MDS

Courtyard Pool_9919.Photograph courtesy of Henry Domke,MD.www.henrydomke.com

Myelodysplastic Syndromes:Diagnosis and Staging

Luca Malcovati, MD, and Stephen D.Nimer, MD Myelodysplastic syndromes (MDS) represent a heterogeneous group of hematologic disorders characterized

by ineffective hematopoiesis and an increased risk of developing acute myelogenous leukemia (AML).Accuratediagnosis of MDS can be difficult,and its classification requires evaluation of cytopenias,bone marrow morphology,blast percentage,and cytogenetics.These factors,as well as patient performance status and redblood cell transfusion dependence,can be used to predict prognosis in MDS.Accurate diagnosis and classificationare essential for subgroup identification and prognostic assessment of patients with MDS.This article reviewsessential criteria for staging and subgroup classification and summarizes prognostic scoring systems that aidin risk stratification and selection of optimal therapy.

Classification systems such as the World Health Organization (WHO) classification are widely used but donot always provide sufficient prognostic information.This limitation led to the creation of the International Prognostic Scoring System (IPSS).However,this system was designed to be used only at diagnosis and may notbe suitable for serial assessment of patients whose disease can evolve over time.The WHO classification-basedprognostic scoring system (WPSS) permits dynamic estimation of survival and risk of AML transformation at multiple time points during the natural course of MDS.Prognostic scoring systems such as WPSS allow for prediction of survival and risk of leukemic evolution at any time during the course of the disease.Such an approachmay provide a useful adjunct for clinical decision making,including selection of appropriate treatment options.From the Department of Hematology,University of Pavia and Fondazione Istituto di Ricovero e Cura a Carattere Scientifico Policlinico San Matteo,Pavia,Italy (LM),and the Division ofHematologic Oncology,Department of Medicine,Memorial Sloan-Kettering Cancer Center,New York,NY (SDN).

Address correspondence to Stephen D.Nimer,MD,Division of Hema-tologic Oncology,Department of Medicine,Memorial Sloan-KetteringCancer Center,1275 York Ave,Box 575,New York,NY 10021-6007.E-mail:s-nimer@mskcc.orgAbbreviations used in this paper:MDS = myelodysplastic syndromes,AML = acute myelogenous leukemia,RBC = red blood cell,WHO =World Health Organization,IPSS = International Prognostic ScoringSystem,WPSS = WHO classification-based prognostic scoring system.4Cancer Control

Introduction

Myelodysplastic syndromes (MDS) constitute a groupof clonal hematopoietic disorders characterized bybone marrow failure,dysplasia,and an increased likeli-hood of evolution to acute myeloid leukemia (AML).MDS is generally classified as “primary”(or de novo)and “treatment-related”(secondary to prior cytotoxicchemotherapy).1These disorders are thought to arisedue to abnormalities in hematopoietic stem cell self-renewal and differentiation.The incidence of MDSincreases with age,perhaps as a result of ongoing acqui-October 2008, Vol.15, No.4 Supplementsition of DNA damage,natural depletion of stem cells,and/or accumulated exposure of the bone marrow toenvironmental stresses or toxins.2

Many different conditions are grouped togetherunder the “MDS”umbrella based on common clinicalcharacteristics,thus accounting for the wide hetero-geneity observed.Diagnosis of patients with this dis-ease can be difficult at times.Similarly,the assigningof prognosis and the selection of appropriate therapyrequire careful application of prognostic scoring sys-tems taking into account clinical characteristics (eg,cytopenias age,performance status) and cytologicalparameters (eg,blast count,morphology,karyotype).3,4Factors such as poor cytogenetics are associated withdecreased survival in MDS.Knowing the specificcytogenetic abnormality can assist in the selection oftherapy for certain subgroups of MDS patients (eg,lenalidomide for patients with 5q– syndrome).Inad-equate diagnosis can lead to delayed or incorrect clas-sification of MDS that could result in inappropriatetreatment decisions,such as relying only on best sup-portive care.Several factors have been identified that can signifi-cantly impact the prognosis and selection of therapy forMDS patients,such as cytogenetics,patient performancestatus,and red blood cell (RBC) transfusion dependence.Numerous studies have shown that patient performancestatus is inversely associated with overall or event-freesurvival in patients receiving intensive chemotherapyfor MDS or AML,particularly older individuals.5-8Currenttreatment guidelines suggest that performance status beconsidered when evaluating therapeutic options sinceit affects the ability of higher-risk patients to safelyundergo aggressive treatments such as intensive induc-tion chemotherapy or hematopoietic stem cell trans-plantation.9Older MDS patients,who are more likely tohave lower performance status and major comorbidi-ties than younger patients,may not be able to toleratesuch intensive therapy.Some patients whose perfor-mance status is suboptimal may benefit from treatmentwith azacitidine or decitabine as a bridging therapy priorto transplantation.Recent advances in the diagnosis and prognosticassessment of MDS,accompanied by a greater numberof effective treatment options,have improved patientoutcomes based on proper selection of therapy.Thisreview summarizes the current classification systemsused to subcategorize MDS.Evolving prognostic scor-ing systems that facilitate assessment of patient survivaland risk of progression to AML at diagnosis as well asduring the course of their disease are discussed.Addi-tionally,factors that can adversely affect survival suchas cytogenetics,performance status,and RBC transfu-sion dependence are examined,along with criteria thatcan be used for the selection of appropriate therapy forindividual MDS subgroups.

October 2008, Vol.15, No.4 SupplementDiagnosis

The differential diagnosis of MDS can be challenging,given its heterogeneous nature and the subjective assess-ment of dysplasia.Accurate diagnosis of MDS-associateddysplasia requires a review of the peripheral blood smearas well as a bone marrow aspirate and biopsy.Morpho-logical analysis is performed to determine the percentageof myeloblasts;it is also performed because specific cyto-logical abnormalities can be found in the blood and bonemarrow.10,11Peripheral blood smears may reveal hypo-granulated neutrophils with hyposegmented nuclei (thepseudo–Pelger-Huet anomaly) and large platelets,whilebone marrow dysplasia affects the erythroid,myeloid,ormegakaryocytic lineages.1The bone marrow is usuallynormocellular or hypercellular,but up to 20% of MDSpatients have hypocellular marrow.3Cytopenias are the most common clinical featureof MDS,particularly anemia.A diagnosis of MDS mustbe distinguished from other causes of anemia such asaplastic anemia or paroxysmal nocturnal hemoglobin-uria (PNH).3Hairy cell leukemia,which can also mimicMDS clinically,should also be excluded by careful eval-uation of the bone marrow.Bone marrow findings canalso exclude a diagnosis of natural killer cell leukemia.Occasionally,human immunodeficiency virus (HIV) cancause cytopenias that mimic MDS.12Unexplained ane-mia that commonly occurs in the elderly also must beexcluded.13Other factors can cause MDS-like bonemarrow dysplasia,including vitamin B12or folate defi-ciencies,viral infections,and exposure to certain toxicagents (eg,chemotherapy,lead,benzene),but theireffects are generally transient.The presence of fibrosisin some patients requires that a diagnosis of primarymyelofibrosis be ruled out.14

Several groups have demonstrated that flow cytom-etry may have diagnostic utility for the immunopheno-typing of MDS.15-17For example,using flow cytometricanalysis,Malcovati et al18found a higher proportion ofimmature erythroid cells and immature myeloid cells inMDS.A significant correlation was found between mul-tiparameter flow analysis and both the InternationalPrognostic Scoring System (IPSS) score and the degreeof morphologic dysplasia.Flow cytometry can also beused to rule out a diagnosis of PNH or large granularlymphocytic leukemia.This technique may serve as auseful adjunct in the diagnosis and staging of MDS,although it is not routinely performed in many diag-nostic pathology laboratories outside of research oracademic institutions.

Appropriate diagnosis and classification of MDSdepends on accurate assessments of both clinical fea-tures and laboratory/pathology findings (eg,blastcount,peripheral blood counts,cytogenetics).To thisend,well-prepared bone marrow smears and biopsyspecimens are essential.Close collaboration betweenthe oncologist and pathologist is key for proper evalu-Cancer Control 5

ation of clinical findings together withbone marrow morphology,karyotype,and any additional data such as flowstudies or immunophenotyping.19

Table 1. — Common Cytogenetic Alterations in MDS

MDS SubtypeRARARSFrequency25%10%Chromosomal Aberrationsdel(5q), del(20q), –Y, –7, +8del(5q), del(20q), –Y, –7, +8, idic(X)(q13)RCMD50%del(5q), –7, +8CytogeneticsRCMD-RS50%del(5q), –7, +8Cytogenetic analysis of the bone mar-RAEB-150%del(5q), –7, +8, del(20q)row is indicated in MDS not only to

detect characteristic chromosomal ab-RAEB-250%–75%del(5q), –7, +8, –17p, del(11q), t(11q23), –13, del(13q)normalities,but also to assess for clonalMDS del(5q)100%del(5q)evolution.Chromosomal abnormalitiesData from Mufti20and Fenaux.22have been documented in 40% to 70%of patients presenting with MDS and in

the majority of patients with treatment-related MDS.with treatment-related MDS.In contrast to primaryA review of the literature,involving more than 3,000MDS,where translocations may be balanced,manycases,indicated an incidence of chromosomal aberra-unbalanced translocations were noted in patients with20

tions of approximately 48% at time of presentation.treatment-related AML.26While the remaining cases appear to have a normalCurrently used MDS prognostic scoring systemskaryotype,technical failures such as inability to obtaininclude cytogenetics as a key variable,consideringsufficient analyzable metaphases may reduce the actualboth the presence and type of chromosomal alter-proportion of abnormal cases.Fluorescence in situhy-ations.In the IPSS,for example,a normal karyotype orbridization (FISH) has been used to assess chromosomalthe presence of –Y,del(5q),or del(20q) is classified aschanges in MDS not detected by standard cytogeneticgood risk,while the presence of three or more abnor-analysis.It is particularly useful in cases where insuffi-malities (ie,complex cytogenetic abnormalities) orcient numbers of metaphases are available for standardchromosome 7 abnormalities connotes poor risk.AllG-banding.FISH has been used to detect 5q– alter-other abnormalities are considered intermediate risk.21

ations in patients with a seemingly normal karyotype;These criteria are also used in the World Health Orga-its greater sensitivity is useful for detecting minimalnization (WHO) classification-based prognostic scoringresidual disease or early relapse.This technique,how-system (WPSS).27Cytogenetic risk factors,in combina-ever,detects alterations only at defined loci and requirestion with other prognostic factors such as bone mar-genetic probes specific for the region(s) of interest.row blasts and cytopenias,are used to assess overallCurrent MDS staging systems do not yet incorporateoutcome as well as risk of transformation to acuteFISH or other molecular analyses.leukemia.Thus,IPSS “low-risk”MDS patients with a

While no chromosomal abnormalities are patho-good-risk karyotype,< 5% blasts,and 0 or 1 cytopenias

gnomonic for MDS,several cytogenetic alterations havehave a median survival of 5.7 years compared with IPSSbeen consistently noted.Chromosomal deletions are“high-risk”patients who have a median survival of onlymost common,although translocations and loss or gain0.4 years.28of whole chromosomes can also occur.22In patientswith primary MDS,the most common cytogenetic alter-5q– Syndromeation consists of interstitial deletion of the long arm ofSome patients with the 5q– deletion are categorized bychromosome 5 (5q–),which is found in up to 30% ofthe WHO classification as having a distinct type of MDScases.22,23Other common changes include trisomy 8(the “5q– syndrome”),characterized by < 5% blasts in(19%) and 7q– or monosomy 7 (15%).Additional report-the bone marrow (with no Auer rods),thrombocytosis,

ed karyotypic aberrations,including loss of chromosometypical dysmegakaryopoiesis,macrocytic anemia,andY,17p–,isochromosome 17q,and interstitial deletions ofan isolated 5q– abnormality.9Such patients,who are3,11,12,13,and 20,are frequently observed in morepredominantly female,typically have a low frequency

22,23advanced disease (Table 1).Between 10% and 20% ofof progression to AML (10%) and favorable survivalpatients with primary MDS have complex karyotypes.24compared with other MDS subgroups.Cytogenetic and

In patients with treatment-related MDS,certainFISH analyses have shown that this deletion is presentcytogenetic alterations are associated with exposure in the pluripotent hematopoietic progenitor cellsto specific leukemogenic agents.Monosomy 7,7q–,(CD34+,CD38–) of MDS patients.29monosomy 5,and 5q– are the most common changesUntil recently,such patients were thought torelated to alkylating agent exposure.Alterations atrespond poorly to available therapy,responding only to11q23 are generally found following treatment withRBC transfusions.However,the immunomodulatory25

topoisomerase inhibitors.These chromosomal aber-drug lenalidomide has been found to be extremely effec-rations are associated with a poor prognosis in patientstive in the treatment of RBC transfusion-dependent,low-6Cancer ControlOctober 2008, Vol.15, No.4 Supplementto intermediate-risk MDS patients with 5q– abnormali-ties (alone or with other cytogenetic alterations).Lenalidomide therapy leads to transfusion independencein more than two-thirds of cases,with frequent cytoge-netic remissions being achieved.In a study of 148 MDSpatients treated with 10 mg of lenalidomide daily,112(76%) had a reduced need for transfusions and 99(67%) became transfusion-independent.30Of 85 evalu-able patients,62 had cytogenetic improvement,38 ofwhom had a complete cytogenetic remission.On thebasis of these results,lenalidomide was approved by theUS Food and Drug Administration for the treatment oftransfusion-dependent patients with low- to intermedi-ate-risk MDS and chromosome 5q deletion.31

While outcomes may vary among patients with dif-ferent 5q deletions,patients with the 5q– syndromegenerally have a longer survival.26One study found thatpatients with 5q– and no other karyotypic changes hada median survival of 76 months compared to 42 monthsfor MDS patients with a normal karyotype.32The pres-ence of additional (complex) chromosomal changes oran increased bone marrow blast count significantlyreduces median survival in patients with a 5q– deletion.In a study of 66 MDS patients,median survival for thosewith only 5q– was 146 months vs 45 months for patientswith additional chromosomal changes (P= .0085).33Survival was shorter still if the amount of blasts wasgreater than 5%.A recent study implicated haploinsufficiency of aspecific gene as causing some of the features of the 5q– syndrome.Using a novel RNA interference (RNAi)screening approach,Ebert et al34found that inactivationof the gene encoding the RPS14 ribosomal proteincould favor hematopoietic cell differentiation in a man-ner consistent with the thrombocytosis and anemiafound in the 5q– syndrome.34The RPS14 protein func-tions in the processing of pre-ribosomal RNA and theformation of the 40S ribosomal subunit.A similardefect in a gene involved in ribosomal processing (theRPS19gene) is seen in another bone marrow failure

syndrome,Diamond-Blackfan anemia.35These resultssuggest that haploinsufficiency of critical ribosomalgene products could account for these disease pheno-types and imply that translational control is dysregulat-ed in hematologic malignancies.

MDS Classification Systems

Staging systems are key for accurate MDS diagnosis andselection of therapy.MDS staging and classificationschemes have evolved significantly over the past fewdecades to address our evolving understanding of thebiology of this disease and its subtypes.A review of theliterature reveals that a variety of classification systemshave been used,thus making comparisons of the typesof patients enrolled in different clinical trials quite chal-lenging.The recent use of uniform classification sys-tems has remedied this problem.However,some sys-tems,developed to assess patients at diagnosis,may notbe well suited for evaluating patients following diseaseprogression.Newer and hopefully improved classifica-tion schemes may advance this situation.

French-American-British ClassificationThe French-American-British (FAB) classification is theoldest and most well-established scheme for the classi-fication of MDS and AML (Table 2).Developed in 1982,the FAB MDS classification scheme includes five sub-types of MDS,based largely on the proportion of blastsin the peripheral blood and bone marrow and the pres-ence or absence of ringed sideroblasts or increased cir-culating monocyte numbers.Median survival and timeto AML transformation vary according to FAB subtype,allowing for assigning prognosis and in some cases pre-dicting response to therapy (eg,RARS patients general-ly have a low rate of response to erythropoietin).28Itwas noted,however,that FAB subtypes are not homo-geneous within each group,suggesting limitations ofthis system.Also,the FAB scheme does not accuratelyreflect the biology of the disease,such as the numberof lineages with dysplasia.

Table 2. — French-American-British (FAB) Classification for MDS

ClassificationRefractory anemia (RA)Refractory anemia with ringed sideroblasts (RARS) Refractory anemia with excess blasts (RAEB) Refractory anemia with excess blasts in transformation (RAEB-t) Chronic myelomonocytic leukemia (CMML)CharacteristicsCytopenia of 1 PB lineage; normo- or hypercellular marrow with dysplasias; <1% blasts in PB and < 5% bone marrow blastsCytopenia and dysplasia, with same percentage blast as RA; >15% ringed sideroblasts in bone marrowCytopenia of ≥2 PB lineages; dysplasia involving all 3 lineages; <5% PB blasts and 5%–20% bone marrow blasts Same hematologic features as RAEB; >5% blasts in PB or 21%–30% blasts in bone marrow or presence of Auer rods in blastsMonocytosis in PB; <5% blasts in PB and up to 20% bone marrow blastsPB = peripheral blood. Adapted from Catenacci DV, Schiller GJ. Myelodysplastic syndromes: a comprehensive review. Blood Rev. 2005;19(6):301-319. Reprinted with permission. © 2008 Elsevier.October 2008, Vol.15, No.4 SupplementCancer Control 7

WHO ClassificationA variety of clinical and morphological issues led to thedevelopment of a new MDS staging system by the WHOto further refine the classification criteria and improvethe assignment of prognosis for intermediate-riskcases.36,37The WHO scheme is based on the FAB systembut better defines each class by use of specific criteriafor assigning dysplasia to one or more lineages in thebone marrow cells.38Refractory anemia (RA) andrefractory anemia with ringed sideroblasts (RARS) areconsidered as erythroid dysplasia with < 10% dysplasiain the myeloid or megakaryocytic lineages.Refractoryanemia with excess blasts (RAEB) is considered asRAEB-1 (5% to 9% blasts) or RAEB-2 (10% to 19% blasts).A new category,refractory cytopenia with multilineagedysplasia,was added.Chronic myelomonocytic leukemia(CMML),atypical chronic myelogenous leukemia(CML),and juvenile myelomonocytic leukemia (JMML)are classified under myelodysplastic/myeloproliferativedisorders.The 5q– syndrome is considered as a separatesubgroup under the WHO.In addition,AML was de-fined based on a lower minimum criterion for percent-age of bone marrow blasts compared to FAB (20% vs30%).The WHO system has demonstrated prognosticvalue in predicting overall and leukemia-free survival(both P< .001).39,40Some clinicians,however,do notsupport the elimination of RAEB-t from the FAB classi-fication.This stems in part from the important hetero-geneity that exists for patients with 20% to 30% blasts.By not distinguishing those with MDS from those withAML,tracking their outcome will be difficult using theWHO system.37,41-45

Prognostic Scoring Systems

Various prognostic systems havebeen proposed aimed at improvingthe ability to predict survival andprogression in MDS patients.Whilethe IPSS has been the most widelyused,the implementation in theclinical practice of the WHO classi-fication compelled a refinement ofprognostic factors and the develop-ment of new prognostic scoringsystems,such as the WPSS.

International Prognostic ScoringSystemFollowing the proposal by the FABcooperative group,several studieshave been performed to improveour ability to evaluate prognosis inMDS.In 1997 an International MDSRisk Analysis Workshop defined theIPSS.The IPSS was derived on aregression model based on 816

8Cancer Controlpatients with primary MDS from seven previous stud-ies.28The scoring system integrates bone marrow blastpercentage scored into four ranges,number of periph-eral cytopenias,and karyotype categorized in three

ggroups.An IPSS score is obtained by summing the indi-vidual scores for the three variables,resulting in strati-fication of patients into one of four risk groups:low

(score of 0),Int-1 (score of 0.5 to 1.0),Int-2 (score of1.5 to 2.0),or high (score > 2.0) (Table 3).The fourrisk groups showed significantly different overall sur-vival and risk of transformation to AML.Median sur-vival ranges from 5.7 years for patients with low risk(score of 0) to 0.4 years for those with high risk (scoreof 2.5 or greater).The IPSS has been extensively vali-dated in independent patient populations and hasbecome a benchmark for clinical trials and clinicaldecision making.

To improve prognostication in MDS,a workingparty of seven Austrian and German groups identifiedelevated lactate dehydrogenase (LDH) as an additionalprognostic variable and proposed a refinement of IPSSby including the LDH.46The IPSS+LDH was capable ofseparating a new group of “very-low-risk”patients.Also,for patients in the Int-2 and high-risk groups,the addi-tion of the LDH results in meaningful stratifications.

The limitation of the IPSS to stratify patients withlower-risk disease (ie,low or Int-1 risk) was alsoaddressed by Garcia-Manero et al.47They developed aprognostic scoring system specifically designed toidentify among patients with low-risk disease thosewith poor prognosis who may be eligible for early inter-ventions or enrollment in a clinical trial.These investi-gators analyzed 856 patients with low or Int-1 IPSS risk,and they developed a model that computes the pres-

Table 3. — International Prognostic Scoring System (IPSS) for MDS

Survival and Acute Myelogenous Leukemia Evolution ScorePrognostic variableBone marrow blasts (%)Karyotype*Cytopenias**0<5Good0 or 10.55–10Intermediate2 or 31.0—Poor—1.511–20——2.021–30——* Good = normal or any 1 of the following: –Y, del(5q), del(20q); Intermediate = other abnormalities;Poor = complex (≥3 abnormalities) or chromosome 7 abnormalities.** Cytopenias: neutrophil count <1,500/μL, platelets < 100,000/μL, hemoglobin <10 g/dL.IPSS Risk CategoryLow Int-1 Int-2 Hih Combined score Leukemic death Median time to AML (yr) Median survival (yr)* 0 19% 9.4 5.7 0.5–1.0 30% 3.3 3.5 1.5–2.0 33% 1.1 1.2≥2.5 45% 0.2 0.4 Int-1 = intermediate-1, Int-2 = intermediate-2, AML = acute myelogenous leukemia. This researchwas originally published in Greenberg P, Cox C, LeBeau MM, et al. International scoring system forevaluating prognosis in myelodysplastic syndromes. Blood. 1997;89(6):2079-2088. Reprintedwith permission. © 2008 American Society of Hematology.October 2008, Vol.15, No.4 Supplementsubgroup,and requirement for RBC transfusion as anindicator of symptomatic anemia.This model estimatesthe relationship between variables repeatedly mea-sured during follow-up and outcome and,therefore,provides dynamic prognostic information throughoutthe clinical course.Patients are assigned a score of 0 to3 according to WHO categories (Table 4).Patients withisolated erythroid dysplasia (ie,RA and RARS accordingto WHO criteria),as well as those with MDS with iso-lated del(5q),have a very good prognosis (score 0)WHO Classification-Based Prognostic compared to those with multilineage dysplasia (RCMDScoring SystemAlthough IPSS is the most commonly used prognosticand RCMD-RS,score 1).Among patients with excess of

26staging system for MDS,it does have limitations,whichbone marrow blasts,the two categories of RAEB iden-were emphasized following introduction of the WHOtify subgroups with significantly different survival andclassification.The IPSS retains significance within therisk of leukemic evolution (RAEB-1 scored 2;RAEB-2WHO subgroups.However,the two systems are redun-scored 3).Karyotype is grouped according to the IPSS

dant mainly because the IPSS blast intervals have beencriteria,but its relative weight in the model is highersubstantially maintained by the WHO classification.Incompared to the original scoring system,with good,addition,the number of peripheral cytopenias fails tointermediate,and poor cytogenetic risk groups beingretain significant value when the number of dysplasticscored from 0 to 2.The third variable in the score ismarrow lineages is included in the analysis.Therefore,transfusion requirement.Patients developing a needwhen accounting for blast percentage using WHO cate-for regular RBC transfusion have a worse prognosis

gories,the only other IPSS variable adding prognostic(score 1) compared to those without a transfusion needinformation is cytogenetics.Finally,the assimilation of(score 0).MDS with more than 20% blasts to acute leukemia result-In contrast with the four risk groups identified by

ed in a significant cutback of the higher-risk IPSS groups.the IPSS,the WPSS stratifies MDS patients into five dif-A further limitation of IPSS is that,as with otherferent risk categories:very low (score = 0),low (1),traditional prognostic scoring systems,it provides esti-intermediate (2),high (3–4),or very high (5–6) (Table

mates of risk and survival based on assessment of clin-4).Significant differences can be seen among the five

ical variables at the time of diagnosis,irrespective ofgroups in overall survival (P< .0001) and risk of AMLdisease progression.Consequently,its application during(P< .0001).In an independent cohort of patients withthe course of the disease could introduce bias.MDS,median survival ranged from 140 months for

27To address these limitations,Malcovati et aldevel-patients with a score of 0 to 10 months for those with

oped a dynamic prognostic model based on the variablesa score of 5 or more.27The most significant improve-introduced by the WHO classification-based prognosticment in prognostic stratification compared with IPSS isscoring system (WPSS).The WPSS score integrates theobserved in patients without excess blasts,primarilythree most important prognostic factors in MDS patientsdue to the strong impact of marrow lineage involve-classified according to WHO criteria:karyotype,WHOment and transfusion dependency.

The WPSS is based on a time-depen-dent regression model,and therefore itTable 4. — WHO-Based Prognostic Scoring System (WPSS) for MDS

provides dynamic prognostic informa-Scoretion throughout the clinical course.

According to this model,a patient isParameter0123classified into a risk group at the time ofWHO categoryRA, RARS, 5q–RCMD, RCMD-RSRAEB-1RAEB-2diagnosis and stays in the same group asKaryotype*GoodIntermediatePoor—long as the score remains unchanged.IfRBC transfusion requirement**NoRegular——the patient progresses,the WPSS cate-RA = refractory anemia, RARS = refractory anemia with ringed sideroblasts, RCMD = refrac-gory changes according to the resultingtory cytopenia with multilineage dysplasia, RCMD-RS = refractory cytopenia with multilineagescore,and the patient will subsequentlydysplasia and ringed sideroblasts, RAEB-1 = refractory anemia with excess of blasts-1, RAEB-2 = refractory anemia with excess of blasts-2.be followed in the new risk group.As a

* Karyotype: Good = normal, –Y, del(5q), del(20q), Poor = complex (≥3 abnormalities), chro-consequence,the interpretation of themosome 7 anomalies, Intermediate = other abnormalities.survival curves resulting from such a** RBC transfusion requirement = having ≥1 RBC transfusion every 8 weeks over a 4-monthperiod. Modified from Malcovati L, Germing U, Kuendgen A, et al. Time-dependent prognosticmodel is different from that used for scoring system for predicting survival and leukemic evolution in myelodysplastic syndromes.traditional survival curves.Non–time-J Clin Oncol. 2007;25(23):3503-3510. Reprinted with permission. © 2008 American Societydependent curves give an estimate ofof Clinical Oncology. All rights reserved.survival and risk of leukemic evolution

October 2008, Vol.15, No.4 SupplementCancer Control 9

ence of poor cytogenetics,age,hemoglobin,platelets,and percentage of marrow blasts.This led to a scoringsystem that allows patients to be classified into threeprognostic groups,with median survival ranging from80.3 months to 14.2 months.For low-risk patients,thisapproach may improve the identification of those witha more aggressive disease who may benefit from appro-priate early therapy or inclusion in clinical trials.

based on data at diagnosis,independent of any furtherevolution of the disease.In contrast,time-dependentsurvival curves provide a risk estimate based on thecurrent clinical features.In fact,a patient is followed inthe risk group assigned at diagnosis as long as the dis-ease remains stable.In the case of disease progression,the patient is re-scored according to the WPSS,and thetime-dependent survival curves for the new risk groupwill provide an updated estimate of survival and risk ofleukemic progression (Fig 1).Such an approach may behelpful for clinical decision making,particularly in low-risk patients who may be candidates for delayed treat-ment strategies.The ability to determine a risk estimateas long as patients have stable disease is of particularinterest in the very low WPSS risk group,in which mor-A1.0tality is not significantly different from that of the gen-eral population.This indicates that patients classified ashaving a very low risk,as long as their disease remainsstable,have a life expectancy not significantly differentfrom that of the general population.

The basis of WPSS is the WHO classification,whichrequires skilled morphologists since the definition ofthe WHO subgroup may be problematic in occasionalpatients due to morphologic features that are not clear-cut.In this regard,flow cytometric immunophenotyp-ing might be of help in the future.

Transfusion Requirements in MDS

The majority of MDS patients are anemic at the time ofdiagnosis,and moderate or severe anemia is observed inBCumulative Probability of Survival0.90.80.70.60.50.40.30.20.10Cumulative RiskRisk GroupLowIntermediate-1Intermediate-2High1.00.90.80.70.60.50.40.30.20.1Risk GroupLowIntermediate-1Intermediate-2High122436486072849610812013214415616818019220421622824001224364860728496108120132144156168180192204216228240Time (months)Risk GroupLowIntermediate-1Intermediate-2HighN0150178 6434N121271414218N24102104174N4862497-N723519--N962310--N12010---Risk GroupLowIntermediate-1Intermediate-2HighN0150178 6434N12124138399Time (months)N241009591N4861442-N723419--N96229--N12010---C1.0D1.00.90.80.90.8Cumulative Probability of Survival0.60.50.40.30.20.10Cumulative Risk0.7Risk GroupVery LowLowIntermediateHighVery High0.70.60.50.40.30.20.1Risk GroupVery LowLowIntermediateHighVery High122436486072849610812013214415616818019220421622824001224364860728496108120132144156168180192204216228240Time (months)Time (months)Fig 1. — Analysis of overall survival (A and C) and time to AML transformation (B and D) in a cohort of 271 MDS patients as analyzed by IPSS at diagnosis(upper panels), or by time-dependent WPSS in an independent cohort of 739 German MDS patients (lower panels). The number (N) of patients at risk at0, 12, 24, 48, 72, 96, and 120 months is shown in panels A and B. From Malcovati L, Germing U, Kuendgen A, et al. Time-dependent prognostic scoringsystem for predicting survival and leukemic evolution in myelodysplastic syndromes. J Clin Oncol.2007;25(23):3503-3510. Reprinted with permission.© 2008 American Society of Clinical Oncology. All rights reserved.

10Cancer ControlOctober 2008, Vol.15, No.4 Supplementapproximately 60% of cases.Severe anemia is usuallyassociated with fatigue,weakness,and other symptoms.Once patients develop symptomatic anemia,standardtherapy consists of supportive care measures such asRBC transfusions aimed at preserving quality of life andpreventing anemia-related morbidity.The use ofhematopoietic growth factors such as erythropoietin,with or without granulocyte colony-stimulating factor(G-CSF),can increase hemoglobin levels without theneed for transfusion in a proportion of MDS patients.48

Transfusion dependency in MDS is significantlyassociated with survival.2In a study of 467 patientswith primary MDS,Malcovati et al49demonstrated thattransfusion-dependent patients had a significantly short-er survival (hazard ratio [HR] = 2.16;P < .001) and high-er risk of leukemic progression (HR = 2.02;P < .001)compared with those not requiring transfusions.Sur-vival is inversely correlated with severity of transfusionrequirement,particularly among patients with low-riskdisease (Fig 2).49Transfusion-dependent patients have ahigher risk of non-leukemic death,which is mainly dueto an increased risk of cardiac failure (51% of cases)compared to patients without need for transfusion (P=.01).Taken together,these data suggest that the negativeeffect of transfusion dependency is mainly due to amore severe anemia and more aggressive disease.Inaddition,indirect evidence suggests that transfusion-related iron overload may also play a role.

All patients undergoing long-term RBC transfusionseventually develop iron overload.However,at presentthere is little evidence for the role of iron in organ dam-age and the impact of iron overload on the outcome ofpatients with MDS.Despite the limited evidence,the

available evidence- and consensus-based guidelines rec-ommend iron chelation in MDS patients but only forthose in whom long-term transfusion therapy is likely.Patients who are candidates for allogeneic stem celltransplantation may also benefit from chelation therapysince iron overload is associated with increased trans-plantation-related complications.

Conclusions

The recent clinical development of multiple newagents for treatment of MDS has resulted in morepatients receiving therapy that can potentially changethe course of their disease and improve survival.Appropriate selection and application of such treat-ments,which are highly dependent on correct diagno-sis and prognostic classification,are key in providingmaximal therapeutic benefit and possibly impactingsurvival.Thus,cytogenetic analysis is required for diag-nosis of patients with 5q– syndrome and subsequenttreatment with lenalidomide.26Further elucidation ofthe precise molecular defects present in various sub-types,such as the mutation affecting RPS14 in 5q– syn-drome,34will allow for more refined prognostic stagingand selection of the most appropriate therapy for indi-vidual patients.Hypomethylating agents such as azaci-tidine,which has shown a survival advantage comparedwith conventional care in Int-2 and high-risk IPSS sub-groups,may be the treatment of choice in thesegroups.50It is likely that additional novel compoundscurrently in development will prove to have greateractivity in select subpopulations as defined by theseprognostic classification schemes.

The MDS prognostic scoring systems describedhere allow physicians to select therapy that ismost appropriate based on each patient’s specif-1.0ic subclassification.So far,the IPSS — based on

0 U PRC/4WK0.9percentage of marrow blasts,cytogenetic pat-1 U PRC/4WK2 U PRC/4WKtern,and number and degree of cytopenias —0.83 U PRC/4WKhas been commonly used for predicting survival0.74 U PRC/4WKand leukemic risk.The use of the WPSS,a novel

0.6prognostic scoring system based on WHO clas-0.5sification,may be more useful than IPSS in clini-cal decision making,particularly among patients0.4without excess blasts.This ability of the WPSS0.3to stratify low-risk MDS patients is mainly due

0.2to the strong impact of single- vs multilineage

0.1dysplasia on patients and the effect of a regular

transfusion requirement as a surrogate of symp-0.0020406080100120140160180tomatic anemia.

Survival Time (months)The impressive heterogeneity of the disease

complicates clinical decision making,not only

Fig 2. — Severity of transfusion requirements predicting survival in a cohort of 426

for the choice of the most appropriate thera-Italian MDS patients who were diagnosed according to WHO criteria. Requirement of

peutic modality but also for the optimal timingRBC transfusions was calculated as the number of packed RBC units (U PRC) needed

over a 4-week period. From Malcovati L, Della Porta MG, Cazzola M. Predicting survivalof intervention.In fact,many patients with low-and leukemic evolution in patients with myelodysplastic syndrome. Haematologica.risk MDS are able to survive a long time without2006;91(12):1588-1590. Obtained from Haematologica/the Hematology Journal websitesigns of disease progression.For these patients,http://www.haematologica.org. © 2008 Ferrata Storti Foundation.

Cumulative Proportion SurvivingCancer Control 11

October 2008, Vol.15, No.4 Supplementthe risks of immediate morbidity and mortality associ-ated with transplantation are often thought to be unac-ceptably high.The use of dynamic prognostic scoringsystems,designed for patient assessment not only atdiagnosis but throughout the course of their disease,isuseful for this type of clinical decision making.

Finally,the presence of comorbid conditions maysubstantially impact the eligibility of MDS patients toundergo potentially curative therapies.Recently,ahematopoietic cell transplantation-specific comorbidi-ty index (HCT-CI) for predicting risk of nonrelapse mor-tality has been developed and validated.The HCT-CIprovides a reliable base of evidence for assessing therisks of patients who are candidates for allogeneic stemcell transplantation.51

Disclosures

Dr Nimer has served as a consultant and advisory board memberfor Celgene Corp,MGI Pharma Inc (now Eisai Inc),Pharmion Corp(now Celgene) and Genzyme Corp.

Dr Malcovati reports no significant relationship with the compa-nies/organizations whose products or services may be referencedin this article.

References

1.Kouides PA, Bennett JM.Understanding the myelodysplastic syn-dromes.Oncologist.1997;2(6):389-401.

2.Cazzola M, Malcovati L.Myelodysplastic syndromes:coping withineffective hematopoiesis.N Engl J Med.2005;352(6):536-538.

3.Catenacci DV, Schiller GJ.Myelodysplastic syndromes:a compre-hensive review.Blood Rev.2005;19(6):301-319.

4.Komrokji RS, Bennett JM.Evolving classifications of the myelodys-plastic syndromes.Curr Opin Hematol.2007;14(2):98-105.

5.Gupta V, Chun K, Yi QL, et al.Disease biology rather than age is themost important determinant of survival of patients > or = 60 years with acutemyeloid leukemia treated with uniform intensive therapy.Cancer.2005;103(10):2082-2090.

6.van Besien K, Artz A, Smith S, et al.Fludarabine, melphalan, andalemtuzumab conditioning in adults with standard-risk advanced acutemyeloid leukemia and myelodysplastic syndrome.J Clin Oncol.2005;23(24):5728-5738.7.Kantarjian H, O’Brien S, Cortes J, et al.Results of intensivechemotherapy in 998 patients age 65 years or older with acute myeloidleukemia or high-risk myelodysplastic syndrome:predictive prognostic mod-els for outcome.Cancer.2006;106(5):1090-1098.

8.Lancet JE, Gojo I, Gotlib J, et al.A phase 2 study of the farnesyl-transferase inhibitor tipifarnib in poor-risk and elderly patients with previous-ly untreated acute myelogenous leukemia.Blood.2007;109(4):1387-1394.

9.National Comprehensive Cancer Center Practice Guidelines inOncology, v.2.2008.Myelodysplastic syndromes.http://www.nccn.org.Accessed May 7, 2008.

10.Heaney ML, Golde DW.Myelodysplasia.N Engl J Med.1999;340(21):1649-1660.11.Chatterjee T, Dixit A, Mohapatra M, et al.Clinical, haematologicaland histomorphological profile of adult myelodysplastic syndrome:study of96 cases in a single institute.Eur J Haematol.2004;73(2):93-97.

12.Kaloutsi V, Kohlmeyer U, Maschek H, et al.Comparison of bone mar-row and hematologic findings in patients with human immunodeficiencyvirus infection and those with myelodysplastic syndromes and infectious dis-eases.Am J Clin Pathol.1994;101(2):123-129.

13.Rothstein G.Disordered hematopoiesis and myelodysplasia in theelderly.J Am Geriatr Soc.2003;51(3 suppl):S22-S26.

14.Steensma DP, Hanson CA, Letendre L, et al.Myelodysplasia withfibrosis:a distinct entity? Leuk Res.2001;25(10):829-838.

15.Elghetany MT.Diagnostic utility of flow cytometric immunophenotyp-ing in myelodysplastic syndrome.Blood.2002;99(1):391-392.

16.Cherian S, Moore J, Bantly A, et al.Peripheral blood MDS score:anew flow cytometric tool for the diagnosis of myelodysplastic syndromes.Cytometry B Clin Cytom.2005;64(1):9-17.

17.Ogata K, Kishikawa Y, Satoh C, et al.Diagnostic application of flowcytometric characteristics of CD34+ cells in low-grade myelodysplastic syn-dromes.Blood.2006;108(3):1037-1044.12Cancer Control18.Malcovati L, Della Porta MG, Lunghi M, et al.Flow cytometry evalu-ation of erythroid and myeloid dysplasia in patients with myelodysplasticsyndrome.Leukemia.2005;19(5):776-783.

19.Vardiman JW.Myelodysplastic syndromes, chronic myeloproliferativediseases, and myelodysplastic/myeloproliferative diseases.Semin DiagnPathol.2003;20(3):154-179.

20.Mufti GJ.Chromosomal deletions in the myelodysplastic syndrome.Leuk Res.1992;16(1):35-41.

21.Rigolin GM, Bigoni R, Milani R, et al.Clinical importance of inter-phase cytogenetics detecting occult chromosome lesions in myelodysplas-tic syndromes with normal karyotype.Leukemia.2001;15(12):1841-1847.22.Fenaux P.Chromosome and molecular abnormalities in myelodys-plastic syndromes.Int J Hematol.2001;73(4):429-437.

23.Mhawech P, Saleem A.Myelodysplastic syndrome:review of the cyto-genetic and molecular data.Crit Rev Oncol Hematol.2001;40(3):229-238.24.Bennett JM, Komrokji R, Kouides PA.The myelodysplastic syn-dromes.In:Abeloff MD, Armitage JO, Niederhuber JE, et al, eds.ClinicalOncology.3rd ed.Philadelphia, Pa:Elsevier;2004:2849-2882.

25.Shali W, Hélias C, Fohrer C, et al.Cytogenetic studies of a series of43 consecutive secondary myelodysplastic syndromes/acute myeloidleukemias:conventional cytogenetics, FISH, and multiplex FISH.CancerGenet Cytogenet.2006;168(2):133-145.

26.Nimer SD.Clinical management of myelodysplastic syndromes withinterstitial deletion of chromosome 5q.J Clin Oncol.2006;24(16):2576-2582.27.Malcovati L, Germing U, Kuendgen A, et al.Time-dependent prog-nostic scoring system for predicting survival and leukemic evolution inmyelodysplastic syndromes.J Clin Oncol.2007;25(23):3503-3510.

28.Greenberg P, Cox C, LeBeau MM, et al.International scoring systemfor evaluating prognosis in myelodysplastic syndromes.Blood.1997;89(6):2079-2088.

29.Nilsson L, Astrand-Grundström I, Arvidsson I, et al.Isolation andcharacterization of hematopoietic progenitor/stem cells in 5q-deletedmyelodysplastic syndromes:evidence for involvement at the hematopoieticstem cell level.Blood.2000;96(6):2012-2021.

30.List A, Dewald G, Bennett J, et al.Lenalidomide in the myelodysplas-tic syndrome with chromosome 5q deletion.N Engl J Med.2006;355(14):1456-1465.31.Melchert M, Kale V, List A.The role of lenalidomide in the treatmentof patients with chromosome 5q deletion and other myelodysplastic syn-dromes.Curr Opin Hematol.2007;14(2):123-129.

32.Knipp S, Hildebrandt B, Jotterand M, et al.Dependency of survivalon additional karyotype anomalies and bone marrow blast count in patientssuffering from myelodysplastic syndromes with 5q deletion.Blood.2007:110(11).Abstract 2459.

33.Giagounidis AA, Germing U, Haase S, et al.Clinical, morphological,cytogenetic, and prognostic features of patients with myelodysplastic syn-dromes and del(5q) including band q31.Leukemia.2004;18(1):113-119.34.Ebert BL, Pretz J, Bosco J, et al.Identification of RPS14 as a 5q- syn-drome gene by RNA interference screen.Nature.2008;451(7176):335-339.35.Flygare J, Kiefer T, Miyake K, et al.Deficiency of ribosomal proteinS19 in CD34+ cells generated by siRNA blocks erythroid development andmimics defects seen in Diamond-Blackfan anemia.Blood.2005;105(12):4627-4634.

36.Bennett JM.World Health Organization classification of the acuteleukemias and myelodysplastic syndrome.Int J Hematol.2000;72(2):131-133.37.Vardiman JW, Harris NL, Brunning RD.The World Health Organiza-tion (WHO) classification of the myeloid neoplasms.Blood.2002;100(7):2292-2302.

38.Harris NL, Jaffe ES, Diebold J, et al.World Health Organization clas-sification of neoplastic diseases of the hematopoietic and lymphoid tissues:report of the Clinical Advisory Committee meeting, Airlie House, Virginia,November 1997.J Clin Oncol.1999;17(12):3835-3849.

39.Malcovati L, Porta MG, Pascutto C, et al.Prognostic factors and lifeexpectancy in myelodysplastic syndromes classified according to WHO criteria:a basis for clinical decision making.J Clin Oncol.2005;23(30):7594-7603.

40.Müller-Berndorff H, Haas PS, Kunzmann R, et al.Comparison of fiveprognostic scoring systems, the French-American-British (FAB) and WorldHealth Organization (WHO) classifications in patients with myelodysplasticsyndromes:results of a single-center analysis.Ann Hematol.2006;85(8):502-513.

41.Heaney ML, Golde DW.Critical evaluation of the World Health Orga-nization classification of myelodysplasia and acute myeloid leukemia.CurrOncol Rep.2000;2(2):140-143.

42.Head DR.Proposed changes in the definitions of acute myeloidleukemia and myelodysplastic syndrome:are they helpful? Curr Opin Oncol.2002;14(1):19-23.

43.Greenberg P, Anderson J, de Witte T, et al.Problematic WHO reclas-sification of myelodysplastic syndromes.Members of the International MDSstudy group.J Clin Oncol.2000;18(19):3447-3452.

44.Germing U, Gattermann N, Strupp C, et al.Validation of the WHOproposals for a new classification of primary myelodysplastic syndromes:aretrospective analysis of 1600 patients.Leuk Res.2000;24(12):983-992.

October 2008, Vol.15, No.4 Supplement45.Strupp C, Gattermann N, Giagounidis A, et al.Refractory anemiawith excess of blasts in transformation:analysis of reclassification accord-ing to the WHO proposals.Leuk Res.2003;27(5):397-404.

46.Germing U, Hildebrandt B, Pfeilstöcker M, et al.Refinement of theinternational prognostic scoring system (IPSS) by including LDH as an addi-tional prognostic variable to improve risk assessment in patients with prima-ry myelodysplastic syndromes (MDS).Leukemia.2005;19(12):2223-2231.47.Garcia-Manero G, Shan J, Faderl S, et al.A prognostic score forpatients with lower risk myelodysplastic syndrome.Leukemia.2008;22(3):538-543.

48.Hellström-Lindberg E, Malcovati L.Supportive care and use of hema-topoietic growth factors in myelodysplastic syndromes.Semin Hematol.2008;45(1):14-22.

49.Malcovati L, Della Porta MG, Cazzola M.Predicting survival andleukemic evolution in patients with myelodysplastic syndrome.Haemato-logica.2006;91(12):1588-1590.

50.Issa JP.Optimizing therapy with methylation inhibitors in myelodys-plastic syndromes:dose, duration, and patient selection.Nat Clin PractOncol.2005;2(suppl 1):S24-S29.

51.Sorror ML, Maris MB, Storb R, et al.Hematopoietic cell transplanta-tion (HCT)-specific comorbidity index:a new tool for risk assessment beforeallogeneic HCT.Blood.2005;106(8):2912-2919.

October 2008, Vol.15, No.4 SupplementCancer Control 13