Guidelines for the investigation and management of Transient Leukaemia of Down Syndrome

Abstract

This guideline was compiled according to the British Society for Haematology (BSH) process at (http://www.bcshguidelines.com). The Grading of Recommendations Assessment, Development and Evaluation (GRADE) nomenclature was used to evaluate levels of evidence and to assess the strength of recommendations. The GRADE criteria can be found at http://www.gradeworkinggroup.org. Ovid MEDLINE and Ovid EMBASE were searched systematically for publications in English from 1980 to the end of 2015 using the key words Transient Abnormal Myelopoiesis, Transient Myeloproliferative Disorder, Transient Leukaemia, and Down Syndrome. Specific searches relating to fetal disease and hepatic parameters were also performed. References from relevant publications were also searched. The guideline group was selected to be representative of UK-based medical experts with invited representatives from the British Association of Perinatal Medicine and the Royal College of Paediatrics and Child Health. Review of the manuscript was performed by the BSH Guidelines General Haematology Task Force, the BSH Guidelines Committee and the General Haematology sounding board of BSH. It was also placed on the members section of the BSH website for comment. Further comments were invited from a sounding board of the Childhood Leukaemia Clinicians'27 Network, the Childhood Cancer and Leukaemia Group (CCLG), the Royal College of Paediatrics and Child Health, the British Association of Perinatal Medicine (BAPM) and patient representatives identified through the Down Syndrome Association; these organisations do not necessarily approve or endorse the contents. The objective of this guideline is to provide healthcare professionals with guidance on the investigation and management of patients with Transient Leukaemia of Down Syndrome (TL-DS). Individual patient circumstances may dictate an alternative approach. This is the first BSH guideline on this topic and is in date at time of publication. Any updates will be posted on the BSH Guidelines website (http://www.bcshguidelines.com). Between 5% and 30% of children with Down syndrome (DS) are born with transient leukaemia of Down syndrome (TL-DS), also known as transient abnormal myelopoiesis (TAM) and transient myeloproliferative disorder (TMD), a clonal disorder characterised by circulating megakaryoblasts and dysplastic changes in peripheral blood (PB) cells (Zipursky, 2003; Pine et al, 2007; Roberts et al, 2013). TL-DS is driven by mutations in the haematopoietic transcription factor gene GATA1 and is only seen in conjunction with trisomy 21, either constitutional or acquired. TL-DS may present with overt clinical features but some cases are only identified through examination of the blood film and/or by GATA1 mutation analysis (Klusmann et al, 2008; Roberts et al, 2013). Although many cases resolve without treatment, TL-DS results in early death in 15–23% cases and 20–23% of survivors will develop acute myeloid leukaemia of Down syndrome (ML-DS) in the first 4 years of life. Overall, TL-DS has an event-free survival of 63–68% (Massey et al, 2006; Klusmann et al, 2008; Gamis et al, 2011). Despite the very significant mortality and morbidity associated with the condition, care has not been standardised in the UK and many children do not receive the specialist care that is standard for all other paediatric malignancies. These guidelines aim to provide an evidence-based approach to the investigation and management of TL-DS and lay out clear treatment pathways to allow all children to receive the best possible care. TL-DS is a congenital leukaemia unique to neonates with DS or mosaic trisomy 21. The terms transient abnormal myelopoiesis (TAM) and transient myeloproliferative disorder (TMD) are also used to describe TL-DS but these terms can give a misleading impression of benignity. TL-DS displays many features of a malignant condition: TL-DS cells spread throughout the body, infiltrating the liver, pleural and pericardial spaces, skin and, to a lesser extent, the bone marrow. Despite its malignant nature, in the UK, care has not always been given by specialist Paediatric Oncology Principal Treatment Centres, potentially leading to mis- or delayed diagnosis, delayed treatment and avoidable death. TL-DS is marked by the presence of an acquired N-terminal mutation in exon 2 or exon 3 of the key haematopoietic transcription factor gene GATA1, resulting in a truncated GATA1 protein (GATA1s) (Groet et al, 2003; Hitzler et al, 2003; Mundschau et al, 2003; Rainis et al, 2003; Xu et al, 2003; Ahmed et al, 2004; Alford et al, 2011). Paired TL-DS and ML-DS samples show the same GATA1 mutation(s), indicating that they are clonally linked conditions (Ahmed et al, 2004; Yoshida et al, 2013). GATA1 mutations are not detected in remission samples after treatment of ML-DS nor are they present in other DS and non-DS leukaemias (Wechsler et al, 2002). Furthermore, GATA1 mutation(s) are not leukaemogenic in cells that are not trisomic for chromosome 21 (Hollanda et al, 2006). Studies using next generation sequencing (NGS) indicate that cases classified clinically as TL-DS (Yoshida et al, 2013) or by blast % (>10%; Roberts et al, 2013) all have detectable GATA1 mutations. The failure to demonstrate GATA1 mutations in clinically suspected TL-DS is likely to be due to one or more technical factors (e.g. a large GATA1 deletion, lack of assay sensitivity or a sample with a low blast %), although some cases are reported where a mutation cannot be demonstrated even after extensive investigation (Schifferli et al, 2015). The World Health Organization (WHO) Classification of Tumours of Haematopoietic and Lymphoid Tissues (Swerdlow et al, 2008) recognises the unique clinical and molecular features and the central role of GATA1, and defines TAM (TL-DS) as ‘increased peripheral blood blast cells in a neonate with Down syndrome'. No definition of increased peripheral blasts is offered. The Oxford-Imperial Down Syndrome Cohort Study (OIDSCS), which included a systematic examination of blood findings in neonates with DS together with sensitive GATA1 mutational analysis, found that 98% of neonates with DS had circulating blasts, the great majority of whom had no clinical features of TL-DS and no detectable GATA1 mutation (Roberts et al, 2013). Of note, no neonate without a detectable GATA1 mutation had either clinical features of TL-DS or subsequently developed ML-DS (Roberts et al, 2013). For these reasons, and as further discussed below, we recommend that, in keeping with other myeloid leukaemias in the WHO Classification, TL-DS is primarily defined on a genetic basis – the presence of a GATA1 mutation in a neonate with DS or mosaic DS – combined with an increased blast count (see below) or features suggestive of TL-DS. There is no internationally agreed definition of a percentage blast threshold that constitutes ‘increased peripheral blood blast cells’. The only prospective study of neonates with DS to evaluate the clinical significance of the blast percentage in neonates with DS is the OIDSCS (Roberts et al, 2013). The interim analysis of the first 200 neonates enrolled in the study, supported by the recently updated analysis, has shown that a threshold of >10% peripheral blood blasts in the first week of life identifies all neonates with clinical features of TL-DS (Roberts et al, 2013; Bhatnagar et al, 2016). However, some neonates with DS with blasts >10% do not have a GATA1 mutation even when very sensitive (NGS)-based methods are used. Using the updated Oxford study data, the sensitivity and specificity of blasts >10% (for the presence of GATA1 mutations) is 74% and 81% respectively (Bhatnagar et al, 2016; and unpublished data). Higher blast thresholds are likely to be more specific for TL-DS based on previously published retrospective studies and on the OIDSC study. In the OIDSC study all neonates with blasts >20% had a GATA1 mutation. Therefore, current data suggest that setting a blast threshold of >10% will identify more cases of TL-DS and that GATA1 mutation analysis is particularly important in neonates with blasts of 10–20% to prevent over-diagnosis of TL-DS. Blast count assessment requires careful examination of a peripheral blood film in the first week of life, ideally in the first 3 days of life, by a haematologist experienced in reviewing neonatal blood films. We recommend referral of blood films from any cases with suspicion of TL-DS for morphology review by a paediatric haematologist. Automated blast counts are not accurate and blasts are often missed. Blast count assessment after the first week of life may underestimate the prevalence of disease as, in our experience, the blast % often falls rapidly after birth and we recommend blast count assessment as soon as possible after birth to prevent delay in diagnosis of clinically relevant and life-threatening cases of TL-DS. Care should also be taken in neonates with intrauterine growth restriction (IUGR) or other history of placental insufficiency (e.g. maternal hypertension, pre-eclampsia or diabetes mellitus) as these babies may have lower blast counts despite large mutant GATA1 clones. From their origin in the fetal liver, megakaryoblastic TL-DS cells can spread locally, spill into the peripheral blood and infiltrate throughout the liver as well as distant tissues. This usually manifests as enlargement of the liver; as malignant effusions in pleural and pericardial spaces; and/or as a papular or vesicopustular rash due to deposits containing TL-DS blast cells in the skin. Skin nodules in TL-DS also occur but reports are rare (Winckworth et al, 2012). Splenomegaly is found in 30% of cases, although this is often due to portal venous obstruction (Gamis & Smith, 2012) as splenic infiltration is rarely reported (Yagahashi et al,1995; Smrcek et al, 2001). Thus, TL-DS can present with a spectrum of abnormalities ranging from a few circulating blast cells in an otherwise well neonate to hyperleucocytosis, hepatic fibrosis and multi-organ failure (Massey et al, 2006; Klusmann et al, 2008; Muramatsu et al, 2008; Gamis et al, 2011). No single clinical feature is entirely specific to TL-DS because each of these features may also occur in the absence of TL-DS (see Table 1). However, there are several characteristic features that are seen relatively frequently in TL-DS but are uncommon in DS neonates without GATA1 mutations, including organomegaly, hepatopathy (raised with skin pericardial and pleural and (Klusmann et al, 2008; Roberts et al, 2013). of one or more of these features in the absence of a clear alternative should to the early of a diagnosis of TL-DS. TL-DS from abnormal in the fetal liver et al, 2008; et al, 2008; et al, 2012). blast cells are often and in keeping with their and are often a feature (Roberts et al, 2013). they have a from other of and myeloid and and as well as of and and low of et al, Klusmann et al, 2008; et al, 2013). The of blast cells in neonates with DS is from that of blast cells in neonates without DS where these are not However, no of blast cells from blasts without GATA1 mutations has been reported in DS (Roberts et al, 2013). examination is not in blast cells are to in the liver and blasts are and in peripheral not with disease (Massey et al, 2006; Klusmann et al, 2008; Gamis et al, 2011). for GATA1 mutation are (Ahmed et al, 2004; Alford et al, and (Roberts et al, 2013). for the sequencing and methods of of these methods have technical and but sequencing and are not sensitive to mutant GATA1 that may be clinically significant that may to (Roberts et al, 2013). Using the prevalence of mutant GATA1 in neonates with DS DS reported from the OIDSCS (Roberts et al, 2013) has been on a sample from the same study Bhatnagar et al, 2016). The OIDSCS found that at of DS neonates with GATA1 mutations do not have blast >10% and have no clinical features of TL-DS. The prevalence of GATA1 mutations in DS neonates with blast of is of these children developed any clinically significant from TL-DS. The of one or more GATA1 mutation with no increased blast percentage in a neonate with DS has been TL-DS or TAM (Roberts et al, 2013; Bhatnagar et al, 2016). neonate with TL-DS has been reported to have to ML-DS (Roberts et al, 2013) but unpublished data from the Oxford study show a of of lower the seen in clinical TL-DS (Massey et al, 2006; Klusmann et al, 2008; Gamis et al, et al, unpublished data). analysis of the prospective OIDSCS has found no cases of ML-DS in neonates with DS without a GATA1 mutation detectable by at birth (Bhatnagar et al, 2016). these data indicate that with DS at birth with a sensitive for GATA1 mutations should identify all children at of However, there is no evidence of the clinical and of an approach. Although cases of TL-DS resolve without prospective studies of clinical TL-DS an early mortality of 15–23% (Massey et al, 2006; Klusmann et al, 2008; Muramatsu et al, 2008; Gamis et al, 2011). in the prospective studies of TL-DS combined with the large retrospective study, neonates with TL-DS Table This with mortality in the first after diagnosis of all in in the UK in 2013). of mortality in the UK from for to for acute myeloid leukaemia 2013). This that TL-DS has an early mortality in of any other in the The of death is a hepatopathy with leading to hepatic and findings show hepatic fibrosis with extensive and a infiltrate of megakaryoblasts et al, due to hepatic fibrosis is reported in of cases, (Massey et al, 2006; Klusmann et al, 2008; Muramatsu et al, 2008; Gamis et al, 2011). to TL-DS occur in of often due to failure associated with malignant pericardial and pleural failure and (Massey et al, 2006; Klusmann et al, 2008; Muramatsu et al, 2008; Gamis et al, 2011). death in neonates with TL-DS may also occur due to not to to disease or other congenital abnormalities (Klusmann et al, 2008; Gamis et al, 2011). of early death are in Table The factor associated with early death in the 3 large studies of TL-DS was (Massey et al, 2006; Klusmann et al, 2008; Gamis et al, 2011). In 2 of the studies a count was associated with early death (Klusmann et al, 2008; Gamis et al, in the study (Massey et al, the count of the with TL-DS early was to in the factor associated with early death in TL-DS is liver In the study by et levels were in with TL-DS early to survivors and all of the with TL-DS early had evidence of liver failure and in the Klusmann et study, all patients with TL-DS with hepatic fibrosis This study also found a of and in the children with TL-DS early to The Oncology Group (Gamis et al, used a of the presence or absence of they defined as life-threatening Table 4 and and the presence of absence of to patients into with no and low or survival for the was and but all in the low and were not to be to TL-DS. Thus, these data suggest that the presence of should identify all children at of TL-DS early death. TL-DS and ML-DS blast cells are sensitive to et al, et al, 2002). This that very low of can be used to blast in TL-DS. et reported that of patients identified to have life-threatening hepatic 3 patients were with low for and all further patient was on in but all not receive treatment et al, 2002). The group treatment with for for any patients with TL-DS with clinical due to of or liver or count of treatment with the treatment group included large of children care in and several with hepatic fibrosis survival in the 2 was very survival that treatment was with when analysis was to patients identified as using analysis or failure to the of death was in the treatment group for the group (Klusmann et al, Further evidence for the of low from a of the study from babies with a count a clear in survival when with – for not et al, 2015). The identified of patients as (see Table and patients were given as a at a of for days (Gamis et al, 2011). of patients 3 or 4 the of in other of the treatment group It is the that there is often to for a that may resolve and that this the of potentially treatment despite evidence of low with of is a relatively in Care In the of may have in rapidly counts in patients with not be to other In the study, of the group were with or 2 no further treatment, 2 further treatment, had and (Gamis et al, 2011). et recently reported their of in TL-DS babies at a of 2 days of the count from a of to a of 21 with clear clinical in all cases, although 2 subsequently and 2 more to in the first The group recently their data from the a prospective study at the of low to prevent with a clinical syndrome of TL-DS and any babies with disease detectable by or molecular disease were with a of at for one week et al, there was a survival in with clinical TL-DS to a there was no in the of ML-DS This is in keeping with the experience, which found no role from low with to of ML-DS (Gamis et al, 2011). the mortality in the evidence of of treatment in some the that cases will and the lack of evidence that treatment of is clear that, based on our current but not all patients should be with low Thus, the in TL-DS is and to The to were due to of or liver or count (Klusmann et al, Of we that is not only not associated with early but is no more in TL-DS is in neonates with DS and no GATA1 mutation (Roberts et al, we not as an to and liver are of hepatic fibrosis and death (Massey et al, 2006; et al, 2007; Klusmann et al, 2008) and are criteria to although the specific used by the – the and the – and low respectively – The to were the presence of (see Table with due to and failure not due to congenital all these of potentially a count has been associated with early death. the presence of was associated with a survival of and all but one of the TL-DS Thus, these to be of hepatic disease is often one of the key of when to but of are The of hepatic disease is that of with a often by leading to and However, the of hepatic disease is and may present at birth together with and There is often hepatic in the of an blood count et al, However, is relatively – Gamis et that was present in of TL-DS patients were not to treatment and in the absence of defined an group were at low of early death to TL-DS. the was to the et reported a of cases of of whom due to hepatic In at was to a of at the time of death. Muramatsu et reported a large retrospective and found that the presence of a was associated with early death on and analysis – and et reported their of with TL-DS a with to the history of liver Of note, all with the on developed on a of at a time when peripheral blood blast counts were had a early one after and the other had a diagnosis of of a in was sensitive nor specific for TL-DS hepatic et used their of hepatic disease in TL-DS to assess Of patients any babies had out of the a on and resulting in or fibrosis on liver However, is not clear this given that a and all be criteria to in assessment is not in the UK, and fibrosis has previously been associated with death in of children should be for evidence of disease and because of the of and In some cases, a single of is not to TL-DS blast counts to but liver disease in the first of life et al, and can be to an of In liver disease often a history to the peripheral blast of should be to where liver and should be that will often to resolve and is not necessarily of Study data are on – in the only a from disease or (Gamis et al, 2011). In cases of TL-DS where there are no is to without treatment as all cases will the found of children of the peripheral blast cells at a of days the other 2 developed and treatment (Gamis et al, 2011). However, some children develop ML-DS without of counts (Klusmann et al, no study has shown any after of (Massey et al, 2006; Klusmann et al, 2008; Gamis et al, 2011). Studies of clinical TL-DS have shown a of to ML-DS of 20–23% of children TL-DS developed ML-DS (Massey et al, 2006; Klusmann et al, 2008; Muramatsu et al, 2008; Gamis et al, 2011). The which used more sensitive for GATA1 mutation that the of in children with DS had a GATA1 mutation detected at birth of all neonates with was to (Roberts et al, 2013). There is no evidence that treatment with to ML-DS (Gamis et al, 2011). of clinical of the blood count and blood film will usually to a of children will have abnormalities or will the first 3 of life with In our experience, GATA1 mutation analysis can be to the diagnosis in these However, the of all cases of TL-DS for the of the GATA1 mutation and/or of assessment of the of any in TL-DS or TL-DS has not been studies have of blast cells based on a (Klusmann et al, 2008) or using time & Pine et al, There is no evidence at present to show that either of these is of In to of cases of TL-DS have been shown to have mutant GATA1 to at (Ahmed et al, 2004; Alford et al, Roberts et al, 2013; Yoshida et al, 2013) and ML-DS may develop from the the mutant (Yoshida et al, 2013). that ML-DS has a in the of life and is rare after the of 4 years et al, 2003; et al, of children with TL-DS can be by the of 4 years the is ML-DS often has an with a with usually with a low percentage of circulating blasts, for many blast cells the of the and of ML-DS in an with a history of TL-DS is a cases will and the is to all cases as ML-DS (Zipursky, 2003; et al, There are no published studies to the of for However, given that cases of ML-DS will present 2 years et al, et al, there is a for more to this Furthermore, the of to ML-DS in cases that a 3 is likely to identify ML-DS as cases of ML-DS will have by 2 we suggest that is to the of after this the is Any significant blood count particularly should GATA1 mutation analysis and early of a bone and bone are frequently due to fibrosis and are for the diagnosis of It is the of this guideline to recommend treatment for ML-DS but give survival of et al, 2016; et al, et al, In the UK, is that ML-DS be according to the ML-DS Despite in is uncommon for TL-DS to present of neonatal cases have been fetal TL-DS is detected on in the with or pericardial abnormalities fetal pleural and peripheral fetal blood is blood films show with blasts and abnormal liver et al, 2016). Of cases identified in the only were at there were of and in in the first cases were few cases all reported in and there is only evidence on which to on have included and intrauterine of cells and be for fetal but may or et al, et al, et al, 2016). There are no cases reported of intrauterine with and is important to that is reported et al, 2016). 2 for a of the approach to investigation and management of TL-DS. the and in this guidance is to be and accurate at the time of to the the British Society for Haematology nor the any for the of this have a of to the BSH Guidelines and Task which may be on of the members of the group has any of to in and at the of Medicine is by with Cancer and the Leukaemia of the group will the group any evidence that the strength of the in this or The will be and from the BSH current guidelines website are an will be published on the BSH Guidelines changes are due to changes in of evidence or significant evidence current a of the current guidance will be on the BSH Guidelines is for this guideline and on the of the BSH of the the and to the and of the