Myelodysplasia and myeloproliferative disorders in childhood: an update

Abstract

The myelodysplastic syndromes (MDS) and the myeloproliferative disorders (MPD) arising during childhood, even if combined, are generally thought to represent an uncommon occurrence of a heterogenous group of clonal stem cell disorders. Previous reported incidence rates, based on small series, have suggested that childhood MDS represents approximately 3% of paediatric haematological malignancies. Only two actual population-based studies have been reported for childhood MDS (62; 73). The recent study based in Denmark (62), found an annual incidence of 4.0 per million corresponding to 9% of all malignant haematological disorders in children. In contrast, 73) found a much lower incidence of childhood MDS in the range of 0.5 per million, but there were also significant differences between these studies for the reported incidences of acute myeloid leukaemia, acute lymphoid leukaemia and chronic myeloid leukaemia. A number of conditions have been identified as predisposing factors for childhood MDS or MPD (6; 52; 62; 64; 89; 98). These conditions include: Down's syndrome, Kostmann's syndrome, Noonan's syndrome, Fanconi anaemia, trisomy 8 mosaicism, neurofibromatosis (as discussed in further detail below), Schwachmann's syndrome, immunodeficiency and familial leukaemia. In addition, previous chemotherapy administered to children for other disorders can be a predisposing factor for the subsequent development of either childhood or adult MDS and MPDs. The incidence of therapy-related MDS is likely to increase in the coming decades as chemotherapy dose escalation is now being attempted for numerous malignant diseases. Therapy-related MDS will not be further addressed in this review but represents a serious problem with few effective therapies at present. In addition to the uncommon occurrence of these diseases, much of the difficulty in assigning incidence rates has arisen from a lack of consensus for a classification scheme. Since these diseases generally occur with a much higher frequency in adults, wherein classification schemes have been at least partially devised, for the most part attempts have been made to apply these adult classifications to the paediatric cases (Table I). Most MDS patients present with signs and symptoms of the cytopenias present in the peripheral blood. Hepatosplenomegaly is uncommon. Dysplastic features are common and the marrow cellularity is normal to increased. If one applies the French–American–British (FAB) group criteria (10, 11) devised for adult MDS to cases of childhood MDS, one notes that refractory anaemia with ringed sideroblasts (RARS) is only very rarely reported in childhood cases, whereas chronic myelomonocytic leukaemia (CMML) in childhood is more common, especially in children under the age of 5 (24; 60; 98; 82; 94; 84). Cases of refractory anaemia (RA), refractory anaemia with excess blasts (RAEB) and refractory anaemia with excess blasts in transformation (RAEB-t) have all been reported to present during childhood. The predominating diagnosis amongst the FAB classified diseases in childhood MDS is childhood CMML, now termed juvenile myelomonocytic leukaemia (JMML) as discussed below. This disease has also been referred to by other names in the past including: juvenile chronic myelogenous leukaemia (JCML) (51; 55), juvenile chronic granulocytic leukaemia (1; 7) and subacute and chronic myelomonocytic leukaemia (24). A similar disorder, also often occurring in early childhood, has been termed infantile monosomy 7 or monosomy 7 syndrome (110; 46; 85; 98). These disorders will be discussed in much greater detail below. A strong male predominance has been noted in several MDS subtypes including CMML. The male to female ratio has ranged in various reports from 1.7:1 in both the French and American studies (24; 84) to 4.8:1 in the British study (98). The factors underlying the male predominance and its potential biological significance are unknown. The adult diseases which 32) first termed as related myeloproliferative disorders are polycythaemia vera, myelofibrosis (agnogenic myeloid metaplasia), essential thrombocythaemia and chronic myelogenous leukaemia. However, there are numerous other disorders that have also been classified previously (118; 39) in the myeloproliferative disorder category (Table I), several of which have been reported to occur in childhood populations. Reports of MPD occurring in childhood have infrequently described polycythaemia vera, essential thrombocythaemia, agnogenic myeloid metaplasia (also termed primary myelofibrosis and idiopathic myelofibrosis) (96; 88; 104), primary familial and congenital polycythaemia (100; 37) and chronic neutrophilic leukaemia (65). Since the majority of these reports of paediatric cases have been anecdotal, traditionally the classification and diagnostic criteria for these disorders in childhood generally has followed the same criteria as developed for adults with these disorders. Unlike the rare paediatric occurrence of the MPDs discussed above, Philadelphia chromosome-positive (bcr/abl-positive) or adult-type, chronic myelogenous leukaemia (ACML) clearly occurs in children with a significant occurrence constituting 2–5% of all leukaemias. It is less common under the age of 2 as compared to other age groups, ranging up to 20 years of age, at which the incidence is < 1 per 100 000. The incidence of ACML increases after the age of 20. Besides the difference in age distribution, ACML in childhood is typically easily distinguishable from JMML and from adult-type CMML (which does occur in children) by several other characteristics as illustrated in Table II. Childhood ACML patients commonly present with symptoms attributable to the splenomegaly that is invariably present. These symptoms include: abdominal distension and discomfort, dysphagia and weight loss. Other common presenting symptoms are malaise, bone and joint pains, fever, night sweats and bleeding diathesis. Leucostasis signs and symptoms including strokes, papilloedema, cerebellar signs and priapism occur slightly more commonly in childhood ACML than in adults with ACML. A definitive diagnosis is made by the demonstration of the translocation between chromosomes 9 and 22, or by detection of the bcr/abl fusion gene. Patients with ACML, JMML or adult type CMML must all be distinguished from the occasional paediatric patient who will present with leucocytosis or a ‘leukaemoid reaction’ resulting from infectious causes, metastatic cancer or congenital heart disease. The latter cases are distinguished from the former cases by normal cytogenetics, lack of associated features as depicted in Table II, and resolution and/or lack of progression of signs and symptoms. The classification of CMML in childhood (now termed JMML) into either the MDS category or the MPD category has been particularly problematic. Likewise, the classification of CMML in adulthood has also fallen more recently under similar scrutiny. Attempts to re-categorize adult CMML into either the MDS or MPD heading, based on the total leucocyte count, have thus far not proved useful in regards to prognostic information (53; 97). Until more knowledge is elucidated regarding the pathogenetic mechanisms of these two disorders we may not be able to clearly classify adult CMML, or its paediatric counterpart JMML, into either the MDS category or the MPD category. For the present, and in order that common nomenclature is used by all to compare similar groups in future studies, the most appropriate classification of these two disorders may be to classify them (as noted by 9) as ‘bridging’ disorders and call them myeloproliferative/myelodysplastic disorders (Table I). Besides childhood ACML, of all of the remaining subtypes of childhood MPD and MDS, JMML and monosomy 7 together make up by far the largest porportion of patients. JMML has several distinguishing characteristics (51; 39; 4; 20; 27). Common clinical features include: marked hepatosplenomegaly, generalized lymphadenopathy, and skin changes including eczematous rashes, café-au-lait macules and dermal neurofibromas. Common laboratory findings include: moderate leucocytosis with monocytosis, anaemia and thrombocytopenia, elevated fetal haemoglobin, and hypergammaglobulinaemia. Pathologic findings include infiltration of various non-haemopoietic organs (skin, lungs, intestines) with leukaemic monocytic cells. Molecular findings include: lack of the bcr/abl fusion gene, and increased incidences of monosomy 7 as well as NF1 and RAS abnormalities. One of the previously most commonly accepted names for JMML was JCML. This term was coined to distinguish this myeloproliferative/myelodysplastic disorder from the adult type (bcr/abl-positive) CML. However, the disorder named JCML never really had sufficient similarities to ACML to warrant the name since a monocytic presence predominated and the disease course was generally much more aggressive than ACML. Similarities to the adult form of CMML are much broader than those to ACML (Table II), but distinct differences are also noted including: lack of incidence of neurofibromatosis in adults and rare elevations in fetal haemoglobin in adults, higher incidence of and more complex karyotypic abnormalities in adults, and differences in in vitro cell growth characteristics between the two disorders (41; 21). As a result, over the last several years discussions have taken place amongst the International JMML Working Group and the European Working Group of MDS in Childhood (EWOG-MDS) and an international consensus has been reached to rename this myeloproliferative/myelodysplastic disorder to JMML (4; 95). To the clinician, JMML can pose a diagnostic dilemma. The initial presentation and early clinical course of these patients can be quite heterogenous. Morphologically, in some cases the cells may show minimal dysplastic changes and there is no evidence of maturation arrest. The issue is further clouded by the fact that the clinical presentation can mimic that of several viral infections including Epstein-Barr virus, cytomegalovirus and human herpes virus-6 (68; 75; 83). Through the ongoing deliberations of the International JMML Working Group and the EWOG-MDS, a consensus has been reached on a set of clinical and laboratory features to be utilized as minimal diagnostic criteria (Table III, adapted from Niemeyer et al, 1998) until the molecular abnormalities defining JMML have been substantially characterized. Some young children with a monosomy 7 karyotype share many of the same clinical, laboratory and pathological features of JMML. Several groups have attempted to classify this group of young children with monosomy 7 into a separate disorder called infant monosomy 7 syndrome (110; 46; 98). This group, when first defined, appeared to initially differ from other JMML patients by a lower fetal haemoglobin level, a longer survival, but a higher rate of transformation to AML. However, many areas of overlap exist amongst these two disorders. Beyond this group of infantile patients, monosomy 7 as a cytogenetic finding is seen in children with JMML, AML and in children with all of the FAB subtypes of MDS (31; 84). Although some reports have advocated classifying these young children with monosomy 7 as a truly separate disorder (110; 19), more recent studies suggest there is no data to support the concept of monosomy 7 as a distinct syndrome, but rather it should be viewed as a ‘cytogenetic opportunist’ (84; 61). The International JMML Working Group and the EWOG-MDS have concluded that these disorders should be treated as the same disorder in treatment protocols, and subjected to subset analysis later. The clonal nature of JMML was inferred by several reports of small numbers of patients who had chromosomal abnormalities (66; 69, 70; 16; 2). Recent methodological improvements in clonality assessment, based on X-chromosome inactivation patterns, have made possible the tracing of the clonal origin of JMML cells to at least the level of the myeloid stem cell, and potentially even the pluripotent stem cell (18). Further, JMML cells appear to maintain clonality for relatively long periods in cell cultures (at least up to 42 d), in contrast to ACML cells (38). JMML mononuclear cells, obtained from either a peripheral blood or bone marrow source, have long been known for their consistent ability to spontaneously form granulocyte-macrophage colonies (CFU-GM) in vitro, with the monocytic forms predominating (1; 44; 55). Other MPD patient cells can occasionally demonstrate spontaneous colony growth but not as consistently as in JMML (111). This spontaneous CFU-GM colony growth is not due to autocrine or paracrine mechanisms of cytokine production (36), but rather is due to the JMML haemopoietic progenitor cells having a selective 10-fold hypersensitivity to granulocyte-macrophage colony-stimulating factor (GM-CSF) (35). Growth factor responsiveness of these cells to either interleukin-3 (IL-3) or to G-CSF is normal. This hypersensitivity of the JMML myeloid series to GM-CSF results in activated monocytes, which in turn produce other cytokines such as IL-1 (7) and tumour necrosis factor (TNF), which can then enhance the morbidity of JMML by suppressing normal haemopoiesis (50). This pathogenetic mechanism of cytokine hypersensitivity is similar to that for the erythropoietin (EPO) hypersensitivity observed in primary familial and congenital polycythaemia (100). However, unlike primary familial and congenital polycythaemia in which EPO receptors are mutated in some families with the disorder (33; 113), no GM-CSF receptor mutations have been reported to date in JMML (119; 49). Cytokine hypersensitivity has emerged as a common pathogenetic mechanism in at least five myeloproliferative disorders (39). If one includes monosomy 7 under the classification of JMML, 40–67% of patients will have normal cytogenetics, approximately 25–33% will have monosomy 7, and only 10–25% will have other chromosomal aberrations. Of this small latter portion, several will have other chromosome 7 abnormalities, leaving only a very small number of JMML patients who have more complex karyotypes (52; 98; 94; 86; 84). Neurofibromatosis is a common autosomal dominant disorder with a general incidence of 1 in 3500 individuals. JMML patients have a strikingly increased incidence of clinical signs of neurofibromatosis, being observed in about 15% of patients (5; 107; 94; 109). In addition, a recent report has demonstrated that an additional 15% of JMML patients harbour mutations in the neurofibromatosis gene, NF1, without manifesting clinical signs of the disease (108). In addition to NF1 mutations, various studies reported a rate of 15–30% of JMML patients demonstrating RAS gene mutations within their haemopoietic cells (74; 91; 108). These groups of patients with NF1 or RAS abnormalities appear to be mutually exclusive (74; 108). This matter becomes of considerable importance because the Ras family of proteins regulates one of the significant signalling pathways for GM-CSF within haemopoietic cells (106). Further, the NF1 gene encodes for neurofibromin, a protein with a domain which functions as a GTPase activating protein which serves to inactivate Ras from its active GTP-bound state (14; 12; 15). The NF1 gene abnormalities in JMML patients represent loss of function mutations suggesting that NF1 acts as a tumour suppressor gene in immature myeloid cells by negatively regulating Ras output (107; 109). Recently, two groups have homozygously deleted the neurofibromatosis gene in mice, Nf1, to simulate the condition in JMML patients who have loss of heterozygosity for NF1 (13; 77). The mice who harboured the homozygous deletion, Nf1−/−, suffered an embryonic lethal event at day 13–14 of gestation. However, when these investigators removed the day 13.5 fetal liver cells (which at this developmental stage contain the majority of haemopoietic cells) from these mouse embryos and subjected them to colony growth assays, they observed the same GM-CSF hypersensitive growth pattern as observed in JMML patient samples. Further, if these same cells were transplanted into irradiated recipient mice, the mice developed a myeloproliferative disorder reminiscent of JMML. These definitive animal studies clearly indicate that a deregulated Ras signalling pathway can result in cells demonstrating hypersensitivity to GM-CSF and delineate at least one potential pathogenetic mechanism for the development of JMML. Whether the reported somatic RAS gene mutations in JMML, the majority of which are either K-RAS or N-RAS activating point mutations, are pathogenetically causative in JMML, remains to be determined. However, the NF1 and RAS mutations remain mutually exclusive in all JMML patients thus far studied (74; 108). Taken together, deregulated Ras signalling appears to be molecularly evident in up to 50–60% of JMML patients. The genetic events occurring in the other 40–50% of JMML cases is an area of active investigation. Childhood cases of RA, RAEB, RAEB-t and even adult-type CMML have been noted in children. RARS appears to occur only very rarely. The clinical course of childhood MDS can be quite variable. An approximate 40% conversion rate to AML is similar to that observed in adult MDS (84). MDS must be distinguished from AML with a low blast count, as the latter appears to have a higher survival rate associated with therapy (25). As in adult cases, some childhood cases of RA have been managed with supportive care, or in some cases treatment with low-dose chemotherapy, types of differentiating agents or haemopoietic growth factors have been attempted. Low-dose chemotherapeutic agents which have been attempted include hydroxyurea, 6-mercaptopurine, busulphan, prednisone, cytarabine and 5′-azacytidine. Unfortunately, most of these attempts with ‘less-intensive’ therapy from all categories have resulted in only temporary responses in a subset of patients. Cases of RAEB, RAEB-t and CMML have been treated with intensive (AML-type) chemotherapy with or without allogeneic bone marrow transplant following remission induction (98; 63; 94; 29). Argument exists as to whether these childhood MDS cases respond as well or worse to this type of therapy as compared to de novo AML. Nevertheless, regimens of intensive chemotherapy which do not contain stem cell transplantation as part of the regimen have limited curative potential (63). It would appear that allogeneic stem cell transplantation, performed early in the course of the disease and from the most closely matched source available, confers the best hope for long-term survival (93). Some series note a disease-free survival for childhood MDS to be as high as 60% following transplantation (93; 81; 3), whereas others note a distinctly lower rate. As we further subtype classes of MDS based on cytogenetics and molecular markers, subgroups may emerge that require different intensities of therapy (8). Children with Down's syndrome can demonstrate several unique haematological manifestations including: (1) a transient myeloproliferative disorder in the newborn period, which normally shows spontaneous resolution, (2) relatively frequent occurrence of MDS, and (3) 20-fold higher incidence of AML with a particular preponderance to develop the megakaryoblastic subtype. Several recent publications all appear to agree on the fact that children with Down's syndrome and haematological disorders represent a distinct subgroup which are more responsive to conventional therapy than children without Down's syndrome who have the same haematological picture (102; 30; 76). Less than standard therapy in these Down's syndrome patients might even be warranted. The three phases of bcr/abl+ CML which are typically observed in most adults over the course of their illness are also typically present in most children with ACML. Most patients begin in chronic phase, lasting on average about 3 years, during which time disease control is achieved with minimal therapy. An accelerated phase may then precede the blast crisis phase, or patients may skip right into the blastic phase. The accelerated phase is characterized by leucocytosis refractory to previous therapeutic doses, increasing splenomegaly, anaemia and weight loss. The blast phase can be either of the myeloid or the lymphoid type. Traditionally, either busulphan or hydroxyurea chemotherapy had been used to control the leucocytosis in chronic-phase ACML. However, advances in the use of interferon-alpha therapy, either alone or combined with chemotherapy such as cytarabine, have largely replaced the use of these traditional agents in children with chronic-phase ACML, just as they have in adults (114; 56). Interferon-alpha therapy can lead to very adequate control of the leucocytosis in ACML. At present though, the only known potentially curative therapy is stem cell transplantation from either matched related or unrelated sources (115; 54). Children can have an 80% or greater chance of long-term survival with a matched, related or unrelated donor, provided they are transplanted early in their disease course (58). Every effort should be made to transplant children with ACML within a year from diagnosis. Duration of disease before transplantation was identified as a predictor of survival in the study by 58). Current recommendations, therefore, would be to begin hydroxyurea therapy when necessary to control the leucocytosis after a definitive diagnosis of ACML has been made. A search for a stem cell donor, either related or unrelated, should commence immediately. If all potential stem cell sources, related and unrelated, and from marrow or cord blood sources, fail to produce an acceptable donor then stem cell transplantation may be discarded as a therapeutic option. If this unlikely scenario does occur, a switch from hydroxyurea therapy to interferon-alpha therapy with or without cytarabine could be considered. Patients who have responses to interferon can have median survivals of 6–10 years, or possibly more if treated concomitantly with cytarabine. A minority of patients treated with interferon-alpha can achieve a major cytogenetic response as defined by marked reductions, or elimination, of the Philadelphia chromosome or the bcr/abl fusion product. However, since these types of responses can take 6–18 months of continuous interferon-alpha therapy before they occur, attempting interferon therapy prior to transplantation could mean missing the window of opportunity for best survival after transplantation. At present, interferon therapy for childhood ACML must be viewed as a second-line agent behind stem cell transplantation. States of erythrocytosis may arise due to primary polycythaemias, secondary polycythaemias or relative polycythaemias. The primary and secondary disorders are characterized by an elevated red cell mass, whereas the relative polycythaemias have a normal red cell mass but a reduced plasma volume. The radioactively labelled red cell mass study should therefore be one of the first tests performed in the work-up of polycythaemic patients. Secondary polycythaemic states arise from a multitude of causes including excessive endogenous erythropoietin levels from several sources, haemoglobins with abnormal oxygen affinity, and decreased erthrocyte diphosphoglycerate. The majority of the secondary polycythaemias can be ruled out by measuring erythropoietin levels, arterial blood oxygen saturation, and P50 levels or haemoglobin dissociation curves. This is one of the two primary polycythaemic states and is a clonal haemopoietic stem cell disorder whose pathogenesis has been linked with hypersensitivity of haemopoietic progenitor cells to insulin-like growth factor-I (IGF-I) in strictly serum-free culture conditions (28). Patients with polycythaemia vera can present with hepatosplenomegaly. However, the majority of symptoms, including dizziness, headaches, fatigue, pruritus and night sweats, arise due to the increased red blood cell mass. Since this is a haemopoietic stem cell disorder, aberrations in the white blood cell and platelet counts can also be encountered. Growth of erythropoietin-independent erythroid colonies in serum-containing cultures can be a useful diagnostic test (99). The criteria developed by the Polycythaemia Vera Study Group should be applied to children as well as adults in making a diagnosis of polycythaemia vera. As with the treatment of young adults with this disorder, most paediatricians are hesitant to use chemotherapy or radioactive phosphorus to control the cell counts because of the leukaemogenic risk with these agents, as patients with polycythaemia vera can expect a long survival. Phlebotomy is often used to control the haematocrit, but clinicians should be aware that the highest period of risk for thrombotic and/or haemorrhagic complications is directly following the phlebotomy. The use of isovolumic erythrocytopheresis is recommended if available (103). Hydroxyurea, interferon-alpha or -gamma and anagrelide are all agents used in adults to control cell counts and may be useful in paediatric cases. Hydroxyurea is more simple to administer and has fewer side-effects than interferon. The leukaemogenic risk of hydroxyurea appears to be low, at least thus far. Anagrelide is probably useful only to lower the platelet count and is the newest and least tested of the three agents. Primary familial and congenital polycythaemia (PFCP) is the other primary polycythaemic disorder. It is distinguished from polycythaemia vera by aberration of only the erythroid series in the former (101). PFCP is usually inherited in an autosomal dominant fashion, but autosomal recessive modes have been reported. Secondary polycythaemic states must be ruled out in making the diagnosis. PFCP erythroid progenitors display erythropoietin hypersensitivity but not erythropoietin-independent growth, thus making it possible to separate it from a polycythaemia vera disorder which rarely occurs in a familial pattern. Most families who have members with PFCP do not need medical intervention. In fact, some patients appear to gain an athletic advantage from their polycythaemic state (33). However, there have been other reported families wherein some members do suffer thrombotic complications from their disease. Most children with this disorder will have older affected family members from which a predictive disease course can be approximated. These disorders must remain in the category of diagnosis by exclusion, since they appear to be extremely rare and since there are a multitude of causes for paediatric thrombocytosis. Not only should the other myeloproliferative disorders including polycythaemia vera and ACML be ruled out, but more common causes including nutritional, inflammatory, infectious and neoplastic disorders should be likewise excluded. Essential thrombocythaemia typically presents with isolated elevated platelet counts. The platelet count must remain elevated over time and while ruling out other potential causes. A recent article suggests that many patients carrying a diagnosis of essential thrombocythaemia may in fact not have a clonal disorder (59). This serves to further strengthen the argument that one must be very deliberate when entertaining a possible diagnosis of essential thrombocythaemia, especially in children. Patients with myelofibrosis will usually have more signs and symptoms referrable to their disease, including other affected cell lineages, splenomegaly, teardrop red blood cell forms on peripheral smear, malaise, night sweats and weight loss. Other potential causes to be ruled out include metabolic and bone diseases, connective tissue disorders, granulomatous diseases and metastatic tumours. If either essential thrombocythaemia or myelofibrosis is in fact diagnosed, and if there is a need for therapeutic intervention, the most recommended therapies are those being attempted in adults with these disorders. Chemotherapy, interferon-alpha and anagrelide are all potential agents. It must be kept in mind that these disorders can have a quite variable natural history. Despite the common feature of the pathogenesis being linked to GM-CSF hypersensitivity, the clinical course of JMML is truly quite heterogenous. Virtually all (>95%) of JMML patients are diagnosed under the age of 6, with males predominating. The majority all present with hepatosplenomegaly, lymphadenopathy, bleeding, anaemia, fevers and recurrent infections. Only a small percentage (∼15%) of JMML patients convert or transform to an acute leukaemia-type blast crisis (84).Some patients will experience a period of transiently stable disease followed by progression whereas others will show immediate rapid progression (24; 23). Several investigators have also noted that a very small minority of patients enjoy long survival with relatively indolent disease. It appears that this small minority can be treated with observation or with low-dose chemotherapy as the situation dictates. What separates this small minority of patients from the rest in terms of pathogenesis is still unclear. Three of three long-term survivors have demonstrated GM-CSF hypersensitivity in their myeloid marrow progenitors (unpublished observations). For the remaining majority of JMML patients, the literature supports that they suffer a fairly rapid disease course followed by demise within a year (24; 57; 94). Morbidity and mortality often occur due to bleeding, infection or non-haemopoietic organ failure due to monocytic infiltration. One feature that has been confirmed by several investigators is that JMML patients can be grouped into two prognostic categories with the most significant determining features being based on platelet count, fetal haemoglobin level and age (24; 98; 94). The worst prognosis group is those patients over the age of 2, with a low platelet count, and a high fetal haemoglobin level. Numerous therapeutic modalities have been attempted in efforts to induce remissions or at least to slow down the aggressive progression of JMML. (1) Intensive AML-type chemotherapy has been attempted to induce complete remissions, similar to its use in AML or in MDS patients with RAEB or RAEB-t. In contrast to the AML and the other MDS patients, JMML patients rarely achieve a remission with intensive chemotherapy regimens, and if they do achieve a remission it is short-lived (26; 43; 47; 63; 94). Because of the associated morbidity and the relative ineffectiveness of this type of therapy, participants at the First International Workshop on MDS in Childhood (Titisee, Germany, 1997) resoundingly agreed that such intensive therapeutics should not be further pursued unless it was in the setting of a pre-transplant regimen. (2) Lower-dose chemotherapy has also been attempted using various agents such as low-dose cytarabine or mercaptopurine (78; 116). Responses have been noted, but effect on outcome is uncertain and no long-term remissions off-therapy have been noted. (3) Interferon-alpha has been noted to have some inhibitory effects on the spontaneous CFU-GM growth in JMML (45). However, when applied to patients in a protocol using doses of interferon effective in ACML, no responses were observed (90; 67). It is possible that other doses of interferon-alpha would show some effectiveness in JMML. (4) 13-cis retinoic acid (CRA) was also shown to inhibit spontaneous JMML CFU-GM colony growth (42). And, contrary to interferon-alpha, CRA produced an overall 40–50% response rate in a pilot trial (23). The pilot trial also demonstrated that the CRA appeared to modulate the GM-CSF hypersensitive response in JMML cells. This 40–50% overall response rate has now been confirmed in a larger phase II clinical trial which has recently been preliminarily reported (22). The CRA is extremely well tolerated at doses as high as 200 mg/m2, and numerous patients have now been maintained on daily CRA for 2–5 years without untoward effects. Unfortunately, just as with low-dose chemotherapy, sustained remissions off-therapy appear to be scarce. (5) Stem cell transplantation is the only therapy that has thus far produced unequivocal sustained remissions. Stem cell sources have ranged from matched related bone marrow, to mismatched unrelated bone marrow, to related or unrelated cord blood stem cells (105; 17; 117; 34; 112; 80, 79; 87). No matter what the source, it appears that a major limiting factor to the success of stem cell transplantation continues to be an inordinately high relapse rate, >50% in most reported studies. The overall survival at 5 years appears to be of the order of 20–30%, or possibly less. Given the rarity of JMML and monosomy 7, most of the trials reported above have been relatively small, making the response percentages somewhat suspect. In any event, more effective therapy is necessary for these myeloproliferative/myelodsyplastic disorders. Over the last few years, JMML basic and clinical research efforts have been unified on an international scale, an approach that is direly needed for rare disorders such as JMML where it is difficult to obtain sufficient numbers of patients at any one institution. The European Working Group on Myelodysplastic Syndrome in Childhood (EWOG-MDS), together with other participating European countries, has implemented a prospective clinical trial employing early intervention with upfront allogeneic stem cell transplantation for JMML and/or monosomy 7. Likewise, the paediatric clinical trial cooperative groups in North America have joined forces and will soon activate a protocol for all newly diagnosed JMML and/or monosomy 7 children in North America. Patients will receive multimodality therapy with CRA, pre-transplant chemotherapy, stem cell transplantation, and then further CRA therapy. In addition to these large prospective trials which both incorporate stem cell transplantation, given the knowledge of Ras signalling deregulation resulting in GM-CSF hypersensitvity, several novel mechanism-based agents are now being explored in vitro and in animal models for potential future phase I testing in JMML. These potential future therapeutics include: (1) GM-CSF antagonist analogues (72, 71), (2) novel retinoid derivatives (92), (3) GM-CSF/diphtheria fusion toxins (48), and (4) farnesyl-protein transferase inhibitors (40). Given the acceptance of basic uniform diagnostic criteria, it is hoped that analysis of trials such as those above will produce more accurate information as to incidence and response to therapy. Until more is known about pathogenesis and/or subset response to therapy, the International JMML Working Group and the EWOG-MDS have agreed to include JMML and monosomy 7 (infantile monosomy 7) under the same treatment programmes. Given these large, internationally agreed upon clinical trials, it is hoped that all newly diagnosed JMML and monosomy 7 patients, at least in Europe and North America, could be entered upon one of these two treatment trials and at the same time entered into the respective registries now being developed or in place. It is likely that only through unified, prospective epidemiologic, pathogenetic and therapeutic investigations that we will advance the treatment of these patients. The preliminary infrastructure of such an approach with uniform diagnostic criteria, international registries, and new protocols has taken shape over the last 5 years in several parts of the world. The impetus now lies with us to advance the science and the therapy for childhood MPD and MDS.