This article was downloaded by: [University of Otago] On: 17 July 2015, At: 14:50 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: 5 Howick Place, London, SW1P 1WG
Leukemia & Lymphoma Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/ilal20
Haematological Differences Between Chronic Granulocytic Leukaemia, Atypical Chronic Myeloid Leukaemia, and Chronic Myelomonocytic Leukaemia D. A. G. Galton
a
a
Honorary Consultant Physician, Hammersmith Hospital; Leukaemia Research Fund Professor Emeritus of Haematological Oncology, Royal Postgraduate Medical School, University of London. Formerly Honorary Director, Medical Research Council's Leukaemia Unit, London, UK Published online: 01 Jun 2015.
To cite this article: D. A. G. Galton (1992) Haematological Differences Between Chronic Granulocytic Leukaemia, Atypical Chronic Myeloid Leukaemia, and Chronic Myelomonocytic Leukaemia, Leukemia & Lymphoma, 7:5-6, 343-350 To link to this article: http://dx.doi.org/10.3109/10428199209049789
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Leukemia and Lymphoma, Vol. 7, pp. 343-350
0 1992 Harwood Academic Publishers GmbH
Reprints available directly from the publisher Photocopying permitted by license only
Printed in the United Kingdom
Haematological Differences Between Chronic Granulocytic Leukaemia, Atypical Chronic Myeloid Leukaemia, and Chronic Myelomonocytic Leukaemia Downloaded by [University of Otago] at 14:50 17 July 2015
D. A. G. GALTON Honorary Consultant Physician, Hammersmith Hospital; Leukaemia Research Fund Professor Emeritus of Haematological Oncology, Royal Postgraduate Medical School, University of London. Formerly Honorary Director, Medical Research Council’s Leukaemia Unit, London, U.K. (Received 25 January 1992)
Chronic myeloid leukaemia (CML) is a generic term that include five apparently distinct entities. The best known form, the classical Ph-positive subtype, accounts for about 90% of all cases of CML. The morphology of its presentation blood film is highly characteristic but is also seen in about half of the remaining 10% of cases, which are Ph-negative. This classical morphological subtype, whether Ph-positive or Ph-negative 1 describe as ‘chronic granulocytic leukaemia’ to refer to the exuberant granulocytic proliferation which is its hallmark. This term is often used indiscriminately and interchangeably with ‘chronic myeloid leukaemia’ and similar terms, just as ‘chronic lymphocytic leukaemia’ was, until recently, used to cover the chronic lymphoid leukaemias in general, but is now used in a specific sensez3. Chronic granulocytic leukaemia (CGL), whether Ph-positive or Ph-negative, is almost always BCR-rearranged and associated with the production of a unique 210-kd protein with enhanced tyrosine kinase activity. Most of the remaining cases of Ph-negative C M L are examples of either chronic myelomonocytic leukaemia (CMML), a subtype almost as homogeneous as CGL, and characterized in its presentation blood film by the presence of monocytes and neutrophils but few immature granulocytes, or atypical CML (aCML), distinct from and less homogeneous than either C G L or CMML, in which some cases also share features with C G L while others share some with CMML. C M M L and aCML d o not show BCR rearrangement and are not associated with the production of p2lOkd. CGL, CMML, and aCML, though characterized on morphological features differ in their clinical features and behaviour, response to treatment and survival. It is therefore clinically important that they are diagnosed accurately. Guidelines for doing so would also permit precise descriptions of each case submitted for molecular studies, thus assisting the working out of the events resulting in each phenotype, and enriching our understanding of the molecular basis of stem-cell differentiation. KEY WORDS:
CGL
atypical CML
CMML.
Ever since the discovery of a ‘minute chromosome’ in chronic granulocytic leukaemia by Nowell and Hungerford in 19601, thought by Baikie et d Z to be a possibly specific chromosomal abnormality for the
disease and named the Philadelphia chromosome (Ph) by them, it has been recognized that in a small proportion of cases of the same or a similar condition the minute chromosome is not present. The relationship between the Ph-positive and Ph-negative has been greatly since the Ph was shown Address for correspondence: Professor D. A. G. Galton, The Old Anchorage, Wiveton, Holt, Norfolk NR25 7TH, England, U.K. to result from a reciprocal translocation rather than 343
D. A. G. GALTON
Downloaded by [University of Otago] at 14:50 17 July 2015
344
a deletion, and since the molecular events underlying the translocation were identified and shown to occur in some cases in which the microscopically visible translocation was not present. In spite of this, there is still confusion about the nosological identity of the Ph-negative cases, in particular those which lack the characteristic molecular arrangements. This is partly because in many publications in which cytogenetics and molecular investigations have been described in detail, it is not possible for the haematologist to identify the kind of disease that was studied because the haematological descriptions are incomplete. The opportunity is therefore lost for understanding the events whereby a disturbed genome is expressed in a particular morphological phenotype. I have had the rare opportunity of examining, in a relatively short period of time, the presentation films of peripheral blood and bone marrow from 97 patients with Ph-negative chronic myeloid leukaemia (CML) (Table 1). 35 patients had been entered into the Medical Research Council's chronic myeloid leukaemia trial into which both Ph-positive and Ph-negative patients were eligible for entry. These slides were paired with slides from Ph-positive patients selected at random from the 512 Ph-positive cases in the trial, and they were examined independently by Dr Patricia Shepherd and me, neither knowing the Ph status. Slides from 41 patients attending Hammersmith Hospital or Burton District Hospital were examined by Drs Isabel Krisnik, P. C. Srivastava, and me: two of these were subsequently shown to be Ph-positive.
Films from 20 patients (including 5 who were Ph-positive) were submitted to me for morphological assessment by Drs Leanne Wiedemann and Dr Li Chan who were determining the BCR status and tyrosine kinase (PTK) activity. Finally films from 7 patients whose BCR status was to be studied by Dr T. S. Ganesan were examined for Professor J. M. Goldman: one of these was later found to have a complex chromosomal rearrangement with a masked t(9;22). The following observations were made on each blood film: first, a differential count on at least 300 cells and, secondly, qualitative assessments of the morphology of the erythrocytes, platelets and cells of the megakaryocytic, granulocytic, and monocytic lineages, with special reference to the occurrence and severity of dysplastic features. The Ph-negative chronic myeloid leukaemias account for about 9% of all cases of CML, and about half of them are indistinguishable from the common form of classical Ph-positive CML. Thus, in a random series of 100 cases of CML only 4 will be examples of the unusual variants. The opportunity to study so many examples of such extremely rare conditions in a relatively short period of time helped us to identify recurring patterns of differential counts and characteristic morphological features and so to characterize three main subtypes of CML. We can add nothing to the descriptions of the extremely rare juvenile CML and chronic neutrophilic leukaemia. The details of our analysis have been published e l ~ e w h e r e ~and - ~ here I discuss our conclusions, first, in relation to practical
Table 1 Morphological diagnosis in 122 cases of chronic myeloid leukaemia Source of slides
Morphological category
MRC*
CGL Ph' CGL PhaCML CMML CNL jCM L trs CGL Ph+ Total
19 18 10 7
H H und BDH**
CWt 5 5
14
6
23 2
1 1
JMGtt
Totul
I
25 26 33 31 3 2 2 122
3 3
2 2 44
41
20
7
CGL: chronic granulocytic leukaemia. aCML: atypical chronic myeloid leukaemia. CMML: chronic myelomonocytic leukaemia. CNL: chronic neutrophilic leukaemia. jCML: juvenile chronic myeloid leukaemia. * MRC: Medical Research Council; observers D. A. G . G. and Dr Patricia Shepherd. ** HH and BDH: Hammersmith Hospital and Burton District Hospital; observers: D. A. G . G., Dr Isabel Krsnik and Dr P. C. Srivastava. t CW: Drs Chan and Wiedemann; observer: D. A. G. G. tt JMG: Professor John Goldman; observer: D. A. G . G.
Downloaded by [University of Otago] at 14:50 17 July 2015
CGL. ATYPICAL CML AND CMML
diagnostic haematology, and secondly in the context of the sequence of events whereby a molecular alteration in the genome is expressed in a particular clinical-morphological phenotype. Before outlining the cytomorphological profiles of the three main subtypes of CML, it should be noted that we found the bone marrow aspirate to have limited if any diagnostic value. In all subtypes the marrow is characteristically hypercellular, usually devoid of fat, and shows marked granulocytic hyperplasia with full maturation. The common type, namely chronic granulocytic leukaemia (CGL), which is BCR + , and usually Ph + , typically has the highest granu1ocyte:nucleated erythroid-cell ratio, and increased numbers of megakaryocytes, a high proportion of which are small hypolobulated forms: chronic myelomonocytic leukaemia (CMML) has the lowest granu1ocyte:nucleated erythroid-cell ratio, while atypical chronic myeloid leukaemia (aCML) often has reduced though usually dysplastic megakaryocytes, but the overlap in these features is too great for them to have diagnostic value. The bone marrow aspirate has been overrated as a diagnostic aid in CML, while the presentation blood film has been undervalued. Martiat et d.’, however, have recently found the percentage of bone marrow erythroblasts to be a significant discriminant between CGL on the one hand and aCML and CMML on the other.
CHRONIC GRANULOCYTIC LEUKAEMIA (CGL) All 25 cases of CGL in chronic phase, Ph-positive, were correctly identified. The haematological features in presentation blood films, and those found by Shepherd in 100 further cases in the MRC series, confirmed the earlier findings of Spiers, Bain and Turner (1977)*.They may be summarized as follows: 1) The mean platelet count is above the upper limit of the normal range. 2) The differential leucocyte count shows prominent peaks in the neutrophils and in the myelocytes plus metamyelocytes, and the mean percentage of these cells increases as the total leucocyte count increases. 3) There is absolute basophilia (upper limit of normal 0.14 x 109/1) 4) The monocyte count is less than 3% and decreases further with increasing leucocyte counts.
345
5) Dysplastic features are minimal or absent in all cell lineages and are found in less than 10% of the cases. This characteristic combination of features is highly distinctive and in most cases CGL can be recognized at once in a well-stained presentation film. However, in early cases, when the total leucocyte count is below 50 x 109/1 the very low monocyte:granulocyte ratio may be less marked and if, by chance, the basophil count is low, confusion with aCML may arise. In 26 out of the 97 cases of Ph-negative CML, a morphological diagnosis of classical CGL was made. These 26 cases could not be distinguished on morphological grounds from the 25 Ph-positive cases. Furthermore, 14 of the 16 whose BCR status was investigated proved to be positive. The conclusion, therefore, is that CGL is a discrete entity with a characteristic morphological profile as seen in the presentation blood film. The general experience is that the vast majority of cases are BCR-positive, including the few that lack the t(9;22). Among the 71 remaining cases of Ph-negative CML, 31 stood out as a homogeneous group with a morphological profile distinct from that of CGL. This was chronic myelomonocytic leukaemia (CMML), now to be discussed.
CHRONIC MYELOMONOCYTIC LEUKAEMIA (CMML) Although so much has been written about CMML, there is still no generally accepted definition, so the range of features included in different definitions is wide, while one possibly crucial diagnostic feature, namely the percentage of immature granulocytes in the blood film at presentation, is rarely referred to. All agree that CMML is a chronic leukaemia predominantly affecting the elderly, that monocytosis is a constant feature, often associated with neutrophilia, thrombocytopenia, or anaemia in various combinations, that the bone marrow, though hyperplastic, frequently shows fewer monocytes than might be expected from the peripheral blood picture, and that about a third of the patients die from acute myeloid leukaemia. Non-specific esterase-stained bone-marrow films, however, may show a higher proportion of esterase-positive cells than would be expected from Romanowsky-stained films. CMML may be discovered by chance, or may present with symptoms caused by anaemia or thrombocytopenia.
Downloaded by [University of Otago] at 14:50 17 July 2015
346
D. A. G. GALTON
A palpable spleen, the commonest physical sign, is present in about a third of patients. Among patients with this range of features, the leucocyte count ranges from leucopenic to high levels of the order found in CML. A second prominent but variable feature is the presence in various combinations and in different degrees, of dysplasia in the megakaryocytic, erythroid or granulocytic series of cells. Indeed, there are so many cases in which, apart from monocytosis, the clinico-pathological picture is that of a myelodysplastic syndrome in any of its named variants, that the FAB Co-operative Group” felt that any case of myelodysplastic syndrome could be diagnosed as CMML when the monocyte count was higher than 1 x 109/l. However, the myelodysplastic syndromes are accepted primarily as conditions in which bone marrow hyperplasia is associated with one or more cytopenias, and the question arises as to whether the prominent neutrophilia often found in CMML is compatible with a myelodysplastic state. To some extent the dilemma may be resolved by considering the biological significance of cell death in attempting to distinguish myelodysplasia from myeloproliferation. Both states involve granulocytic hyperplasia, but in dysplasia early death of neutrophils renders the hyperplasia ineffective with resultant neutropenia, whereas in myeloproliferation the neutrophils survive, even though they may be abnormal, and neutrophilia ensues. If early cell death is considered a fundamental property of dysplasia, CMML with neutrophilia cannot be accepted as a myelodysplastic syndrome, but if early cell death is regarded as a secondary phenomenon, there would be no reason for excluding CMML from the MDS because of the neutrophilia and no inconsistency in accepting CMML with neutrophilia as both myelodysplastic and myeloproliferative4. However, the issue does not depend only on the interpretation of neutrophilia. The extent of dyserythropoiesis and of dysgranulopoiesis and of dysmegakaryocytopoiesis varies greatly in CMML. It is sometimes claimed that trilineage myelodysplasia is more frequent in leucopenic cases, and that many patients with high leucocyte counts have minimal or no dysplastic features in any cell lineage, or dysplastic features are confined to the granulocytic lineage. The implication here is that the patients with leucopenia and those with leucocytosis belong to two separate diagnostic groups and possibly to biologically distinct entities. The latter might be accepted as true CMML, the former simply as MDS with monocytosis. The issue is not merely academic because myeloprolifera-
tive conditions might turn out to respond to different types of treatment from that effective in myelodysplastic states. Even now CMML with leucocytosis is often treated with drugs, such as hydroxyurea or mercaptopurine, that would not be considered for MDS, while cytarabine at low dosage, or 13-cis-retinoic acid are not thought appropriate for myeloproliferative disease. The practical importance of resolving the dilemma is apparent in considering for example, the assessment of a new treatment. In designing a randomized trial of a new treatment for CMML, the eligibility criteria would have to be precisely defined. At present there are no guidelines for separating ‘CMML’ from ‘MDS with monocytosis’: cut-off points would necessarily be arbitrary and could lack biological justification, for it is not yet known whether essential differences exist between the ‘high’ and ‘low’ count cases. High-count CMML clearly qualifies as a ‘chronic myeloid leukaemia’ because it shows the cardinal features of bone-marrow hyperplasia with full granulocytic maturation, leucocytosis with neutrophilia, and a considerable risk of transformation to a blast-cell leukaemia. It has been considered to account for many of the cases reported as Ph-negative CML”.’’. Seven of our cases (with leucocyte counts up to 256 x 109/1) were admitted to the MRC CML trial. Our remaining 24 cases included those with leucocyte counts from 3.9 x 109/1 to 44 x 109/l. Our enquiry was limited in scope. It was designed to examine the morphological relationship between Ph-positive and Ph-negative CMLs. In the event, CMML has emerged as a remarkably homogeneous entity apart from the wide range of leucocyte counts. O n morphological grounds alone clear differences between the low-count and highcount cases have not emerged, and other methods will be required to reveal differences. The major morphological features in our series of 31 cases of CMML were as follows:
1) monocytosis with a low neutrophi1:monocyte ratio. Up to 10% of the monocytes may have azurophilic granules. 2) immature granulocytes in the peripheral blood at presentation do not exceed 15% and account for less than 5% of total leucocytes in a majority of cases. 3) the neutrophils may be either normal or d ysplastic. 4) the absolute basophil count is almost always within the normal range. These features, collectively, contrast markedly with
CGL, ATYPICAL CML AND CMML
Downloaded by [University of Otago] at 14:50 17 July 2015
those of CGL. The most surprising finding was the consistently low percentage of immature granulocytes: the characteristic appearance of the CMML blood film was a monotonous mixture of mature neutrophils and monocytes with some lymphocytes, and with only an occasional myelocyte or metamyelocyte. 33 of the remaining 40 cases of Ph-negative CML shared features that excluded them from both the CGL and CMML categories, though some cases also shared features with either CGL or CMML. This group was thus less homogeneous than either CGL or CMML, but nevertheless its constellation of features warranted its designation as a distinct entity which, for some years, we have provisionally named ‘atypical chronic myeloid l e ~ k a e m i a ’ ~’. - ~ . ~
ATYPICAL CHRONIC MYELOID LEUKAEMIA The major morphological features in our series of 33 cases of aCML were as follows: 1) basophils are absent or few in the majority of presentation blood films. 2) immature granulocytes and mature neutrophils are markedly dysplastic. In the histogram of the differential counts, there are neutrophil and myelocyte/metamyelocyte peaks as in CGL, but in a minority of cases the immature granulocytes (including promyelocytes) are below 15%. 3) the monocyte counts range from less than 1 x 109/1 to 50 x 109/1 but exceed 4% in two thirds of the cases. When the monocyte counts are 4% or higher, confusion with CMML is possible, and when they are less than 4%, confusion with CGL is possible. The most distinctive profile for the presentation blood film is the combination of low basophil and high monocyte percentages with high percentages of immature granulocytes: the immature granulocytes and mature neutrophils are markedly dysplastic. Equally distinctive but less common is the same combination except for the monocyte percentage which is low. Less common still are cases in which the monocyte percentage is below 4 and the immature granulocyte percentage below 15. When granulocytic dysplasia was prominent we diagnosed aCML. However, two of our patients proved to be Ph-positive and the prominent granulocytic dysplasia was associated with early transformation. In the absence of basophilia or
341
of blast cells it may not be possible to distinguish aCML from transformed CGL. An inherent difficulty in the recognition of aCML arises from the fact that so many of its features, like those of CGL and CMML, are continuous variables which overlap at the extremes of their ranges4. In a particular case of CGL, for example, the differential count may suggest aCML rather than with CGL; substandard staining may suggest dysgranulopoiesis thus supporting, erroneously, a diagnosis of aCML. This may be the most likely explanation for the occasional reports of BCR positivity in cases of Ph-negative aCML3. If this is correct a small number of wrong diagnoses will be unavoidable. But it is also possible that better criteria for the diagnosis of aCML may reduce the number of errors. This matter is being examined by the FAB Co-operative Group. An obvious criticism of the concept of aCML as an entity is that it shows a range of variability in respect of several features, particularly the percentages of monocytes and of immature granulocytes, which at one end of the range merge with values seen in CGL and at the other with those of CMML. This could imply that the three entities are artificial constructs and that in reality there is a continuum ranging between the extreme expressions of a single biological entity. Martiat et al.’ have compared 35 cases of Ph-, BCR- CML (aCML) with 55 of Ph-, BCR’ CGL, and 100 of CMML. Their conclusions essentially confirm our findings, but they speculate that aCML and CMML may be ‘two aspects of the same disorder’, distinct from CGL. Recently a group of seven Ph-negative patients described as ‘chronic neutrophilic leukaemia with dysplastic features’ has There . was been reported by Symeonides et ~ 1 . ’ ~ persistent neutrophilia, no monocytosis, or immature granulocytes, and there were marked dysplastic features in the granulocytic cells and megakaryocytes. The haematological profile in these cases may represent a fully mature variant of the low-monocyte group of aCML, clearly far removed from CGL, but close to chronic neutrophilic leukaemia. Evidence from other sources supports the sharp separation of CGL from aCML and CMML.
Clinical and haematological evolution of CGL, CMML, and aCML Neither aCML nor CMML shows the regular pattern of evolution of CGL in which there is a close relationship between the size of the spleen, the height of the leucocyte count, and the haemoglobin
Downloaded by [University of Otago] at 14:50 17 July 2015
348
D. A. G . GALTON
scales but patients with very high neutrophil counts other than those with CGL were not investigated. However, if the ratio was assumed to have been the same at all neutrophil counts, those for CGL proved to be abnormal, with the monocyte counts much lower than would have been expected in relation to the neutrophil counts, suggesting that the control mechanism that normally gears the production of neutrophils to the level of circulating monocytes is lost or much weakened in CGL. In our studies4 the neutrophi1:monocyte ratio was the same in Phpositive and Ph-negative CGL in contrast to that for CMML which was abnormal in the opposite direction, with the monocyte counts much higher than would have been expected in relation to the neutrophil counts. In aCML the neutrophi1:monocyte ratio was also abnormal. When the monocyte counts were 4% or higher, the neutrophi1:monocyte ratio, as in CMML, was low, though the neutrophil count was somewhat higher in relation to the monocyte count. However, when the monocyte counts were less than 4%, the ratio was abnormal in the reverse sense, as in CGL, with neutrophil counts abnormally high in relation to the monocyte counts5. This observation suggests that the difference between the high-monocyte count and Survival in aCML low-monocyte count aCML is determined by a switch The number of cases of aCML in this series is too mechanism controlling granulocyte/monocyte differsmall to make valid comparison of survival in this entiation. CGL, whether Ph-positive or Ph-negative, is disease with that for CGL, but ih the MRC CML trial4 the survival was markedly wqrse and may have strongly associated with low monocyte and high been worse than that of CMML. It has, of course, basophil counts, while in CMML the reverse is found. been known for many years that the survival in However, in aCML there was no clear Dattern in the Ph-negative CML is inferior to that in Ph-positive cases. CML'6*'7;because about half of the cases would have been examples of Ph-negative CGL, in which the Molecular biology survival is the same as in Ph-positive CGL, the survival of the aCML cases among them would have CGL, whether Ph-positive or Ph-negative, is strongly associated with BCR positivity, and with the been even shorter. production of the hybrid 210-kd protein with enhanced tyrosine kinase activity6. In contrast, BCR Other features rearrangement is not found in CMML, or in the great The relationships between the presentation neutrophil majority of cases of aCML; the few exceptions are and monocyte counts, and between the monocyte likely to be diagnostic mistakes, but more studies are counts in CGL, aCML, and CMML suggest that the needed. The 210-kd protein has not been found in regulation of monocyte/granulocyte differentiation CMML or aCML. and maturation, and the production of basophils differ in the three conditions. Bain and Wickramasinghe" Myelodysplasia in aCML investigated the neutrophil: monocyte ratios at various levels of neutrophil counts in normal subjects Morphological evidence of dysplasia, though charincluding pregnant women, and in patients with acteristic of the myelodysplastic syndromes, is often neutrophilia. The ratio was linear on logarithmic absent in refractory anaemia (RA) and RA with ring
concentration". For example, among the 169 patients admitted to the Medical Research Council's randomized trial of splenectomy in Ph-positive CGL' 5 , none of those whose presentation leucocyte counts were less than 50 x lo9/], and only 4 of the 20 whose counts were between 50 and 100 x 109/1, had a spleen 10cm or more below the costal margin, while the haemoglobin concentration was below 10 g/dl in only one of the 30 patients whose leucocyte counts were below 100 x lo9/]. In aCML the relationship is far less consistent so that aCML may be suspected even before the blood film has been examined, and associations of leucocyte counts, spleen size, and haemoglobin concentration unlikely in CGL are encountered. The impression is strengthened if the platelet count is below 150 x 109/1, as was the case in 12 of the 14 aCML patients in the Hammersmith/ Burton series, in contrast to just under 10% in the MRC trial patient^'^, 6 of the eight aCML patients whose leucocyte counts were below 56 x lo9/] had a haemoglobin concentration between 3.5 and 9.8 g/dl. The neutrophil alkaline phosphatase score, low in 95% of cases of CGL, was high in one third of the cases of aCML.
CGL, ATYPICAL CML AND CMML
sideroblasts (RARS). Some authors’’ exclude such cases from the MDS, while others” would accept as early MDS cases in which persistent unexplained macrocytosis is the only abnormality, because subsequent follow-up of these patients shows that they evolve to conventionally characterized MDS. The current debate on the grounds for excluding high-count CMML from the MDS derives from the recognition of two separate components of the disease, namely the myelodysplastic and the myeloproliferative elements, as already discussed. aCML has at least an equal claim to that of CML to be included as a myelod ysplastic/myeloproliferative disorder2 All granulocytic series cells show dysplastic features in varying degree: deficient or absent primary granulation in promyelocytes, deficient or absent secondary granulation in myelocytes, metamyelocytes, and neutrophils, nuclear abnormalities including PelgerHiiet forms, bizarre hypersegmentation, and abnormal chromatin clumping, are all common in aCML. Dysmegakaryocytopoiesis is the rule in aCML, and dyserythropoiesis is often seen. No case with a leucocyte count below 20 x 109/1 appeared in our series. Dysplasia is a major feature that distinguishes aCML from CGL in which dyserythropoiesis and dysgranulopoiesis are minimal or absent in the chronic phase, although the long recognized small megakaryocytes22, often resembling the micromegakaryocytes of MDS are characteristic. The significant point is that however the prominent myelodysplastic features of aCML and CMML are interpreted in relation to the myelodysplastic syndromes, they add to the list of features that separate aCML and CMML from CGL.
Downloaded by [University of Otago] at 14:50 17 July 2015
’.
349
confirmed, it is to be hoped that detailed recommendations for the definition of the subtypes will be forthcoming. These would be important for two reasons. First, by improving diagnostic accuracy, the comparability of reports from different sources would be much greater than it is now: secondly, the correlation between work at the molecular level with that at the morphological level would be strengthened. This would benefit the hard task of unravelling the sequential processes that link the genome with its ultimate expression in the observed phenotype. Understanding of the molecular biology of CGL and the definition of its phenotype are at a more advanced stage than is the case with any other haematological neoplastic disease, and the existence of three somewhat similar but distinct phenotypes that appear to have a different molecular genetic basis offers an unrivalled opportunity for seeking the explanation of the molecular events responsible for the divergent phenotypes.
SUMMARY
Chronic myeloid leukaemia (CML) is a generic term that include five apparently distinct entities. The best known form, the classical Ph-positive subtype, accounts for about 90% of all cases of CML. The morphology of its presentation blood film is highly characteristic but is also seen in about half of the remaining 10% of cases, which are Ph-negative. This classical morphological subtype, whether Ph-positive or Ph-negative I describe as ‘chronic granulocytic leukaemia’ to refer to the exuberant granulocytic proliferation which is its hallmark. This term is often used indiscriminately and interchangeably with ‘chronic myeloid leukaemia’ and similar terms, CONCLUSIONS just as ‘chronic lymphocytic leukaemia’ was, until recently, used to cover the chronic lymphoid Larger numbers of cases than those reviewed here leukaemias in general, but is now used in a specific must be studied to see whether the grouping of CML sense23. Chronic granulocytic leukaemia (CGL), into three main subtypes, namely CGL, aCML and whether Ph-positive or Ph-negative, is almost always CMML, with two excessively rare subtypes, juvenile BCR-rearranged and associated with the production CML and chronic neutrophilic leukaemia, holds of a unique 210-kd protein with enhanced PTK good, and whether the features characteristic of the activity. less common types can be precisely defined. In our Most of the remaining cases of Ph-negative CML study, aCML and CMML were characterized on are examples of either chronic myelomonocytic morphological features alone, but the characterization leukaemia (CMML), a subtype almost as homogenis supported by evidence from clinical, haematologi- eous as CGL, and characterized in its presentation cal, and negatively, cytogenetic and molecular blood film by the presence of monocytes and biological sources, as described above. neutrophils but few immature granulocytes, or If the classification of CML suggested here is atypical CML (aCML), distinct from and less
D. A. G. GALTON
350
homogeneous than either CGL or CMML, in which some cases also share features with CGL while others share some with CMML. CMML and aCML do not show BCR rearrangement and are not associated with the production of p2lOkd. CGL, CMML, and aCML, though characterized on morphological features differ in their clinical features and behaviour, response to treatment and survival. It is therefore clinically important that they are diagnosed accurately. Guidelines for doing so would also permit precise descriptions of each case submitted for molecular studies, thus assisting the working out of the events resulting in each phenotype, and enriching our understanding of the molecular basis of stem-cell differentiation.
8. Spiers, A. S. D., Bain, B. J. and Turner, J. E. (1977) The 9. 10.
11.
12. 13.
Downloaded by [University of Otago] at 14:50 17 July 2015
14.
Acknowledgements It is a pleasure to thank Dr Barbara Bain for her helpful comments and critical reading of the text.
15.
REFERENCES
peripheral blood film in chronic granulocytic leukaemia. Scandinavian Journal of Haematology, 18, 25-38. Bain. B. J. and England, J. M. (1975) Normal haematological values: sex difference in the neutrophil count. British Medical Journal, 11, 473475. Bennett, J. M., Catovsky, D., Daniel, M. T., Flandrin, G., Galton, D. A. G., Gralnick, H. R. and Sultan, C. (1982) Proposals for the classification of the myelodysplastic syndromes. British Journal of Haematology, 51, 189-199. Pugh, W. C., Pearson, M., Vardiman, J. W. and Rowley, J. D. (1985) Philadelphia-chromosome negative chronic myeloid leukaemia: a morphological reassessment. British Journal of Haematology, 60,456-467. Travis, L. B., Pierre, R. V. and De Wald, G. W. (1986) Ph’ negative chronic granulocytic leukemia: a non-entity. American Journal of Clinical Pathology, 85, 186193. Galton, D. A. G. (1982) The Chronic Leukaemias. In Blood and its Disorders. Eds Hardisty, R. M. and Weatherall, D. J. pp. 877-897. Blackwell, Oxford. Symeonides, A., Kovrakli, A,, Katevas, P., Pagonis, S., Zikos, P., Aktipi, A,, Tiniakou, M., Matsouka, P. and Zoumbas, N. (1991) Chronic neutrophilic leukaemia with dysplastic features (CNL-D). A report of seven patients. Abtstract AS. Second International Conference on Myelodysplastic Syndromes. Leukaemia Research, 15, supplement. MRC (1983) Randomized trial of splenectomy in Ph positive chronic granulocytic leukaemia including an analysis of prognostic features. British Journal of Haematoloqy. 54,
1. Nowell, P. C. and Hungerford, D. A. (1960) A minute
415430. 16. Canellos, G., Eagam, R. T., Whang, J. and Carbone, P. P. (1968)
chromosome in human chronic granulocytic leukemia. Abstract. Science, 132, 1497. Baikie, A. G., Court-Brown, W. M., Buckton, K. E., Harnden, D. G., Jacobs, P. A. and Tough, I. M. (1960) A possible specific chromosome abnormality in human chronic myeloid leukaemia. Nature, 188, 1165-1166. Ganesan, T. S., Rassoul, F., Guo, A. P., Th’ng, K. H., Dowding, C., Hibbin, J. A,, Young, B. D., White, H., Kumaran T. O., Galton, D. A. G. and Goldman, J. M. (1986) Rearrangement of the bcr gene in Philadelphia-chromosome negative chronic myeloid leukemia, Blood, 68, 957-960. Shepherd, P., Ganesan, T. S. and Galton, D. A. G. (1987) Haematological classification of the chronic myeloid leukaemias. In Baillibre’s Clinical Haematology: chronic myeloid leukaemia. 887-906. Ed. Goldman, J. M., Baillitre Tindall, London. Krsnic, I., Srivastava, P. C and Galton, D. A. G. (1992) Chronic myelomonocytic leukaemia and atypical chronic myeloid leukaemia. In press. Wiedemann, L. M., Karhi, K. K., Shinji, M. K., Rayter, S. I., Pegram, S. M., Dowden, G., Bevan, D., Will, A,, Galton, D. A. G. and Chan, L. C. (1988) The correlation of breakpoint cluster region rearrangement and p210 phl/abl expression with morphological analysis of Ph-negative chronic myeloid leukaemia and other myeloproliferative diseases. Blood, 71,
Clinical significanceof the Philadelphia chromosome in chronic granulocytic leukemia. Annals of Internal Medicine, 68, 1166. Ezdinli, E. Z., Sokal, J. E., Crosswhite, L. and Sandberg, A. A. (1970) Philadelphia chromosome-positive and -negative chronic myelocytic leukemia. Annals of Internal Medicine. 72, 175-1 82. Bain, B. J. and Wickramasinghe, S. N. (1976) Relationship between concentration of neutrophils and monocytes in venous blood. Acta Haematologica, 55, 89-94. Tricot, G., Vlietinck, R., Boognaerts, M. A,, Hendrickx B., De Wolf-Peeters, C., Van den Berghe, H. and Verwilghen, R. L. (1985) Prognostic factors in the myelodysplastic syndromes: importance of initial data on peripheral blood counts, bone marrow cytology, trephine biopsy and chromosomal analysis. British Journal of Haematology, 60,19-32. May, S. J., Smith, S. A., Jacobs, A., Williams, A. and Bailey-Wood, R. (1985) The myelodysplastic syndromes: analysis of laboratory characteristics in relation to the FAB classification. British Journal of Haematology, 59, 3 11-3 19. Mufti, G. J. and Galton, D. A. G. (1986) Myelodysplastic syndromes: natural history and features of prognostic importance. In Clinical Haematology: Myelodysplastic syndromes. Ed. Griffin, J. D., W. B. Saunders Co., London. Franzen, S., Strenger, G . and Zajicek, J. (1961) Microplanimetric studies on megakaryocytes in chronic granulocytic leukaemia and polycythaemia Vera. Acta Haematoloyica. 26,
2.
3.
4.
5. 6.
349-355. 7. Martiat, P., Michaux, J. L. and Rodhain, J., for the Groupe Francais de Cytogenenetique Hematologique (1991) Phila-
delphia-negative (Ph-) chronic myeloid leukemeia (CML); comparison with Ph+CML and chronic myelomonocytic leukemia. Blood, 78, 205-21 1.
17. 18. 19.
20.
21.
22.
182-193. 23. Bennett, J. M., Catovsky D., Daniel, M. T., Flandrin, G., Galton, D. A. G., Gralnick, H. R. and Sultan, C. (1989) Proposals for the classification of chronic (mature) B and T lymphoid leukaemias. Journal of Clinical Pathology. 42, 567-584.
Leukemia & Lymphoma Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/ilal20
Haematological Differences Between Chronic Granulocytic Leukaemia, Atypical Chronic Myeloid Leukaemia, and Chronic Myelomonocytic Leukaemia D. A. G. Galton
a
a
Honorary Consultant Physician, Hammersmith Hospital; Leukaemia Research Fund Professor Emeritus of Haematological Oncology, Royal Postgraduate Medical School, University of London. Formerly Honorary Director, Medical Research Council's Leukaemia Unit, London, UK Published online: 01 Jun 2015.
To cite this article: D. A. G. Galton (1992) Haematological Differences Between Chronic Granulocytic Leukaemia, Atypical Chronic Myeloid Leukaemia, and Chronic Myelomonocytic Leukaemia, Leukemia & Lymphoma, 7:5-6, 343-350 To link to this article: http://dx.doi.org/10.3109/10428199209049789
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Leukemia and Lymphoma, Vol. 7, pp. 343-350
0 1992 Harwood Academic Publishers GmbH
Reprints available directly from the publisher Photocopying permitted by license only
Printed in the United Kingdom
Haematological Differences Between Chronic Granulocytic Leukaemia, Atypical Chronic Myeloid Leukaemia, and Chronic Myelomonocytic Leukaemia Downloaded by [University of Otago] at 14:50 17 July 2015
D. A. G. GALTON Honorary Consultant Physician, Hammersmith Hospital; Leukaemia Research Fund Professor Emeritus of Haematological Oncology, Royal Postgraduate Medical School, University of London. Formerly Honorary Director, Medical Research Council’s Leukaemia Unit, London, U.K. (Received 25 January 1992)
Chronic myeloid leukaemia (CML) is a generic term that include five apparently distinct entities. The best known form, the classical Ph-positive subtype, accounts for about 90% of all cases of CML. The morphology of its presentation blood film is highly characteristic but is also seen in about half of the remaining 10% of cases, which are Ph-negative. This classical morphological subtype, whether Ph-positive or Ph-negative 1 describe as ‘chronic granulocytic leukaemia’ to refer to the exuberant granulocytic proliferation which is its hallmark. This term is often used indiscriminately and interchangeably with ‘chronic myeloid leukaemia’ and similar terms, just as ‘chronic lymphocytic leukaemia’ was, until recently, used to cover the chronic lymphoid leukaemias in general, but is now used in a specific sensez3. Chronic granulocytic leukaemia (CGL), whether Ph-positive or Ph-negative, is almost always BCR-rearranged and associated with the production of a unique 210-kd protein with enhanced tyrosine kinase activity. Most of the remaining cases of Ph-negative C M L are examples of either chronic myelomonocytic leukaemia (CMML), a subtype almost as homogeneous as CGL, and characterized in its presentation blood film by the presence of monocytes and neutrophils but few immature granulocytes, or atypical CML (aCML), distinct from and less homogeneous than either C G L or CMML, in which some cases also share features with C G L while others share some with CMML. C M M L and aCML d o not show BCR rearrangement and are not associated with the production of p2lOkd. CGL, CMML, and aCML, though characterized on morphological features differ in their clinical features and behaviour, response to treatment and survival. It is therefore clinically important that they are diagnosed accurately. Guidelines for doing so would also permit precise descriptions of each case submitted for molecular studies, thus assisting the working out of the events resulting in each phenotype, and enriching our understanding of the molecular basis of stem-cell differentiation. KEY WORDS:
CGL
atypical CML
CMML.
Ever since the discovery of a ‘minute chromosome’ in chronic granulocytic leukaemia by Nowell and Hungerford in 19601, thought by Baikie et d Z to be a possibly specific chromosomal abnormality for the
disease and named the Philadelphia chromosome (Ph) by them, it has been recognized that in a small proportion of cases of the same or a similar condition the minute chromosome is not present. The relationship between the Ph-positive and Ph-negative has been greatly since the Ph was shown Address for correspondence: Professor D. A. G. Galton, The Old Anchorage, Wiveton, Holt, Norfolk NR25 7TH, England, U.K. to result from a reciprocal translocation rather than 343
D. A. G. GALTON
Downloaded by [University of Otago] at 14:50 17 July 2015
344
a deletion, and since the molecular events underlying the translocation were identified and shown to occur in some cases in which the microscopically visible translocation was not present. In spite of this, there is still confusion about the nosological identity of the Ph-negative cases, in particular those which lack the characteristic molecular arrangements. This is partly because in many publications in which cytogenetics and molecular investigations have been described in detail, it is not possible for the haematologist to identify the kind of disease that was studied because the haematological descriptions are incomplete. The opportunity is therefore lost for understanding the events whereby a disturbed genome is expressed in a particular morphological phenotype. I have had the rare opportunity of examining, in a relatively short period of time, the presentation films of peripheral blood and bone marrow from 97 patients with Ph-negative chronic myeloid leukaemia (CML) (Table 1). 35 patients had been entered into the Medical Research Council's chronic myeloid leukaemia trial into which both Ph-positive and Ph-negative patients were eligible for entry. These slides were paired with slides from Ph-positive patients selected at random from the 512 Ph-positive cases in the trial, and they were examined independently by Dr Patricia Shepherd and me, neither knowing the Ph status. Slides from 41 patients attending Hammersmith Hospital or Burton District Hospital were examined by Drs Isabel Krisnik, P. C. Srivastava, and me: two of these were subsequently shown to be Ph-positive.
Films from 20 patients (including 5 who were Ph-positive) were submitted to me for morphological assessment by Drs Leanne Wiedemann and Dr Li Chan who were determining the BCR status and tyrosine kinase (PTK) activity. Finally films from 7 patients whose BCR status was to be studied by Dr T. S. Ganesan were examined for Professor J. M. Goldman: one of these was later found to have a complex chromosomal rearrangement with a masked t(9;22). The following observations were made on each blood film: first, a differential count on at least 300 cells and, secondly, qualitative assessments of the morphology of the erythrocytes, platelets and cells of the megakaryocytic, granulocytic, and monocytic lineages, with special reference to the occurrence and severity of dysplastic features. The Ph-negative chronic myeloid leukaemias account for about 9% of all cases of CML, and about half of them are indistinguishable from the common form of classical Ph-positive CML. Thus, in a random series of 100 cases of CML only 4 will be examples of the unusual variants. The opportunity to study so many examples of such extremely rare conditions in a relatively short period of time helped us to identify recurring patterns of differential counts and characteristic morphological features and so to characterize three main subtypes of CML. We can add nothing to the descriptions of the extremely rare juvenile CML and chronic neutrophilic leukaemia. The details of our analysis have been published e l ~ e w h e r e ~and - ~ here I discuss our conclusions, first, in relation to practical
Table 1 Morphological diagnosis in 122 cases of chronic myeloid leukaemia Source of slides
Morphological category
MRC*
CGL Ph' CGL PhaCML CMML CNL jCM L trs CGL Ph+ Total
19 18 10 7
H H und BDH**
CWt 5 5
14
6
23 2
1 1
JMGtt
Totul
I
25 26 33 31 3 2 2 122
3 3
2 2 44
41
20
7
CGL: chronic granulocytic leukaemia. aCML: atypical chronic myeloid leukaemia. CMML: chronic myelomonocytic leukaemia. CNL: chronic neutrophilic leukaemia. jCML: juvenile chronic myeloid leukaemia. * MRC: Medical Research Council; observers D. A. G . G. and Dr Patricia Shepherd. ** HH and BDH: Hammersmith Hospital and Burton District Hospital; observers: D. A. G . G., Dr Isabel Krsnik and Dr P. C. Srivastava. t CW: Drs Chan and Wiedemann; observer: D. A. G. G. tt JMG: Professor John Goldman; observer: D. A. G . G.
Downloaded by [University of Otago] at 14:50 17 July 2015
CGL. ATYPICAL CML AND CMML
diagnostic haematology, and secondly in the context of the sequence of events whereby a molecular alteration in the genome is expressed in a particular clinical-morphological phenotype. Before outlining the cytomorphological profiles of the three main subtypes of CML, it should be noted that we found the bone marrow aspirate to have limited if any diagnostic value. In all subtypes the marrow is characteristically hypercellular, usually devoid of fat, and shows marked granulocytic hyperplasia with full maturation. The common type, namely chronic granulocytic leukaemia (CGL), which is BCR + , and usually Ph + , typically has the highest granu1ocyte:nucleated erythroid-cell ratio, and increased numbers of megakaryocytes, a high proportion of which are small hypolobulated forms: chronic myelomonocytic leukaemia (CMML) has the lowest granu1ocyte:nucleated erythroid-cell ratio, while atypical chronic myeloid leukaemia (aCML) often has reduced though usually dysplastic megakaryocytes, but the overlap in these features is too great for them to have diagnostic value. The bone marrow aspirate has been overrated as a diagnostic aid in CML, while the presentation blood film has been undervalued. Martiat et d.’, however, have recently found the percentage of bone marrow erythroblasts to be a significant discriminant between CGL on the one hand and aCML and CMML on the other.
CHRONIC GRANULOCYTIC LEUKAEMIA (CGL) All 25 cases of CGL in chronic phase, Ph-positive, were correctly identified. The haematological features in presentation blood films, and those found by Shepherd in 100 further cases in the MRC series, confirmed the earlier findings of Spiers, Bain and Turner (1977)*.They may be summarized as follows: 1) The mean platelet count is above the upper limit of the normal range. 2) The differential leucocyte count shows prominent peaks in the neutrophils and in the myelocytes plus metamyelocytes, and the mean percentage of these cells increases as the total leucocyte count increases. 3) There is absolute basophilia (upper limit of normal 0.14 x 109/1) 4) The monocyte count is less than 3% and decreases further with increasing leucocyte counts.
345
5) Dysplastic features are minimal or absent in all cell lineages and are found in less than 10% of the cases. This characteristic combination of features is highly distinctive and in most cases CGL can be recognized at once in a well-stained presentation film. However, in early cases, when the total leucocyte count is below 50 x 109/1 the very low monocyte:granulocyte ratio may be less marked and if, by chance, the basophil count is low, confusion with aCML may arise. In 26 out of the 97 cases of Ph-negative CML, a morphological diagnosis of classical CGL was made. These 26 cases could not be distinguished on morphological grounds from the 25 Ph-positive cases. Furthermore, 14 of the 16 whose BCR status was investigated proved to be positive. The conclusion, therefore, is that CGL is a discrete entity with a characteristic morphological profile as seen in the presentation blood film. The general experience is that the vast majority of cases are BCR-positive, including the few that lack the t(9;22). Among the 71 remaining cases of Ph-negative CML, 31 stood out as a homogeneous group with a morphological profile distinct from that of CGL. This was chronic myelomonocytic leukaemia (CMML), now to be discussed.
CHRONIC MYELOMONOCYTIC LEUKAEMIA (CMML) Although so much has been written about CMML, there is still no generally accepted definition, so the range of features included in different definitions is wide, while one possibly crucial diagnostic feature, namely the percentage of immature granulocytes in the blood film at presentation, is rarely referred to. All agree that CMML is a chronic leukaemia predominantly affecting the elderly, that monocytosis is a constant feature, often associated with neutrophilia, thrombocytopenia, or anaemia in various combinations, that the bone marrow, though hyperplastic, frequently shows fewer monocytes than might be expected from the peripheral blood picture, and that about a third of the patients die from acute myeloid leukaemia. Non-specific esterase-stained bone-marrow films, however, may show a higher proportion of esterase-positive cells than would be expected from Romanowsky-stained films. CMML may be discovered by chance, or may present with symptoms caused by anaemia or thrombocytopenia.
Downloaded by [University of Otago] at 14:50 17 July 2015
346
D. A. G. GALTON
A palpable spleen, the commonest physical sign, is present in about a third of patients. Among patients with this range of features, the leucocyte count ranges from leucopenic to high levels of the order found in CML. A second prominent but variable feature is the presence in various combinations and in different degrees, of dysplasia in the megakaryocytic, erythroid or granulocytic series of cells. Indeed, there are so many cases in which, apart from monocytosis, the clinico-pathological picture is that of a myelodysplastic syndrome in any of its named variants, that the FAB Co-operative Group” felt that any case of myelodysplastic syndrome could be diagnosed as CMML when the monocyte count was higher than 1 x 109/l. However, the myelodysplastic syndromes are accepted primarily as conditions in which bone marrow hyperplasia is associated with one or more cytopenias, and the question arises as to whether the prominent neutrophilia often found in CMML is compatible with a myelodysplastic state. To some extent the dilemma may be resolved by considering the biological significance of cell death in attempting to distinguish myelodysplasia from myeloproliferation. Both states involve granulocytic hyperplasia, but in dysplasia early death of neutrophils renders the hyperplasia ineffective with resultant neutropenia, whereas in myeloproliferation the neutrophils survive, even though they may be abnormal, and neutrophilia ensues. If early cell death is considered a fundamental property of dysplasia, CMML with neutrophilia cannot be accepted as a myelodysplastic syndrome, but if early cell death is regarded as a secondary phenomenon, there would be no reason for excluding CMML from the MDS because of the neutrophilia and no inconsistency in accepting CMML with neutrophilia as both myelodysplastic and myeloproliferative4. However, the issue does not depend only on the interpretation of neutrophilia. The extent of dyserythropoiesis and of dysgranulopoiesis and of dysmegakaryocytopoiesis varies greatly in CMML. It is sometimes claimed that trilineage myelodysplasia is more frequent in leucopenic cases, and that many patients with high leucocyte counts have minimal or no dysplastic features in any cell lineage, or dysplastic features are confined to the granulocytic lineage. The implication here is that the patients with leucopenia and those with leucocytosis belong to two separate diagnostic groups and possibly to biologically distinct entities. The latter might be accepted as true CMML, the former simply as MDS with monocytosis. The issue is not merely academic because myeloprolifera-
tive conditions might turn out to respond to different types of treatment from that effective in myelodysplastic states. Even now CMML with leucocytosis is often treated with drugs, such as hydroxyurea or mercaptopurine, that would not be considered for MDS, while cytarabine at low dosage, or 13-cis-retinoic acid are not thought appropriate for myeloproliferative disease. The practical importance of resolving the dilemma is apparent in considering for example, the assessment of a new treatment. In designing a randomized trial of a new treatment for CMML, the eligibility criteria would have to be precisely defined. At present there are no guidelines for separating ‘CMML’ from ‘MDS with monocytosis’: cut-off points would necessarily be arbitrary and could lack biological justification, for it is not yet known whether essential differences exist between the ‘high’ and ‘low’ count cases. High-count CMML clearly qualifies as a ‘chronic myeloid leukaemia’ because it shows the cardinal features of bone-marrow hyperplasia with full granulocytic maturation, leucocytosis with neutrophilia, and a considerable risk of transformation to a blast-cell leukaemia. It has been considered to account for many of the cases reported as Ph-negative CML”.’’. Seven of our cases (with leucocyte counts up to 256 x 109/1) were admitted to the MRC CML trial. Our remaining 24 cases included those with leucocyte counts from 3.9 x 109/1 to 44 x 109/l. Our enquiry was limited in scope. It was designed to examine the morphological relationship between Ph-positive and Ph-negative CMLs. In the event, CMML has emerged as a remarkably homogeneous entity apart from the wide range of leucocyte counts. O n morphological grounds alone clear differences between the low-count and highcount cases have not emerged, and other methods will be required to reveal differences. The major morphological features in our series of 31 cases of CMML were as follows:
1) monocytosis with a low neutrophi1:monocyte ratio. Up to 10% of the monocytes may have azurophilic granules. 2) immature granulocytes in the peripheral blood at presentation do not exceed 15% and account for less than 5% of total leucocytes in a majority of cases. 3) the neutrophils may be either normal or d ysplastic. 4) the absolute basophil count is almost always within the normal range. These features, collectively, contrast markedly with
CGL, ATYPICAL CML AND CMML
Downloaded by [University of Otago] at 14:50 17 July 2015
those of CGL. The most surprising finding was the consistently low percentage of immature granulocytes: the characteristic appearance of the CMML blood film was a monotonous mixture of mature neutrophils and monocytes with some lymphocytes, and with only an occasional myelocyte or metamyelocyte. 33 of the remaining 40 cases of Ph-negative CML shared features that excluded them from both the CGL and CMML categories, though some cases also shared features with either CGL or CMML. This group was thus less homogeneous than either CGL or CMML, but nevertheless its constellation of features warranted its designation as a distinct entity which, for some years, we have provisionally named ‘atypical chronic myeloid l e ~ k a e m i a ’ ~’. - ~ . ~
ATYPICAL CHRONIC MYELOID LEUKAEMIA The major morphological features in our series of 33 cases of aCML were as follows: 1) basophils are absent or few in the majority of presentation blood films. 2) immature granulocytes and mature neutrophils are markedly dysplastic. In the histogram of the differential counts, there are neutrophil and myelocyte/metamyelocyte peaks as in CGL, but in a minority of cases the immature granulocytes (including promyelocytes) are below 15%. 3) the monocyte counts range from less than 1 x 109/1 to 50 x 109/1 but exceed 4% in two thirds of the cases. When the monocyte counts are 4% or higher, confusion with CMML is possible, and when they are less than 4%, confusion with CGL is possible. The most distinctive profile for the presentation blood film is the combination of low basophil and high monocyte percentages with high percentages of immature granulocytes: the immature granulocytes and mature neutrophils are markedly dysplastic. Equally distinctive but less common is the same combination except for the monocyte percentage which is low. Less common still are cases in which the monocyte percentage is below 4 and the immature granulocyte percentage below 15. When granulocytic dysplasia was prominent we diagnosed aCML. However, two of our patients proved to be Ph-positive and the prominent granulocytic dysplasia was associated with early transformation. In the absence of basophilia or
341
of blast cells it may not be possible to distinguish aCML from transformed CGL. An inherent difficulty in the recognition of aCML arises from the fact that so many of its features, like those of CGL and CMML, are continuous variables which overlap at the extremes of their ranges4. In a particular case of CGL, for example, the differential count may suggest aCML rather than with CGL; substandard staining may suggest dysgranulopoiesis thus supporting, erroneously, a diagnosis of aCML. This may be the most likely explanation for the occasional reports of BCR positivity in cases of Ph-negative aCML3. If this is correct a small number of wrong diagnoses will be unavoidable. But it is also possible that better criteria for the diagnosis of aCML may reduce the number of errors. This matter is being examined by the FAB Co-operative Group. An obvious criticism of the concept of aCML as an entity is that it shows a range of variability in respect of several features, particularly the percentages of monocytes and of immature granulocytes, which at one end of the range merge with values seen in CGL and at the other with those of CMML. This could imply that the three entities are artificial constructs and that in reality there is a continuum ranging between the extreme expressions of a single biological entity. Martiat et al.’ have compared 35 cases of Ph-, BCR- CML (aCML) with 55 of Ph-, BCR’ CGL, and 100 of CMML. Their conclusions essentially confirm our findings, but they speculate that aCML and CMML may be ‘two aspects of the same disorder’, distinct from CGL. Recently a group of seven Ph-negative patients described as ‘chronic neutrophilic leukaemia with dysplastic features’ has There . was been reported by Symeonides et ~ 1 . ’ ~ persistent neutrophilia, no monocytosis, or immature granulocytes, and there were marked dysplastic features in the granulocytic cells and megakaryocytes. The haematological profile in these cases may represent a fully mature variant of the low-monocyte group of aCML, clearly far removed from CGL, but close to chronic neutrophilic leukaemia. Evidence from other sources supports the sharp separation of CGL from aCML and CMML.
Clinical and haematological evolution of CGL, CMML, and aCML Neither aCML nor CMML shows the regular pattern of evolution of CGL in which there is a close relationship between the size of the spleen, the height of the leucocyte count, and the haemoglobin
Downloaded by [University of Otago] at 14:50 17 July 2015
348
D. A. G . GALTON
scales but patients with very high neutrophil counts other than those with CGL were not investigated. However, if the ratio was assumed to have been the same at all neutrophil counts, those for CGL proved to be abnormal, with the monocyte counts much lower than would have been expected in relation to the neutrophil counts, suggesting that the control mechanism that normally gears the production of neutrophils to the level of circulating monocytes is lost or much weakened in CGL. In our studies4 the neutrophi1:monocyte ratio was the same in Phpositive and Ph-negative CGL in contrast to that for CMML which was abnormal in the opposite direction, with the monocyte counts much higher than would have been expected in relation to the neutrophil counts. In aCML the neutrophi1:monocyte ratio was also abnormal. When the monocyte counts were 4% or higher, the neutrophi1:monocyte ratio, as in CMML, was low, though the neutrophil count was somewhat higher in relation to the monocyte count. However, when the monocyte counts were less than 4%, the ratio was abnormal in the reverse sense, as in CGL, with neutrophil counts abnormally high in relation to the monocyte counts5. This observation suggests that the difference between the high-monocyte count and Survival in aCML low-monocyte count aCML is determined by a switch The number of cases of aCML in this series is too mechanism controlling granulocyte/monocyte differsmall to make valid comparison of survival in this entiation. CGL, whether Ph-positive or Ph-negative, is disease with that for CGL, but ih the MRC CML trial4 the survival was markedly wqrse and may have strongly associated with low monocyte and high been worse than that of CMML. It has, of course, basophil counts, while in CMML the reverse is found. been known for many years that the survival in However, in aCML there was no clear Dattern in the Ph-negative CML is inferior to that in Ph-positive cases. CML'6*'7;because about half of the cases would have been examples of Ph-negative CGL, in which the Molecular biology survival is the same as in Ph-positive CGL, the survival of the aCML cases among them would have CGL, whether Ph-positive or Ph-negative, is strongly associated with BCR positivity, and with the been even shorter. production of the hybrid 210-kd protein with enhanced tyrosine kinase activity6. In contrast, BCR Other features rearrangement is not found in CMML, or in the great The relationships between the presentation neutrophil majority of cases of aCML; the few exceptions are and monocyte counts, and between the monocyte likely to be diagnostic mistakes, but more studies are counts in CGL, aCML, and CMML suggest that the needed. The 210-kd protein has not been found in regulation of monocyte/granulocyte differentiation CMML or aCML. and maturation, and the production of basophils differ in the three conditions. Bain and Wickramasinghe" Myelodysplasia in aCML investigated the neutrophil: monocyte ratios at various levels of neutrophil counts in normal subjects Morphological evidence of dysplasia, though charincluding pregnant women, and in patients with acteristic of the myelodysplastic syndromes, is often neutrophilia. The ratio was linear on logarithmic absent in refractory anaemia (RA) and RA with ring
concentration". For example, among the 169 patients admitted to the Medical Research Council's randomized trial of splenectomy in Ph-positive CGL' 5 , none of those whose presentation leucocyte counts were less than 50 x lo9/], and only 4 of the 20 whose counts were between 50 and 100 x 109/1, had a spleen 10cm or more below the costal margin, while the haemoglobin concentration was below 10 g/dl in only one of the 30 patients whose leucocyte counts were below 100 x lo9/]. In aCML the relationship is far less consistent so that aCML may be suspected even before the blood film has been examined, and associations of leucocyte counts, spleen size, and haemoglobin concentration unlikely in CGL are encountered. The impression is strengthened if the platelet count is below 150 x 109/1, as was the case in 12 of the 14 aCML patients in the Hammersmith/ Burton series, in contrast to just under 10% in the MRC trial patient^'^, 6 of the eight aCML patients whose leucocyte counts were below 56 x lo9/] had a haemoglobin concentration between 3.5 and 9.8 g/dl. The neutrophil alkaline phosphatase score, low in 95% of cases of CGL, was high in one third of the cases of aCML.
CGL, ATYPICAL CML AND CMML
sideroblasts (RARS). Some authors’’ exclude such cases from the MDS, while others” would accept as early MDS cases in which persistent unexplained macrocytosis is the only abnormality, because subsequent follow-up of these patients shows that they evolve to conventionally characterized MDS. The current debate on the grounds for excluding high-count CMML from the MDS derives from the recognition of two separate components of the disease, namely the myelodysplastic and the myeloproliferative elements, as already discussed. aCML has at least an equal claim to that of CML to be included as a myelod ysplastic/myeloproliferative disorder2 All granulocytic series cells show dysplastic features in varying degree: deficient or absent primary granulation in promyelocytes, deficient or absent secondary granulation in myelocytes, metamyelocytes, and neutrophils, nuclear abnormalities including PelgerHiiet forms, bizarre hypersegmentation, and abnormal chromatin clumping, are all common in aCML. Dysmegakaryocytopoiesis is the rule in aCML, and dyserythropoiesis is often seen. No case with a leucocyte count below 20 x 109/1 appeared in our series. Dysplasia is a major feature that distinguishes aCML from CGL in which dyserythropoiesis and dysgranulopoiesis are minimal or absent in the chronic phase, although the long recognized small megakaryocytes22, often resembling the micromegakaryocytes of MDS are characteristic. The significant point is that however the prominent myelodysplastic features of aCML and CMML are interpreted in relation to the myelodysplastic syndromes, they add to the list of features that separate aCML and CMML from CGL.
Downloaded by [University of Otago] at 14:50 17 July 2015
’.
349
confirmed, it is to be hoped that detailed recommendations for the definition of the subtypes will be forthcoming. These would be important for two reasons. First, by improving diagnostic accuracy, the comparability of reports from different sources would be much greater than it is now: secondly, the correlation between work at the molecular level with that at the morphological level would be strengthened. This would benefit the hard task of unravelling the sequential processes that link the genome with its ultimate expression in the observed phenotype. Understanding of the molecular biology of CGL and the definition of its phenotype are at a more advanced stage than is the case with any other haematological neoplastic disease, and the existence of three somewhat similar but distinct phenotypes that appear to have a different molecular genetic basis offers an unrivalled opportunity for seeking the explanation of the molecular events responsible for the divergent phenotypes.
SUMMARY
Chronic myeloid leukaemia (CML) is a generic term that include five apparently distinct entities. The best known form, the classical Ph-positive subtype, accounts for about 90% of all cases of CML. The morphology of its presentation blood film is highly characteristic but is also seen in about half of the remaining 10% of cases, which are Ph-negative. This classical morphological subtype, whether Ph-positive or Ph-negative I describe as ‘chronic granulocytic leukaemia’ to refer to the exuberant granulocytic proliferation which is its hallmark. This term is often used indiscriminately and interchangeably with ‘chronic myeloid leukaemia’ and similar terms, CONCLUSIONS just as ‘chronic lymphocytic leukaemia’ was, until recently, used to cover the chronic lymphoid Larger numbers of cases than those reviewed here leukaemias in general, but is now used in a specific must be studied to see whether the grouping of CML sense23. Chronic granulocytic leukaemia (CGL), into three main subtypes, namely CGL, aCML and whether Ph-positive or Ph-negative, is almost always CMML, with two excessively rare subtypes, juvenile BCR-rearranged and associated with the production CML and chronic neutrophilic leukaemia, holds of a unique 210-kd protein with enhanced PTK good, and whether the features characteristic of the activity. less common types can be precisely defined. In our Most of the remaining cases of Ph-negative CML study, aCML and CMML were characterized on are examples of either chronic myelomonocytic morphological features alone, but the characterization leukaemia (CMML), a subtype almost as homogenis supported by evidence from clinical, haematologi- eous as CGL, and characterized in its presentation cal, and negatively, cytogenetic and molecular blood film by the presence of monocytes and biological sources, as described above. neutrophils but few immature granulocytes, or If the classification of CML suggested here is atypical CML (aCML), distinct from and less
D. A. G. GALTON
350
homogeneous than either CGL or CMML, in which some cases also share features with CGL while others share some with CMML. CMML and aCML do not show BCR rearrangement and are not associated with the production of p2lOkd. CGL, CMML, and aCML, though characterized on morphological features differ in their clinical features and behaviour, response to treatment and survival. It is therefore clinically important that they are diagnosed accurately. Guidelines for doing so would also permit precise descriptions of each case submitted for molecular studies, thus assisting the working out of the events resulting in each phenotype, and enriching our understanding of the molecular basis of stem-cell differentiation.
8. Spiers, A. S. D., Bain, B. J. and Turner, J. E. (1977) The 9. 10.
11.
12. 13.
Downloaded by [University of Otago] at 14:50 17 July 2015
14.
Acknowledgements It is a pleasure to thank Dr Barbara Bain for her helpful comments and critical reading of the text.
15.
REFERENCES
peripheral blood film in chronic granulocytic leukaemia. Scandinavian Journal of Haematology, 18, 25-38. Bain. B. J. and England, J. M. (1975) Normal haematological values: sex difference in the neutrophil count. British Medical Journal, 11, 473475. Bennett, J. M., Catovsky, D., Daniel, M. T., Flandrin, G., Galton, D. A. G., Gralnick, H. R. and Sultan, C. (1982) Proposals for the classification of the myelodysplastic syndromes. British Journal of Haematology, 51, 189-199. Pugh, W. C., Pearson, M., Vardiman, J. W. and Rowley, J. D. (1985) Philadelphia-chromosome negative chronic myeloid leukaemia: a morphological reassessment. British Journal of Haematology, 60,456-467. Travis, L. B., Pierre, R. V. and De Wald, G. W. (1986) Ph’ negative chronic granulocytic leukemia: a non-entity. American Journal of Clinical Pathology, 85, 186193. Galton, D. A. G. (1982) The Chronic Leukaemias. In Blood and its Disorders. Eds Hardisty, R. M. and Weatherall, D. J. pp. 877-897. Blackwell, Oxford. Symeonides, A., Kovrakli, A,, Katevas, P., Pagonis, S., Zikos, P., Aktipi, A,, Tiniakou, M., Matsouka, P. and Zoumbas, N. (1991) Chronic neutrophilic leukaemia with dysplastic features (CNL-D). A report of seven patients. Abtstract AS. Second International Conference on Myelodysplastic Syndromes. Leukaemia Research, 15, supplement. MRC (1983) Randomized trial of splenectomy in Ph positive chronic granulocytic leukaemia including an analysis of prognostic features. British Journal of Haematoloqy. 54,
1. Nowell, P. C. and Hungerford, D. A. (1960) A minute
415430. 16. Canellos, G., Eagam, R. T., Whang, J. and Carbone, P. P. (1968)
chromosome in human chronic granulocytic leukemia. Abstract. Science, 132, 1497. Baikie, A. G., Court-Brown, W. M., Buckton, K. E., Harnden, D. G., Jacobs, P. A. and Tough, I. M. (1960) A possible specific chromosome abnormality in human chronic myeloid leukaemia. Nature, 188, 1165-1166. Ganesan, T. S., Rassoul, F., Guo, A. P., Th’ng, K. H., Dowding, C., Hibbin, J. A,, Young, B. D., White, H., Kumaran T. O., Galton, D. A. G. and Goldman, J. M. (1986) Rearrangement of the bcr gene in Philadelphia-chromosome negative chronic myeloid leukemia, Blood, 68, 957-960. Shepherd, P., Ganesan, T. S. and Galton, D. A. G. (1987) Haematological classification of the chronic myeloid leukaemias. In Baillibre’s Clinical Haematology: chronic myeloid leukaemia. 887-906. Ed. Goldman, J. M., Baillitre Tindall, London. Krsnic, I., Srivastava, P. C and Galton, D. A. G. (1992) Chronic myelomonocytic leukaemia and atypical chronic myeloid leukaemia. In press. Wiedemann, L. M., Karhi, K. K., Shinji, M. K., Rayter, S. I., Pegram, S. M., Dowden, G., Bevan, D., Will, A,, Galton, D. A. G. and Chan, L. C. (1988) The correlation of breakpoint cluster region rearrangement and p210 phl/abl expression with morphological analysis of Ph-negative chronic myeloid leukaemia and other myeloproliferative diseases. Blood, 71,
Clinical significanceof the Philadelphia chromosome in chronic granulocytic leukemia. Annals of Internal Medicine, 68, 1166. Ezdinli, E. Z., Sokal, J. E., Crosswhite, L. and Sandberg, A. A. (1970) Philadelphia chromosome-positive and -negative chronic myelocytic leukemia. Annals of Internal Medicine. 72, 175-1 82. Bain, B. J. and Wickramasinghe, S. N. (1976) Relationship between concentration of neutrophils and monocytes in venous blood. Acta Haematologica, 55, 89-94. Tricot, G., Vlietinck, R., Boognaerts, M. A,, Hendrickx B., De Wolf-Peeters, C., Van den Berghe, H. and Verwilghen, R. L. (1985) Prognostic factors in the myelodysplastic syndromes: importance of initial data on peripheral blood counts, bone marrow cytology, trephine biopsy and chromosomal analysis. British Journal of Haematology, 60,19-32. May, S. J., Smith, S. A., Jacobs, A., Williams, A. and Bailey-Wood, R. (1985) The myelodysplastic syndromes: analysis of laboratory characteristics in relation to the FAB classification. British Journal of Haematology, 59, 3 11-3 19. Mufti, G. J. and Galton, D. A. G. (1986) Myelodysplastic syndromes: natural history and features of prognostic importance. In Clinical Haematology: Myelodysplastic syndromes. Ed. Griffin, J. D., W. B. Saunders Co., London. Franzen, S., Strenger, G . and Zajicek, J. (1961) Microplanimetric studies on megakaryocytes in chronic granulocytic leukaemia and polycythaemia Vera. Acta Haematoloyica. 26,
2.
3.
4.
5. 6.
349-355. 7. Martiat, P., Michaux, J. L. and Rodhain, J., for the Groupe Francais de Cytogenenetique Hematologique (1991) Phila-
delphia-negative (Ph-) chronic myeloid leukemeia (CML); comparison with Ph+CML and chronic myelomonocytic leukemia. Blood, 78, 205-21 1.
17. 18. 19.
20.
21.
22.
182-193. 23. Bennett, J. M., Catovsky D., Daniel, M. T., Flandrin, G., Galton, D. A. G., Gralnick, H. R. and Sultan, C. (1989) Proposals for the classification of chronic (mature) B and T lymphoid leukaemias. Journal of Clinical Pathology. 42, 567-584.
