in blood in 2000 and reanalysis in 2021 the NF2 diagnosis could not be confirmed. Furthermore, no other genetic cause could be identified. The multiple intracranial tumours caused neurological dysfunction requiring surgery. NF2 analysis by NGS of the two removed meningiomas and a previously removed schwannoma revealed a PV (NM_000268.3: c.592C>T p.(Arg198Ter)) in VAFs ranging from 45% to 70%. All three lesions showed LOH of the NF2 locus. In retrospect, NF2 mosaicism was confirmed in blood (1% (14/1555) of the reads). Her son, who underwent NF2 screening as a child and young adult, opted for testing and did not inherit the PV. Discussion In three independent cases with a strong clinical suspicion of a TPS, we identified a low-grade mosaic VHL, NF1 and NF2 PV via multiple tissue sampling combined with deep sequencing. These cases illustrate that recent developments in somatic DNA testing can be complementary to germline DNA testing when a germline DNA test fails to detect the causative PV in blood despite a strong clinical suspicion for a TPS. Mosaicism, described in various TPS,16,18 can be diagnostically challenging, as illustrated by the lengthy journey to diagnosis in our three cases. The first challenge is to identify tissues containing the causative PV. As the presence of the PV in mosaic cases may be limited to specific tissues, a potential PV can be missed if only a blood sample is investigated. In some TPS, the affected tissues can easily be recognised and tested, like the neurofibromas in NF1. For other tumours, like paragangliomas in VHL, imaging techniques might help to identify affected tissues. However, when specific clues for affected tissue are absent, detection of mosaicism is more difficult, and to detect low-level mosaicism, analysing multiple tissue is often necessary. Another challenge can be to obtain a sample of the tissue of interest. For example, to reach intracranial tumours in NF2 or cerebellar haemangioblastomas in VHL might be difficult or even impossible. Also, sometimes for other tumours the affected tissue might not (yet) be available due to an intended wait-and-scan strategy. NF2 has a high rate of mosaicism. Evans et al found a high rate of NF2 mosaics at VAFs between 0.2% and 30% in blood in 1055 NF2 cases who met clinical criteria for NF2 with a negative family history.12 Mosaic cases in which the PV is absent in peripheral blood or those with a VAF <10% are likely to be missed in routine genetic diagnostics, as these tests are focused on detecting heterozygous PV with VAFs ~50%. As a result, filter settings often filter out variants with a low VAF. In addition, the lower number of reads generated in routine diagnostics is generally not suitable to detect low level mosaicism. The three cases presented here illustrate the complementary use of somatic DNA testing. Our VHL case adds to the few reported case studies on VHL mosaicism.1619,21 Recently, no VHL PV was identified in the blood of an 11-year-old with an apparently sporadic brainstem haemangioblastoma. Somatic VHL testing after developing additional haemangioblastoma confirmed VHL mosaicism.20 These VHL cases illustrate that multiple tissue testing with combined deep sequencing should be considered early in the diagnostic process in case of a high clinical suspicion for VHL, as it can significantly shorten the time to diagnosis. This is especially important in young patients in which the early clinical presentation is suspicious of VHL. Mosaicism has also been described in other TPS, like Li-Fraumeni syndrome, tuberous sclerosis complex and PTEN hamartoma tumour syndrome.16 18 Therefore, the combination of multiple tissue testing and deep sequencing can also be used to search for mosaicism in other TPS.16 Accordingly, we recommend a similar approach involving multiple tissue sampling and deep sequencing for patients with a high clinical suspicion of other TPS. Detection of mosaicism can have important medical and psychological consequences for patients and their families. First, confirmation of mosaicism can provide more accurate genetic counselling for patients themselves. Second, it can give patients of childbearing ages more information regarding the risk of passing on the TPS, which in general is considered to be lower in mosaic cases than in full heterozygous cases.16 In addition to this, it provides these patients with the option of invasive prenatal testing for the PV.12 13 16 20 Third, identification of the mosaic PV provides children of affected patients the opportunity to perform a DNA test to determine whether they inherited the PV. Only descendants who inherited the NF2 or VHL mutation will then be advised screening instead of screening all children of parents with a clinical diagnosis of NF2 and VHL lacking a molecular diagnosis. In addition, confirmation of a mosaic PV is also informative for siblings. As this means it is a de novo PV and screening is no longer recommended for siblings. In the era of personalised medicine with the development of new treatment options for TPS-associated tumours, detection of mosaicism also becomes more important. In addition, but equally important, a clear genetic diagnosis offers patients acknowledgement and confirmation of their symptoms with subsequent psychological benefits. Lastly, shortening the time to diagnosis can save patients the psychological burden of a diagnostic odyssey. An important limitation of the suggested approach is that multiple tissue testing combined with deep sequencing might not be possible for every patient suspected of having a mosaic TPS. As described above, it is not always possible to retrieve multiple tissues from a patient for somatic DNA testing. In patients with only one tumour or tissue of interest (eg, CAL) available for sampling, it might still be possible to detect mosaicism. A possible strategy in those cases could be to investigate that tissue for somatic mutations, followed by deep sequencing (>100 reads) of DNA extracted from blood, with special attention to the variants identified in the tissue. If one of these variants is detected in the DNA extracted from blood at a low percentage of the reads, this confirms mosaicism. In addition, depending on the technique used for genetic analysis, the number of reads available for analysis might be limited. For instance, in our NF2 case, the variant was only detected in 14 of the 1555 reads. Unfortunately, the resources for such deep sequencing are not always available. Lastly, the costs of this combined approach can also be a limitation. Given the limitations and costs, this approach should currently be reserved for cases strongly suggestive of mosaic TPS, especially in patients who meet the clinical diagnostic criteria for these TPS, without affected siblings or parents. For example, for NF1 and VHL, this approach can also be considered if one of the clinical criteria is met. For NF2, we would suggest this approach in case of a unilateral vestibular schwannoma and a second NF2-related tumour. In conclusion, recent advances in NGS along with the opportunity of multi-tissue sampling are an effective strategy in the detection of low-level mosaicism, enabling earlier and more accurate diagnosis, refined genetic counselling and optimised clinical management for patients and family members. Footnotes Funding: The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors. Patient consent for publication: Consent obtained directly from patient(s). Provenance and peer review: Not commissioned; externally peer reviewed. References