Dysgranulopoiesis is an independent adverse prognostic factor in chronic myeloid disorders with an isolated interstitial deletion of chromosome 5q

Myelodysplastic syndromes (MDS) are acquired clonal bone marrow disorders characterized by peripheral blood cytopenias and abnormal bone marrow hematopoiesis. An interstitial deletion of the long arm of chromosome 5 [del(5q)], either in isolation or associated with other chromosomal abnormalities, is the most frequent karyotypic abnormality identified in MDS. A distinct subset of MDS with isolated del(5q) was recognized first by Van den Berghe et al.1 and later by others2, 3, 4, 5 owing to its unique clinical and pathologic features ('5q- features'), which included a female preponderance, macrocytic anemia, frequent thrombocytosis, monolobated megakaryocytes and favorable clinical outcomes. This clinical entity was subsequently renamed by the 2001 World Health Organization (WHO) Classification of Hematopoietic Neoplasms6 as '5q- syndrome' and required only the presence of an isolated del(5q) and less than 5% bone marrow blasts. This definition was primarily based on several studies2, 3, 4 that showed the adverse prognostic impact of blast count in peripheral blood or bone marrow. However, the WHO classification did not address two issues: first, the prototypic clinical and pathologic '5q- features,' which are not a prerequisite in the WHO diagnostic criteria, are frequently absent in cases that fulfill the WHO definition of '5q- syndrome'. The prognostic significance of the prototypic '5q- features' is unclear. Second, the prognostic impact of dysgranulopoiesis in the 5q- syndrome was not addressed. Both issues are frequently encountered in clinical practice but have remained unanswered in the literature. To address both issues, we retrospectively studied the clinical and pathologic features of patients with isolated del(5q)-associated chronic myeloid disorders from the pre-Lenalidomide era. We searched our institutional cytogenetic database (1990–2005) for patients whose peripheral blood or bone marrow had a clonal isolated del(5q), which was defined as two or more metaphases with the interstitial chromosome deletion involving 5q11–q35. We identified a total of 68 patients, 38 women and 30 men, with a median age of 72 years (range 8–91) at the time of initial diagnosis (Table 1). Patients with blast percentages greater than 20% or who were diagnosed with MDS at an outside institution greater than 6 months before an initial consultation at Mayo Clinic were excluded. The type of break points and frequency of abnormal metaphases of each case were summarized in Tables 1 and 3. The three most frequently encountered chromosome 5 break points were q13q33 (46%), q15q33 (41%) and q22q33 (6%). There were no correlations between the type or frequency of del(5q) and clinical and pathologic features or overall survival, which is consistent with the findings from others.5 The hematological and morphologic findings of the isolated del(5q)-associated myeloid disorders varied significantly (summarized in Table 2). At diagnosis, most patients had anemia; among those with anemia, less than half had macrocytic red blood cell indices. Neutropenia and thrombocytopenia were common. Thirteen patients had pancytopenia (19%). Neutrophilia, thrombocytosis and monocytosis were infrequent at the time of initial presentation. The spectrum of bone marrow morphologic findings was broad. Most patients had a hypercellular bone marrow (median cellularity 50%, range 20–95%). Dyserythropoiesis was present in 29 cases, and half of those showed more than 15% ringed sideroblasts. Dysgranulopoiesis characterized by nuclear hyposegmentation or cytoplasmic hypogranulation in more than 10% granulocytic precursors was present in almost half of the cases (n=31, 46%). Most of those cases also had significantly increased dual -naphthyl butyrate esterase and chloroacetate esterase (BE/CLE)-positive abnormal myeloid precursors (Figure 1a) as defined earlier.7 In addition, three cases (4%) were classified as having dysgranulopoiesis based solely on the presence of markedly increased dual BE/CLE-positive abnormal myeloid precursors. Small and monolobated megakaryocytes were the most consistent morphologic finding. However, rare cases showed clusters of large and bizarre megakaryocytes (Figure 1b). Increased myeloblasts in either peripheral blood or bone marrow (5–19%) were present in 26 cases (39%). Reticulin fibrosis was present in nine cases. On the basis of the current WHO classification,6 36 cases were classified as '5q- syndrome' (isolated del(5q) and <5% blasts) and 25 cases were classified as refractory anemia with excess blasts (RAEB) (detailed in Table 3). Two cases were best classified as chronic myelomonocytic leukemia (CMML) based on a persistent monocytosis. Both patients, one female and one male, did not have the prototypic '5q- features,' and both had aggressive clinical courses with survivals of less than 15 months. It remains unclear if such cases should be excluded from 5q- syndrome category. Five patients were best classified as myeloproliferative neoplasms (MPN) based on the presence of prominent clustering of large megakaryocytes in addition to scattered small monolobated megakaryocytes (Figure 1b) (five of five patients), reticulin fibrosis (four of five patients, grades 1–3) and splenomegaly (two of five patients). These five cases were reported earlier in a larger study.8 For the 36 cases that were classified as 5q- syndrome by WHO criteria, there was not a uniform morphologic appearance. As the morphologic findings and classification typically precede the cytogenetic karyotyping results, we also attempted to classify these cases using the current WHO morphologic classification scheme (Table 3). Five cases exhibited no morphologic features of myelodysplasia and were morphologically classified as 'normal' bone marrows; all five patients were anemic and had no other clinical or laboratory abnormality. Two cases with increased (>15%) ringed sideroblasts as the primary morphologic abnormality were classified as refractory anemia with ringed sideroblasts. Most cases (21 patients, 31%) were classified as refractory cytopenia with multilineage dysplasia (RCMD) and had varying degrees of dyserythropoiesis, dysgranulopoiesis and dysplastic megakaryocytes in conjunction with at least a bi-cytopenia. Eight of these 21 cases also had increased ringed sideroblasts (>15%) and were thus further classified as RCMD with ringed sideroblasts. Eight cases with only dysplastic megakaryocytes were classified as MDS-unclassified (MDS-U), the terminology used for a uni-lineage, non-erythroid dysplasia. Dysgranulopoiesis was present in 6 of 11 RAEB-1 cases (55%) and 13 of 14 RAEB-2 cases (93%), but was less common in the cases that had less than 5% blasts (14 of 36 patients, 39%). Of the 36 cases that fulfilled the diagnostic definition of 5q- syndrome, only 7 cases had the prototypic '5q- features,' which included macrocytic anemia, normal or elevated platelet count and monolobated megakaryocytes. All of the seven patients were female and had no evident dysgranulopoiesis. The findings in these 36 patients address the two initially raised issues are as follows: (1) most cases that fulfill the WHO definition of 5q- syndrome do not have the prototypic '5q- features' and (2) the finding of dysgranulopoiesis is not uncommon in 5q- syndrome. To evaluate the prognostic significance of the various morphologic variables, we first determined overall survival with the Kaplan–Meier univariant analysis by grouping the cases on the basis of blast counts (<5%, 5–9%, and 10–19%), the presence or absence of dysplasia (dyserythropoiesis, ringed sideroblasts 15%, dysgranulopoiesis and dysplastic megakaryocytes) and myelofibrosis. We excluded the seven patients with CMML or MPN from these analyses. Only an elevated blast count of 10% and dysgranulopoiesis demonstrated statistical significance (P=0.002) (Figure 1c). None of the other morphologic parameters yielded any statistically significant differences in overall survival, including the patients with 5–9% bone marrow blasts (RAEB-1). We then performed a multivariate analysis using the stepwise Cox proportional hazards model and demonstrated that dysgranulopoiesis was the only significant independent adverse prognostic indicator in this group of patients (P=0.002) (Figure 1d). It is the first time that dysgranulopoiesis is demonstrated to have a potential adverse prognostic impact on myeloid disorders with an isolated del(5q), which could provide a possible explanation for the variably reported median survivals of patients with 5q- syndrome.2, 3, 4, 5 The higher prevalence of dysgranulopoiesis in RAEB cases implies that dysgranulopoiesis could be a preceding event for disease progression to a higher grade MDS or even acute myeloid leukemia in patients with an isolated del(5q). Alternatively, the presence of dysgranulopoiesis could represent a separate disease process that, when identified, could be the basis for exclusion from the 5q- syndrome category. There were no consistent types or frequencies of 5q break points that could further explain this finding (Table 3). However, the seemingly homogeneous isolated del(5q) chromosomal abnormality does not exclude the possibilities of concomitant cryptic cytogenetic abnormalities or gene mutations that can only be detected by more sensitive molecular methods. Using the Kaplan–Meier univariate analysis, we also compared the overall survival of the 7 patients who had all of the prototypic '5q- features' with the 15 patients who had an isolated del(5q), less than 5% blasts, and an absence of dysgranulopoiesis. Even though the statistical power was limited by the relative small case numbers, there was no difference in overall survival between these two groups (P=0.5). In summary, this retrospective study addressed several important issues that are unanswered by the current WHO classification of 5q- syndrome. (1) Chronic myeloid disorders with isolated del(5q) exhibit a heterogeneous clinical and pathologic spectrum, encompassing both MDS and MPN presentations. (2) Most cases that fulfill the WHO definition of 5q- syndrome do not have the prototypic clinical and pathologic '5q- features.' Our data also suggest that the presence of prototypic '5q- features' by itself is prognostically insignificant. Thus, the use of the term '5q- syndrome,' which implies a uniform disease process, may not accurately describe this heterogeneous group of chronic myeloid disorders. (3) Dysgranulopoiesis is not an uncommon occurrence in patients with 5q- syndrome and, when present, indicates an unfavorable clinical outcome. Nevertheless, in this study, all the patients were identified from the pre-Lenalidomide era. Whether these findings will remain valid after the Lenalidomide treatment is yet to be determined. Our results point out the necessity for future studies to specifically assess dysgranulopoiesis in patients with isolated del(5q).