(CANCER RESEARCH 46, 3698-3700, July 1986]

Factors Influencing Leukemic Transformation in Refractory Anemias with Excessof Blasts, with Ringed Sideroblasts, and without Ringed Sideroblasts1

Shigeru Oguma,2 YátaroYoshida, Haruto l'chino, and Tadashi Maekawa

The First Division, Department of Internal Medicine, Faculty of Medicine, Kyoto University, Kyoto 606 fS. O., Y. Y., H. U.J, and The Third Division, Department ofInternal Medicine, Faculty of Medicine, Gumma University, Maebashi 371 [T. M.], Japan

ABSTRACT

The association between leukemic transformation and various featuresrecorded at presentation in patients with refractory anemia with excessof blasts and with or without ringed Sideroblasts was analyzed in 255patients using the proportional hazard model. Features associated withhigher transformation rates were: higher values of blasts in peripheralblood or bone marrow; serum haptoglobin; vitamin BI2; megakaryocytesin bone marrow; morphological abnormalities in granulo- or megakary-ocyte series; male sex; circulating megakaryocytes in peripheral blood;older age; and lower ringed sideroblast proportion. Multivariate analysiswas also performed using the following predictor variables: presence orabsence of refractory anemia with excess of blasts; sex; abnormal granules in granulocytes; age; and mononuclear large megakaryocytes. Patients were divided arbitrarily into low (hazard ratio, <0.45), intermediate(hazard ratio, 0.45-1.85) and high (hazard ratio, > 1.85) risk groups. Thecumulative leukemia-free rates in the low and intermediate risk groupsshowed long plateau phases at 95 and 71%, respectively, while in thehigh risk group, the rate was 10% at 5 years. For clinical purposes, thelow risk group should be considered to have nonpreleukemia and thehigh risk group to have preleukemia.

INTRODUCTION

RAEB,3 RARS, and RA are subgroups of myelodysplasticsyndromes proposed by the French-American-British Cooperative Group ( 1). These refractory anemias have many commonfeatures such as anemia, neutropenia, thrombocytopenia, andmorphological abnormalities of the erythroid, granulocytic, andmegakaryocytic series. Transformation into ANLL occurs atvarying rates in RAEB, RARS, and RA. This raises a questionof whether some proportion of patients are destined to developleukemia, or alternately whether the leukemic transformationis a stochastic process where subgroups affect the rates ofleukemic transformation. A definite answer cannot be given atpresent, but it is a major clinical concern whether a particularpatient will develop leukemia or not. Thus, we used biostatist-ical methods to clarify the risk factors for leukemic transformation and to divide patients into nonpreleukemic and preleu-kemic subgroups.

Patients with secondary refractory anemia or possible congenital hematological disorders were excluded to avoid possiblebias.

Virtually all cases of RAEB in transformation develop intoANLL, and a considerable proportion received antileukemictherapy; therefore, they were not included. Most of the patientswith chronic myelomonocytic leukemia have increased bone

Received 10/17/85; revised 2/27/86: accepted 3/17/86.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1This work was supported in part by Grants from the Ministry of Health and

Welfare and the Ministry of Education, Japan.2To whom requests for reprints should be addressed, at The First Division,

Department of Internal Medicine, Faculty of Medicine, Kyoto University, 54-Shogoin-Kawaramachi, Sakyo-ku. Kyoto 606, Japan.

3The abbreviations used are: RAEB, refractory anemia with excess of blasts;

RARS, refractory anemia with ringed Sideroblasts; RA, refractory anemia withoutringed Sideroblasts; ANLL, acute nonlymphoblastic leukemia: PB, peripheralblood; BM, bone marrow.

marrow culture findings, results quite different from those inrefractory anemias. This suggests that it is better to distinguishchronic myelomonocytic leukemia from refractory anemias, atleast when analyzing risk factors for leukemic transformation.

MATERIALS AND METHODS

Patients. The Japanese Refractory Anemia Study Group performeda nationwide retrospective study of potential cases of RAEB, RARS,and RA in 1984. All patients suspected of having refractory anemiaswere entered, after which the subclassification or exclusion was done.A total of 366 cases were entered. Clinical and hematological findingsof these cases will be reported separately.

Subclassification and exclusion were done using previously reportedcriteria (2). RAEB and transformation into ANLL were diagnosisaccording to the FAB proposals, and RARS according to the proposalof Sultan et al. (3), i.e., 20% or more ringed Sideroblasts per totalerythroblasts. Those not meeting the criteria for RAEB or RARS butshowing dysplastic changes of the blood cells and/or hyper- or normo-cellular marrow were diagnosed as RA. Patients meeting the criteriafor both RAEB and RARS were classified as RAEB. Those youngerthan 10 years were excluded.

Based on these considerations, 104 were diagnosed as RAEB, 49 asRARS, and 111 as RA.

Statistical Methods. Our data included many variables of the ordinalor nominal scales, and almost all variables of interval scale did not fîtthe hypothesis of normal distribution. Thus, we used the followingnonparametric methods for all of the statistical analyses: (a) the Kaplan-Meier estimate (4) for cumulative leukemia-free rate and Kalbfleischand Prentice modification of Greenwood's formula (5) for calculatingthe confidence interval of the curves; (b) Cox's proportional hazard

model (6) for estimate of prognostic significance for leukemic transformation. For calculation of the likelihood of tied observations, Breslow's

asymptotic method (7) was used.All calculations were done using computers at Kyoto University Data

Processing Center and Educational Center for Information Processing.Factors Analyzed. Nineteen categoric and 49 metric features at

presentation were recorded when available: age; sex; serum levels oflactic dehydrogenase; total and indirect bilirubin; vitamin B,:; muram-idase; haptoglobin; urine muramidase; ferrokinetic findings; hematological findings of PB and BM; cytochemical assays; morphologicalabnormalities of the erythroid, granulocytic, and megakaryocytic series;karyotypic abnormalities; and in vitro bone marrow culture assays.

In this analysis, failure time was the interval from presentation toleukemic transformation. Leukemia-free survivors and those who diedwithout leukemic transformation were treated as censored observations.

RESULTS

Of the 264 patients in this series, 9 were excluded becausethe date of leukemic transformation or censoring was notknown. The remaining 255 patients ranged in age from 15 to85 years (median, 59 years). There were 167 men and 88women. The age and sex distributions in terms of RAEB,RARS, and RA are listed in Table 1. Leukemic transformationoccurred in 68 patients: 49 RAEB; 4 RARS; and 15 RA.

Cumulative Leukemia-free Rates. Fig. 1 shows the cumulativeleukemia-free rates in RAEB, RARS, and RA. All transformations occurred within 1810 days from presentation. Leukemia-

3698

on April 21, 2019. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

LEUKEMIC TRANSFORMATION IN REFRACTORY ANEMIAS

Table 1 Age and sex distribution

% of patients at following agesMale/female

I0-29yr 30-49 yr 50-69 yr 70> yr ratio

RAEBRARSRA48142314244843472535152.71.61.5

Table 2 Risk factors for leukemic transformationVariables are listed in the order of x2 value up to x2 > 5.02 (P < 0.025 with

d.f. = 1).

Myeloblasts in BM(%)Blastsin PB(abs.°)Blasts

in PB(%)Abnormalgranules ingranulo-cyte

seriesEosinophilsin PB(abs.)Serum

haptoglobinSexPromonocytes

in BM(%)Eosinophilsin PB(%)Megakaryocytes

in BM(abs.)Circulatingmegakaryocyte inPBAbnormal

nucleus ingranulocyteseriesGiantplateletsVitaminU,..Mononuclear

largemegakary-ocytesAgeRinged

sideroblasts*No.

ofcases25525525524625571255237255143231246239180232254214x261.624.321.319.715.412.911.88.47.17.06.66.46.16.05.85.45.4Signs

associatedwith higherriskHigh

valueHighvalueHighvaluePositiveHigh

valueHighvalueMaleHigh

valueHighvalueHighvaluePositivePositivePositiveHigh

valuePositiveOld

ageLowvalue

" abs., absolute count.* Proportion to total erythroblasts.

RflRS

RflEB

?5 >ie 12 96 120MONTHS

Fig. 1. Cumulative leukemia-free rates in RAEB, RARS, and RA.

free rates and their 95% confidence intervals were: 28.9% (14.2-45.4) in RAEB; 89.6% (73.6-96.1) in RARS; and 74.4% (59.2-84.3) in RA.

Our criteria for RA did not exclude patients with hypoplasticmarrow if dysplastic changes were apparent. There are ninesuch patients in this series. Their leukemia-free rate for thesame time period was 77.8%, essentially the same rate as seenfor all RA patients.

Analysis of Individual Factors. The association between leukemic transformation and the various risk factors was examinedby the proportional hazard model (Table 2). Because of multiplecomparisons in this study, we set the significance level slightlymore conservatively (P < 0.025) than the conventional level (P< 0.05), to lessen type I errors.

Multivariate Analysis. Multivariate life table analysis wasperformed on the 202 patients on whom data including sex,age, hemoglobin, leukocyte counts, blasts in PB and BM, andmorphological abnormalities were all available at the time ofpresentation. These factors and the subclassifications of RAEB,RARS, and RA were the items analyzed. Some variables thathad high significance for leukemic transformation (e.g., hapto-

globin) were not included in this analysis because if they had

been included, the number of available observations would havebeen markedly diminished. The actual process was progressive;i.e., the number of variables put into the model was increaseduntil no significant increase of maximum likelihood was obtained. Five variables, i.e., the presence or absence of RAEB,sex, abnormal granules in granulocytes, age, and mononuclearlarge megakaryocytes, were selected. The resulting formula tocalculate expected relative hazard is given in Table 3. Thesignificance level of this analysis was that the global null hypothesis "all of the regression coefficients = 0" was abandonedat the level of P < 0.5 x 10~7.

According to the hazard ratio calculated from the formula inTable 3, patients were divided into three groups: high risk(hazard ratio, >1.85); intermediate risk (hazard ratio, 0.45-1.85); and low risk (hazard ratio, <0.45) groups. The low andintermediate risk groups showed long plateau phases of leukemia-free rates (Fig. 2). The interval from presentation to thesephases and cumulative leukemia-free rates at these phases were322 days and 95.4% for low risk groups and 1011 days and70.9% for intermediate risk groups, respectively. On the otherhand, there was no plateau phase in the high risk group; themedian leukemia-free duration of the group was 351 days.

Of the five variables in the formula (Table 3), four aredichotomized data. Thus, there are 16 possible combinationsof these 4 variables. Table 4 shows our clinical staging systemfor each of the combinations.

DISCUSSION

RAEB, RARS, and RA are all considered as preleukemicdisorders. However, the rate of leukemic transformation differs

Table 3 Formula to calculate hazard ratio

whereHazard ratio = exp(z,$i + Z2j32+ z3/33+ z4ßt

Diagnosis" - 0.36Z| =

and

0.48Sex" - 1.36

048Abnormal granules in granulocytesc —0.5

05Age (yr) - 56.2

16.7Mononuclear large megakaryocytes^ - 0.43

049

0, = 0.87 ±0.14 (S Error of ft)

ft = -0.42 ±0.16

ß,= 0.36 ±0.15

ß,= 0.40 ±0.15

ß,= 0.29 ±0.14°1 for RAEB and 0 for RARS or RA.* 1 for male and 2 for female.c 0 for negative and 1 for positive.

LOW RISK

96 IÕOMONTHS

Fig. 2. Cumulative leukemia-free rates in high, intermediate, and low riskgroups.

3699

on April 21, 2019. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

LEUKEMIC TRANSFORMATION IN REFRACTORY ANEMIAS

Table 4 Clinical staging system for predicting leukemic transformationPatients younger than the lower age limit of intermediate risk are low risk,

and those older than the upper age limit are high risk.

Abnormalgranules in

Diagnosis SexgranulocytesRAEB

MMM

+M+FFF

+F+RARS,

RA MMM

+M+FFF

+F+Mononuclear

Range of agelarge megakar- for intermediate

yocytes risk(yr)28-52+

18-4216-39+

"-3043-66+

33-5631-54+

21-4458-82+

48-7246-70+

36-6073-»+

63-»61-»+

51-75" No lower age limit for intermediate risk."'No upper age limit for intermediate risk.

dramatically among them. Furthermore, a considerable proportion (28.9%) of the patients with RAEB in our series remainednonleukemic and reached a plateau phase after about 60 monthsfrom presentation.

Our major concern was to what extent leukemic transformation could be predicted from various features at presentation,and whether preleukemic in a prospective sense and nonleukemic subgroups could be identified at the time of presentationor not.

Several features had already been suggested as factors associated with a more rapid development of ANLL (8): presenceand severity of dysplasia in more than one cell line (1, 9);increased number of blast cells ( 1, 9); abnormal clustering ofimmature myeloid precursors centrally in the marrow (10); andclonal chromosomal abnormality or a "leukemic" pattern of

growth in marrow cultures in vitro or in diffusion chamberexperiments (9, 11). Alternatively, some patients with isolated5q—syndromes with persistence of a normal cell line have been

followed for more than 6 years without leukemic transformation(12).

In our series, several additional features were identified asrisk factors: higher eosinophil counts in PB; higher serumhaptoglobin levels; male sex; higher megakaryocyte counts inBM; circulating megakaryocytes in PB; higher vitamin B|2levels; older age; and lower ringed sideroblast proportion. Theassociation with sex might be explained in part by the findingthat the relative proportion of males in the RAEB group wasgreater than those in the RARS or RA groups (Table 1). Theclinical significance of vitamin Bn is doubtful because of thepossibility that the agent had been prescribed elsewhere for"unexplained anemia" before the diagnosis of refractory anemia

was made. The significance of ringed sideroblasts is probablyderived from the finding that patients with RAEB had both a

higher leukemic transformation rate and lower ringed sideroblasts than those with RARS. Higher haptoglobin, circulatingmegakaryocytes in PB, and higher megakaryocyte counts inBM might reflect the malignant nature of the disease.

By multivariate analysis, the patients could be divided intothree distinctly different groups with respect to the leukemictransformation rate (Fig. 2). The low and high risk groupscontained considerable numbers of patients, 58 (28.7%) and 70(34.6%), respectively. For clinical purposes, the low risk groupwas considered to be nonleukemic and the high risk group wasconsidered to have preleukemia in a prospective sense. Thus,in the high risk group, an initial treatment approach withassumed greater degree of risk could be justified; while in thelow risk group, conventional intensive anticancer therapy is notrecommended.

We hope our formula and the clinical staging system forleukemic transformation will be useful in clinical trials as astaging method.

ACKNOWLEDGMENTS

This study was made possible by the participation of the followinghematologists: Dr. Ornine of Gumma University, Maebashi, Japan,who provided 30 patients; Dr. Ichirnaru of Nagasaki University, Nagasaki, Japan, who provided 29 patients; Dr. Yamada of NagoyaUniversity, Nagoya, Japan, who provided 27 patients; Dr. Nomura ofNihon Medical College, Tokyo, Japan, who provided 20 patients; andalso Drs. Matsumoto, Mizoguchi, Kamada, Miura, Yasunaga, Amano,Hirashima, Yawata, Sato, Toyama, Nagai, Takaku, Shirakawa, Abe,Miwa, Mikuni, Matsutani, Shibata, Kawakita, Yasuda, and Fujioka.

REFERENCES

1. Bennet, J. M., Catovsky, D., Daniel, M. T., Flandrin, G., Galton, D. A. G.,Gralnick, H. R., and Sultan, C. Proposals for the classification of themyelodysplastic syndromes. Br. J. Haematol., 51: 189-199, 1982.

2. Uchino, H., Yoshida, Y., Yamagishi, M., et al. Proposals of the diagnosticcriteria for refractory anemias. Jpn. J. Clin. Hematol., 27: 36-42, 1986.

3. Sultan, C. Dysmyelopoietic syndromes. Ann. Intern. Med., 87: 740-753,1977.

4. Cox, D. R., and Oakes, D. Analysis of Survival Data, pp. 48-61. London:Chapman and Hall, 1984.

5. Kalbfleisch, J. D., and Prentice, R. L. The Statistical Analysis of FailureTime Data, pp. 1-20. New York: John Wiley and Sons, Inc., 1980.

6. Cox, D. R. Regression models and life-tables. J. R. Statistics Soc. B, 34:187-202, 1972.

7. Breslow, N. Covariance analysis of censored survival data. Biometrics, 30:89-99, 1974.

8. Larson, R. A. Management of myelodysplastic syndromes. Ann. Intern.Med., 103: 136-138, 1985.

9. Jacobs, R. H., Cornbleet, M. A., Vardian, J. W., and Rowley, J. D. Correlation between morphologic subtype, clinical characteristics, and cytogeneticabnormalities in patients with de novo dysmyelopoietic syndromes. Blood,64(Suppl. 1): 147a, 1984.

10. Tricot, G., De Wolf-Peeters, C., Vlietinck, R., and Verwilghen, R. L. Bonemarrow histology in myelodysplastic syndromes: II. Prognostic value ofabnormal localization of immature precursors in MDS. Br. J. Haematol.,58: 217-225, 1984.

11. Gold, E. J., Conjalka, M., Pelus, L. M., et al. Marrow cytogenetic and cellculture analysis of the myelodysplastic syndromes: insights into pathophys-iology and prognosis. J. Clin. Oncol., /: 627-634, 1983.

12. Kerkhofs, H., Hagemeijer, A., Leeksma, C. H. W., et al. The 5q- chromosome abnormality in haematological disorders: a collective study of 34 casesfrom the Netherlands. Br. J. Haematol., 52: 365-381, 1982.

3700

on April 21, 2019. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1986;46:3698-3700. Cancer Res   Shigeru Oguma, Yataro Yoshida, Haruto Uchino, et al.   without Ringed SideroblastsAnemias with Excess of Blasts, with Ringed Sideroblasts, and Factors Influencing Leukemic Transformation in Refractory

  Updated version

  http://cancerres.aacrjournals.org/content/46/7/3698

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/46/7/3698To request permission to re-use all or part of this article, use this link

on April 21, 2019. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from