Page 56 - 2021_03-Haematologica-web

S. Vantyghem et al.
of next-generation sequencing (NGS) platforms by the Institut National du Cancer (INCa) has enabled physicians to investigate these markers since 2013. Non-acute myeloid malignancies, i.e., myelodysplastic syndromes (MDS) and myeloproliferative neoplasms (MPN) represent good indi- cations for NGS because of their relatively high incidence of mutations and the ease with which they can be assessed in peripheral blood samples1-3 in time frames compatible with decision-making.
More than 40 genes are now known to be potentially recurrently mutated in MDS,1 mixed MDS/MPN syn- dromes2 and MPN.3 In a context of idiopathic cytopenia of undetermined significance (ICUS), the presence of somatic mutations can support the diagnosis of clonal cytopenia of undetermined significance (CCUS).4,5 In MDS, several mutations, including those of TP53, EZH2, ETV6, RUNX1 or ASXL1, have been shown to be strongly associated with decreased overall survival,6 independently of the risk group according to the Revised International Prognostic Scoring System (IPSS-R).7 Other studies have shown that the addition of molecular testing allows for a better prog- nostication of the outcome of patients with chronic myelomonocytic leukemia (CMML)8 or MPN.3 It has therefore been suggested that the IPSS-R and Dynamic International Prognostic Scoring System (DIPSS)9 scores could be improved by incorporating information on molecular abnormalities.10 This is possibly pertinent since targeted therapies are now emerging for the treatment of MDS and MPN, which was long limited to a few drugs of moderate efficacy, making the search for somatic muta- tions of relevance in order to be able to propose the best available treatment.11
Given the potential advantages of the broad screening capacity of NGS, this technique is now often used with- out proper published international guidelines. In order to explore the added value of somatic mutation testing by NGS, we conducted an observational multicenter study aimed at assessing the impact of somatic mutation testing by NGS in a real-life setting, focusing on the impact of implementation of this testing on diagnosis, prognosis and treatment in selected populations.
Methods
Patients
In our University Hospital, all NGS requests must be validated during regional multidisciplinary meetings. Indications for NGS analysis include challenging cases of suspected MDS or MPN in which the detection of somatic mutations could help the diagno- sis, prognostic assessment or search for theranostic markers. This occurs after morphology/pathology, flow cytometry and cytoge- netics have already been performed in order to apply World Health Organization (WHO) recommendations. All patients for whom NGS testing was prescribed between October 2014 and March 2019 were included in the present study. All provided their consent according to local ethical rules. In a first cohort of patients (group A), the indication for NGS was to confirm or rule out a sus- pected diagnosis. In a second cohort, (group B), therapeutic deci- sions were expected to be supported by the detection or not of prognostic somatic mutations.
Next-generation sequencing analysis
In order to detect somatic mutations, a customized, targeted panel of 34 genes (145 kbp) was applied to DNA extracted from
peripheral blood or bone marrow samples (Online Supplementary Table S1). The selection of these 34 genes followed INCa recom- mendations published in 2013 and updated in 201612 and was completed by a review of the literature. DNA libraries, built with the Haloplex® target enrichment protocol (Agilent Technologies, Santa Clara, CA, USA), were paired-end sequenced with a MiSeq® Instrument (Illumina, San Diego, CA, USA). Data were analyzed using an in-house pipeline including trimmed reads alignment to the GRCh34 human reference genome (February 2009 assembly) with BWA-MEM, tumor variant detection by three variant callers (GATK HaplotypeCaller, VarScan and SAMTools) and variant annotations with public databases (gnomAD, COSMIC, dbSNP, ClinVar) and in silico predictors in the case of unknown variants (CADD, SIFT, PolyPhen-2, MutationTaster), using updated ver- sions whenever available. Mutations were considered significant if they reached a good quality score (DP4), a minimum variant allele frequency of at least 1% and a minimum of 20 reads supporting the variant for hotspots or 50 reads for non-hotspot variants. Variants of undetermined significance (VUS) with no clear associ- ation to a disease were defined according to Li et al.13 Final reports were delivered to clinicians in real time, without mentioning vari- ants identified as polymorphisms.
Group A: the impact of next-generation sequencing on diagnosis
Clonal hematopoiesis was defined by the presence of at least one somatic mutation. The detection of clonal hematopoiesis was integrated with suspected diagnoses of MDS or MPN according to the criteria detailed below.
Idiopathic cytopenia of undetermined significance and clonal cytopenia of undetermined significance
In the case of ICUS with morphological evidence of dysplasia but without MDS-related cytogenetic abnormalities, the detection of clonal hematopoiesis was used to retain a diagnosis of MDS. In the context of ICUS14 without morphological evidence of dyspla- sia and without MDS-related cytogenetic abnormalities, the pres- ence of somatic mutations defined a diagnosis of CCUS.5
Chronic myelomonocytic leukemia
In cases of suspected CMML with or without minimal evidence of dysplasia, WHO criteria include the presence of acquired clonal cytogenetic or molecular genetic abnormalities (i.e., somatic muta- tions) as supportive of the diagnosis.15
Aplastic anemia
In a context of bone marrow hypoplasia, the absence of clonal hematopoiesis was used to exclude a diagnosis of hypoplastic MDS (hMDS). In addition, the detection of PIGA somatic muta- tions helped to make a diagnosis of aplastic anemia (AA).15,16
Myeloproliferative neoplasms
In cases of suspected MPN without BCR-ABL fusion and with- out any of the three classic driver mutations (JAK2V617F, CALR and MPL) as detected by classical methods, the WHO classification recommends searching for the most frequent accompanying mutations (e.g., ASXL1, EZH2, TET2, IDH1, IDH2, SRSF2 and/or SF3B1) to help determine the clonal nature of the disease.15
Group B: the impact of next-generation sequencing on prognosis
In group B, the diagnoses of MDS, MPN/MDS - including CMML - and MPN were established according to the WHO clas- sification.15
Scores of the IPSS-R,7 CMML-specific Prognostic Scoring
702
haematologica | 2021; 106(3)