Genetic testing in the diagnosis and management of patients with atypical myeloproliferative neoplasms

Following our summary on the recommendations for genetic testing in common myeloproliferative neoplasms (MPN) based on the British Society for Haematology (BSH) Good-Practice Paper published by Cross et al.,¹ the present summary will focus on the atypical MPN subtypes including chronic eosinophilic leukemia (CEL), myeloid/lymphoid neoplasms with eosinophilia (MLN-eo), chronic neutrophilic leukemia (CNL), MPN-unclassifiable (MPN-U), mastocytosis, and myelodysplastic syndromes (MDS)/MPN overlap.

Briefly, Good-Practice Papers aim to recommend good practice in areas where the evidence is limited but a degree of consensus or uniformity may improve patient care using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) nomenclature.

CEL and MLN-eo

When a persistent eosinophilia (at least 1.5 × 10⁹/L) with no secondary cause is identified, initial investigation for FIP1L1-PDGFRA on peripheral blood or bone marrow (BM) by fluorescence in situ hybridization (FISH) or nested reverse transcriptase (RT) polymerase chain reaction (PCR) is recommended (Grade 1B).

• It should be noted that either technique alone may miss cases, and so both or an additional testing should be considered if there is a high clinical suspicion.
• If FIP1L1-PDGFRA is not detected in case of a suspected myeloid neoplasm, then BM cytogenetics should be performed.
• Due to the diversity of fusions, effective targeted RT-PCR may be unlikely, but an increasing number of cases are being detected by broad/targeted RNAseq screens
  • This approach is currently too expensive to be a general screening tool and so should be reserved for exceptional cases.
• Break-apart FISH analysis for specific loci (PDGFRA, PDGFRB, FGFR1, JAK2 for MLN-eo; ABL1, FLT3, ETV6, other tyrosine kinase [TK] genes for CEL) may also be used to identify disruption of key loci.

BM cytogenetics or FISH are recommended to screen for FIP1L1-PDGFRA and other fusion genes, but this must then be confirmed by molecular methods (Grade 1B).

• The timing of tests should follow tests recommended for chronic myeloid leukemia (CML) and more frequent testing should be considered for those attempting treatment-free remission.

In case of elevated serum tryptase levels without a TK gene fusion, mastocytosis should be considered, and examination of BM histology is essential in this context. Myeloid gene panel and KIT D816V testing should be considered for patients with persistent unexplained eosinophilia who test negative for fusion genes (Grade 2B) (Table 1).

Table 1. Markers of clonality associated with eosinophilia*

CNL, MPN-U

CNL is strongly, but not exclusively, associated with CSF3R mutations which are key diagnostic features, but there is also significant overlap between CNL and MDS/MPN.

In the context of mutated genes, testing for CSF3R variants, preferably as part of a more comprehensive myeloid panel, is recommended for all patients with suspected CNL (Grade 2B).

MPN-U is a rare subtype, mostly consisting of patients who do not meet the diagnostic criteria for a specific MPN subtype, or those with features overlapping with ≥2 subtypes. Most patients test positive for JAK2 V617F, CALR or other myeloid driver mutations.

Mastocytosis

KIT D816V mutation is present in most (up to 90%) adult patients with systemic mastocytosis (SM); however, variant allele frequency (VAF) is often too low for detection by next-generation sequencing (NGS). Therefore, more sensitive techniques (i.e., RT-PCR or digital PCR) should be used on peripheral blood or BM samples. It is also possible to consider standard mutation analysis on purified mast cell samples.

Sensitive KIT D816V testing is recommended for all patients with a suspected mastocytosis (Grade 1B). If KIT D816V is negative, screening for other KIT mutations should be considered in adult patients and it is recommended in pediatric patients (Grade 1B), as KIT D816V is only seen in 30–50% of children with mastocytosis.

• KIT D816V can be detected in peripheral blood in most patients. BM analysis should be considered if analysis of peripheral blood is negative and the index of suspicion is high.
• If BM analysis is negative for KIT D816V but the index of suspicion is high, then clinicians should consider wider screening for D816 variants or other KIT mutations.

Apart from the KIT driver mutation, additional somatic mutations are noted in 70–90% of advanced SM cases, and most will have an associated hematologic neoplasm (AHN) which is usually of an MDS/MPN subtype.

• Mutations are less frequent (<20%) in patients with indolent SM (ISM) and mutations in ASXL1, RUNX1, and/or DNMT3A with high VAF (≥30%) have been associated with poor prognosis; however, the value of routine testing is unclear.
• Mutations in SRSF2, ASXL1, RUNX1, EZH2, and NRAS are found in advanced SM and associated with an adverse prognosis, so molecular profiling can be useful when there are transplant decisions to be made.
• Some patients with SM-AHN (~25%) harbor an abnormal karyotype which can be associated with poor prognosis (poor-risk karyotype, e.g., monosomy 7, complex karyotype). However, there is not a consensus if this is independent of mutation status.

Certain mutations are associated with adverse prognosis, and as a result, profiling can inform decision making regarding therapy and transplantation.

Myeloid panel analysis is advised in patients with advanced SM who are candidates for allogenic stem cell transplantation (allo-SCT) (Grade 1B).

Myeloid panel analysis should be considered for other SM patients if the behavior of the disease may influence therapy (Grade 2B). Myeloid panel and/or BM cytogenetics should be considered to allow characterisation of the AHN component of SM-AHN (Grade 2B).

MDS/MPN overlap

Adult MDS/MPN overlap syndromes include chronic myelomonocytic leukemia (CMML), atypical CML BCR-ABL1-negative (aCML), MDS/MPN with ring sideroblasts and thrombocytosis (MDS/MPN-RS-T), and MDS/MPN-U.

• Diagnosis is heavily reliant on BM morphology and clinical assessment.
• Molecular genetics can help provide key information to aid diagnosis, subclassification, and prognostic information.

Initial investigations in suspected MDS/MPN overlap syndromes

It is essential to exclude BCR-ABL1 in all cases, and rearrangements of PDGFRA, PDGFRB, FGFR1, or PCM1-JAK2 in rare cases with an associated eosinophilia (Grade 1B). Cytogenetics should be performed at time of diagnosis as it can demonstrate clonality and provides information on prognosis, and can be useful to rule out any rare TK fusions which can mimic MDS/MPN.

• Cytogenetic abnormalities are seen in 30–50% of cases using conventional karyotyping.
  • The most common abnormalities include +8, +9, −7, del7q, del20q, del13q, and isochromosome 17q, and FISH is recommended as a minimum requirement for chromosomes 7 and 8.
  • If not undertaken or if the result has failed, then SNP array analysis should be considered as this can increase the yield of detecting an abnormality to 75%.

Testing for additional somatic driver mutations with myeloid gene panels

In patients with inconclusive morphology

Somatic mutations are reported in >90% of cases across the MDS/MPN overlaps. Therefore, the presence of additional mutations can provide supportive evidence of clonality and inform diagnostic discussions.

• There is concern over mutational analysis as there are frequent somatic mutations present in the healthy older population.
  • Myeloid gene panels are advised in challenging diagnostic cases, and the presence of two mutations, one of which with VAF over 20%, is supportive of a diagnosis.
• Genes included in the National Genomic Test Directory (NGTD)² for suspected MDS/MPN overlaps and CMML are the following:
  • KRAS, NRAS, TP53, JAK2, CALR, MPL, ASXL1, CBL, ZRSR2, CHEK2, CSF3R, CUX1, DNMT3A, EZH2, IDH1, IDH2, IKZF1, KIT, NFE2, SF3B1, SH2B3, SRSF2, TET2, U2AF1, HRAS, RUNX1, SETBP1, ZRSR2, BCOR, PTPN11, FLT3, NF1, and NPM1.

• The minimum genes recommended for investigation of patients with suspected CMML include: TET2, ASXL1, DNMT3A, EZH2, IDH1, IDH2, BCOR, CBL, KRAS, NRAS, NF1, JAK2, RUNX1, SETBP1, NPM1, FLT3, SRSF2, SF3B1, U2AF1, and ZRSR2.

• In addition to the NGTD requirements, larger panels may provide additional information (e.g., NPM1 abnormalities are uncommon in MDS/MPN but do identify cases that may transform to AML).

Confirmed diagnosis of MDS/MPN

A myeloid gene panel and BM cytogenetics or SNP array are recommended for patients diagnosed with MDS/MPN or cases with suspected MDS/MPN but with indeterminate morphology (Grade 1B).

The most frequent genetic mutations seen in MDS/MPN are not specific to these disorders and patients can be further classified investigating genotypic/phenotypic correlations.

• SRSF2, TET2, and ASXL1 are the most common mutations in CMML, and a combination of TET2 mutation and either SRSF2 or ZRSR2 is highly specific for the myelomonocytic phenotype.
• An aCML diagnosis can be supported by the presence of mutations in SETBP1 and/or ETNK1 (which are less common in CMML or MDS/MPN-U) and a lack of MPN driver mutations (JAK2, CALR, and MPL).
• SETBP1 is also mutated in CNL.

• SF3B1 and JAK2 mutations are reported in MDS/MPN-RS-T in up to 90% and 57% of cases, respectively, where CALR or MPL mutations are rare.
• Elevated tryptase and/or mast cell abnormalities in MDS/MPN is suggestive of SM-AHN, and diagnosis can be supported by investigation of KIT D816V.

Gene panels can provide information on prognosis and are incorporated into prognostic scoring systems across these diseases.

• ASXL1, NRAS, RUNX1, and SETBP1 are independently associated with a worse overall survival (OS) in CMML. The number of mutations is another risk factor for poor OS. ASXL1 and/or NRAS mutations are associated with worse survival following stem cell transplantation (SCT).
• ASXL1 and SETBP1 mutations are also associated with poor prognosis across other MDS/MPN.

• SETBP1 associated with adverse clinical phenotype and significantly worse OS in aCML
• SETBP1 and ASXL1 mutations are associated with poor survival in MDS/MPN-RS-T

Based on the above information, a targeted sequencing panel is recommended in patients diagnosed with an MDS/MPN overlap disorder, particularly in patients who are candidates for active treatment or transplant. This analysis can also inform prognosis and treatment decisions; therefore, it may be useful even if a patient is on supportive care.

Also, due to a strong correlation between mutations identified in the peripheral blood and BM, especially in CMML, mutational analysis can be a valuable tool in elderly patients or those unfit for BM biopsy.

Conclusions

Mutational analysis represents a valuable component of the diagnostic tool kit to classify patients with atypical MPN subtypes further, and to exclude certain disorders. Investigations using gene panels can also inform decisions around prognosis, and treatment. Genomic and genetic profiling are useful to shed a light on the complex nature of these disorders.