British Journal of Haematology, 1990. 75, 26-33

Bone marrow biopsy in myelodysplastic syndromes: morphological characteristics and contribution to the study of prognostic factors A. Rios, M. C. C A R I Z O , M. A. S A N Z , * T. V A L L E S P i , t G . SANZ,*M. T O R R A B A D E L LF.A ,GOMIS,* ~ C. R u I z t AND J . F. SANMIGUEL Hospital Clinico, Salamanca, *Hospital ‘La Fe’, Valencia, and ?Hospital Valle de Hebron, Barcelona, Spain Received 12 July 1989; accepted for publication 8 January 1990

Summary. Ten characteristics of bone marrow (BM) biopsies in paraffin sections, obtained at diagnosis from patients with myelodysplastic syndromes (MDS) classified according to the FAB criteria, were analysed to identify both the most relevant morphologic data and any possible influence on survival. Agreement between two observers was obtained for 94% of the data. BM cellularity was increased in 63%of the cases and was higher in refractory anaemia with excess of blasts (RAEB), RAEB in transformation (RAEB-t) and chronic myelomonocytic leukaemia (CMML) (P=0-001). Dysmegakaryopoiesis and dyserythropoiesis were present respectively in 83% and 72% of the cases, with slight differences among the FAB subtypes. Abnormal localization of immature precursors (ALIP)was found in more than half of the cases and somewhat more frequently seen in the RAEB RAEBt +CMML group (P= 0.07). Eosinophilia. plasmacytosis and reticulin fibrosis were evident in 26%, 18% and 47% of the

cases respectively. Cellularity ( P = 0.006), eosinophilia (P= 0.009) and, to some extent, dysmegakaryopoiesis (P=0.07) bore a certain relationship with survival on univariate analysis. The presence of ALIP was not seen to affect the outcome. Multivariate analysis showed that the cellularity and presence of dysmegakaryopoiesis, in BM biopsy, added significant independent prognostic information to that achieved with age, platelet count and proportion of blast cells in BM aspirate, three variables with proven prognostic value in MDS patients. Using a regression model including these five characteristics we have stratified the patients into low, intermediate and high-risk groups with different survivals (P= 0.00001). The present findings show that BM biopsy is able to provide both morphological characteristics and information about the prognosis of survival, and should thus be included in the initial evaluation of MDS.

The diagnosis of myelodysplastic syndromes (MDS) is mainly based on morphological studies of the cells obtained from peripheral blood and bone marrow (BM) aspirates. In 1982 the FAB (French-American-British) Cooperative Group, using precise quantitative and qualitative criteria, established five subtypes (Bennett et al, 1982) whose prognostic value has been demonstrated (Garcia et al, 1988: Sanz et al, 1989). BM biopsy has proved to be of value not only for diagnosis but also in predicting the outcome of patients with several haematological disorders such as chronic lymphocytic leukaemia (Rozman et al, 1984). aplastic anaemia (Rozman et al, 1985 ) , myeloproliferative disorders (Lazzarino et al, 1986; Ellis et al, 1986) and non-Hodgkin lymphoma (Bart1 et al.

1982). In 1972 Duhamel pointed out the possible prognostic value of BM biopsy in patients with so-called ‘refractory anaemia’. Since then, reports in MDS have been scanty (Tricot et al, 1984; Martin et al, 1986). The present work was designed to analyse the morphological characteristics of BM biopsy in 126 patients with MDS. In addition it also aimed at exploring, by using both univariate and multivariate methodology, the possible value of the histopathological findings.

+

MATERIAL A N D METHODS

Patients. From 1971 to 1987 a total of 370 consecutive patients with MDS, diagnosed according to FAB criteria, were studied in three different hospitals in Spain. BM biopsy, at the time of diagnosis, was available for evaluation in 165 patients. Thirty-nine were considered technically unsatisfactory, so the final number of biopsies studied was 126.

Correspondence: Dr A. Rios. Servicio de Hematologia, Hospital Clinic0 Universitario. Paseo de San Vicente 108, 3 7007 Salamanca, Spain.

26

Bone Marrow Biopsy in MDS

27

Table I. MDS: histological data analysed on bone marrow biopsy Degree

Total cellularity (%)* Quantitative megakaryopoiesis* Quantitative erythropoiesis'

3

4

0

1

240

41-60

61-80

>SO

abs.

dec.

abs.

dec.

nor. nor.

inc. inc.

m. inc. rn. inc.

2

Dysmegakaryopoiesis The presence of at least two of following: High nuclear/cytoplasmatic ratio Immature nucleus with nucleoli Monolobated nuclei Micromegakaryocytes Dyserythropoiesis The presence of at least two of following: Megaloblastic maturation Defective maturation: some stage of differentiation Predominance of proerythroblasts Abnormal localization of immature precursors (ALIP) Presence of at least three clusters in each section of three to five myeloid precursors (myeloblasts and promyelocytes) clustering centrally Marked eosinophdia Presence of eight or more eosinophils/microscopicfield at medium-power ( x 400) Reticulin fibres Presence of diffuse fibre network Lymphatic follicles Presence of aggregates of small lymphocytes Marked plasmacytosis Presence of two or more plasma celb/microscopic field at medium-power ( x 400)

Abbreviations:abs. =absent: dec =decreased: nor. =normal: inc. =increased m. inc. =markedly increased. * Subjective evaluation by eye. According to FAB criteria, 38 patients (30%) had refractory anaemia (KA), 1 6 (13%) refractory anaemia with ring sideroblasts (RARS), 40 (32%) refractory anaemia with excess of blasts (RAEB), 10 (8%) RAEB in transformation (KAEB-t) and 22 cases (17%) chronic myelomonocytic leukaemia (CMML). B M biopsy. Biopsies were obtained from anterior or posterior iliac crest with a Jamshidi needle. Samples were fixed, embedded in paraffin and stained with haematoxylineosin and the Wilder technique and, in some cases, with Giemsa, periodic acid-Schiff (PAS) and the Masson technique (Hernandez-Nieto & Rozman, 1980). All samples were examined independently by two observers, who were unaware of the patients' identification and FAB subtype. Discordant data were re-examined by each observer, and if disparity persisted examination was performed jointly. When agreement was not reached the data were excluded from later analysis. Characteristics recorded from each biopsy sample included total cellularity, quantitative megakaryopoiesis

and erythropoiesis, dysmegakaryo- and dyserythropoiesis features, abnormal localization of immature progenitor cells (ALIP) and presence of eosinophilia, reticular fibrosis, lymphoid follicles and plasmacytosis. The criteria for evaluation and grading of these characteristics are shown in Table I. The total data anaiysed in 126 biopsies was 1207, because reticulin could not be evaluated in 53 cases. Examples of BM sections are shown in Figs 1-4. Statistical methods. The Chi-square test, or the Fisher exact test when appropriate, were used for comparison of proportions. Survival curves were plotted according to the actuarial method of Kaplan & Meier (1958). Confidence intervals for the median survival time were calculated by the BrookmeyerCrowley method (1982). Statistical comparisons between different actuarial curves were based on the long-rank test or, if applicable. the test for trend (Peto et nl. 1977). Further multivariate analysis by means of the proportional hazards regression method was used in order to identify the most significant independent prognostic factors (Cox, 1972). The BM biopsy data selected for possible inclusion in the regres-

28

A. Rios et a1

Fig 1 . BM section showing megakaryocytes with immature monolobated nucleus. (H&E x 400.)

Fig 2. BM section showing megaloblastic erythropoiesis. (H&E x 400.)

sion method were those for which there was some indication of a significant association with survival in univariate analysis (P0.05). Eosinophilia was found in 26% of the biopsies, with no appreciable differences in FAB subtypes, except in RAEB-t in which it was not observed in any patient. Reticulin was increased in 34 out of the 73 biopsies studied (43%),ranging from 63% in the CMML to 17% in the RAEB-t (P>O.l). Lymphoid follicles were found in 10% with a similar distribution in the FAB subtypes. An appreciable degree of plasmacytosis was observed in 23 cases (IS%), being specially frequent among RAEB-t patieents (50%) ( P = 0.02).

' j \i,

.Y

p = 0.01

L 0

li

24

36

48

60

72

84

MONTHS Fig 6. Survival curves according to total cellularity.

t p = 0,001

c

!

I

01

0

I

I

I

I

I

I

I

I?

24

36

48

60

72

84

MONTHS Fig 7. Survival according to presence (+) or absence ( - ) dysrnegakaryopoiesis.

of

Prognostic factors analysis Univariate analysis showed that only two of the characteristics studied, BM cellularity and eosinophilia, were related significantly to survival (P= 0.006 and P=0.009, respectively). The presence of eosinophilia and decreased cellularity were associated with a more favourable outcome. Of the remaining variables only dysmegakarypoiesis also seemed to show a trend suggesting an unfavourable relationship with survival length (P=047). The best combination of patient and disease characteristics selected by means of the Cox proportional hazard regression method in the overall group of patients is given in Table 111. The variables are listed in the order entered by the forward stepwise modelling procedure. Besides the platelet count, the percentage of blast cells in BM aspiration, and age, the current analysis revealed the importance of the increased cellularity (P=0.07) and dysmegakaryopoiesis (P=O.OOl), indicative of a poorer prognosis (Figs 6 and 7). The remaining

32

A. Rim et a1

data obtained from BM biopsy did not provide significant prognostic information. Using cut-off points of 0.7 and 1.4 in the hazard function defined by the regression model, the series of 126 patients was divided into low-. intermediate- and high-risk groups. Actuarial median survial for the three risk groups was 43.2 months (95% confidence limits 26 and 60.4), 1 5 months (7.5 and 24) and 7.1 months (4 and 10.5) respectively (Fig 8). The survival curves were significantly different (P’0~0000l). DISCUSSION The value of BM biopsy in the study of MDS patients has received little attention in the past. Some of the previous reports preceded the FAB classification (Duhamel et al, 1976) or are on small series of patients (Martin et a/, 1986). Only Tricot et a/ (1984), and to a certain extent Ricardi et al (1988),have examined its prognostic value impact. In the present study, disagreement between the two observers was found only in 6% of the data analysed. The factors for such disagreement may be partially technical particularly in the case of ALIP and some features of dysmegakaryopoiesis (micromegakaryocytes). The use of paraffin samples instead of plastic-embedded ones may frustrate the assessment of ALIP. The BM biopsy in our patients was in general normo- or hypercellular. Only 15% of the cases had a cellularity lower than 40% (normal) in accordance with others authors (Tricot et a/, 1984; Martin et a/, 1986). The normal or increased thrombopoiesis detected in 77% of cases is coincidental with that observed by Martin et al(1986). The incidence of dysmegakaryopoiesis (83%)was similar to thatreportedbyTricotetaZ(1984)andMartinetal(1986) The dyserythropoiesiswas in our cases (77%)higher than the 58% reported by Martin et al (1986), although in the latter series megaloblastosis was not considered as diagnostic criterion. The presence of ALIP, observed in 5 5% of the BM biopsies, was slightly lower than that observed by Tricot e t a / (1984), and could be related to the use by these authors of plasticembedding techniques that provided more precise cytologic details. Eosinophilia has only been analysed previously by Duhame1 (1974).who found it in 3 7% of the cases. In our series its incidence was lower (26%). This difference is more striking when one considers that the cases studied by these authors were classed as ‘refractoryanaemia with partial myeloblastosis’ and in our series eosinophilia was found in only 22% of the RAEB and was not observed in any of the RAEB-t cases. The incidence of reticulin fibrosis (47%) was lower than that reported by Tricot et a/ (1984) and Martinet al(1986). while Duhamel et al(1976) observed fibrosis in only 28% of the cases. The presence of lymph follicles in 10%of the cases should be considered within normal ranges in older subjects (Williams, 1939; Duhamel, 1974). The characteristics of the BM biopsy have been shown to have significant prognostic influence in the outcome of

several haematologic diseases (Rozman et a/. 1984. 1985; Lazzarino et al, 1986; Ellis et a/, 1986; Bartl et a!, 1982). In the present study we demonstrate that two characteristics of the BM biopsy-dysmegakaryopoiesis and totaI cellularityafforded additional independent prognostic information. The resultant regression model divided patients into three separate groups with significantly different median survival rates (Fig 8). By contrast, eosinophilia, which in univariate analysis appeared to have a favourable effect (P=0*009), was not selected in the multivariate regression procedure. The adverse influence of increased BM cellularity was also observed by Ricardi et al (1988). On the other hand, the unfavourable influence of dysmegakaryopoiesisis not reproduced in Ricardi et al’s series, although these authors only evaluated the micromegakaryocytes, whereas we included other features of dysthrombopoiesis.The unfavourable effect of dysthrombopoiesishas also been pointed out by Varela et a/ (198 5), although this was in the analysis of BM aspirate data. It has been claimed that the incidence of ALIP is higher in patients with RAEB and RAEB-t and that ALIP is related to an increased risk of leukaemic transformation and shorter survival (Tricotet a/, 1984).In our series it was not possible to confirm this influence. Nevertheless, its incidence was slightly higher in RAEB, RAEB-t and CMML, and survival for ALIP-positive patients was somewhat shorter. Moreover, upon analysing the influence of ALIP in patients with less than 5%blast cells in BM aspirate, in RA and RARS the ALIPpositive cases had a lower survival than the ALIP-negative patients (20 versus 3 0 months) without significant differences. In summary, from the present results it can be concluded that some characteristics of the BM biopsy afford not only descriptive but also prognostic information for predicting the outcome of patients with MDS. ACKNOWLEDGMENT We thank Miss M. Carmen Payo for typing the manuscript. REFERENCES Bartl, R.,Frisch, B., Burkhardt. R., Kettner, G.. Mahl, G.. FathedMoghadan. A. & Sund, M. (1982) Assessment of bone marrow histology in the malignant lymphomas (non-Hodgkin):correlation with clinical factors for diagnosis, prognosis, classification and staging. British Journal of Huematology, 51, 51 1-530. Bennett, J.M., Catowsky, D.. Daniel, M.T.. Flandrin, G., Galton. D.A.G.. Gralnick, H.R. &Sultan, C. (1982) The French-AmericanBritish (FAB)CooperativeGroup. Proposals for the classificationof myelodysplastic syndromes. British Jouriiul of Haematology. 5 1, 189-199. Brookmeyer, R. & Crowley, J.A. (1982) A confidence interval for the median survival time. Biometrics, 38, 17-27. Cox, D.R. (1972) Regression model of life tables. Journal of the Royal Statistical Society, 34, 187-220. Dixon, W.J. (198 3) BMDP Statistical Software. University of California Press, Berkeley. Duhamel, G. (1974) Histopatologiedela Moelfe Osseuse. Masson et Cie. Pans. Duharnel, G., Muratore, R.& Bryon, D. (1976) Anemies refractaires avec myeloblastose partialle. Analyse d’un protocole portant SLU

Bone Marrow Biopsy in MDS 77 cas. Resultats de I’examen des biopsies medullaires. Nouvelle Revue Prancaise d’Hematologie,16, 81-86. Ellis, J.T., Peterson, P.. Geller, S.A. & Rappaport. H. (1986) Studies of the bone marrow in polycythemia Vera and the evolution of myelofibrosis and second hematologic malignancies. Seminars in Hematology, 23, 144-155. Garcia, S., Sanz. M.A.. Amigo, V., Colomina, P., Carrera, M.D., Lorenzo, J.I. & Sanz, G.F. (1988) Prognostic factors in chronic myelodysplasticsyndromes. A multivariate analysis in 107 cases. American Journal of Hematology, 27, 163-168. Hernandez-Nieto. L. & Rozman, C. (1980) Biopsia Medular en la Clinica Hernatoldgica. Salvat Ed.. Barcelona. Kaplan, R. & Meier. P. (1958) Non parametric estimation from incomplete observation. Journal of the American Statistical Association. 53, 457-481. Lazzarino. M., Morra, E., Castello. A., Inverardi, D., Coci, A., Pagnucco, G.. Magrini, U., Zei, G. & Bernasconi, A. (1986) Myelofibrosis in chronic granulocytic leukemia: clinicopathologic correlation and prognostic significance. British Journal of Haematology, 64, 227-240. Martin, E.. Woessner, S., Lafuente, R.. Florensa. L., Vilalta, J. & SansSabrafen. J. (1986)La biopsia medular en el estudio de las anemias refractarias adquiridas. Medicina Clinica, 87, 491-495. Peto, R., Pike, M.C., Armitage, P., Breslow, N.E., Cox, D.R., Howard, S.V.. Mantel, N.. McPherson, K.. Peto, J. & Smith, P.G. (1977) Dosing and analysis of randomized clinical trials requiring prolonged observation of each patient. BritishJournal ofCancer, 3 5 , l 39. Ricardi. A., Giordano. M., Giordano, P.. Cassano, E.. Cirino, M.. Cazzola, M., Scivetti, P.. Danova, M. & Ucci, G. (1988) Prognostic parameters in myelodysplastic syndromes: a multiple regression analysis. European Journal of Hematology. 40, 158-162.

33

Rozman, C., Monserrat, E., Rodriguez-F, J.M.. Ayats, R.. Vallespi, T.. Parody, R., Rios, A.. Prados, D., Morey. M., Gomis, F., Alcala, A.. Gutierrez. M., Maldonado, J., Gonzalez, C., Giralt, M., HernandezNieto. L., Cabrera, A. & Fernandez- Raiiada, J.M. (1984) Bone marrow histologic-pattern. The best singIe prognostic parameter in chronic lymphocytic leukemia. A multivariate survival analysis of 329 cases. Blood, 64, 642-648. Rozman, C., Martin, P., BruguBs, R. & Feliu, E. (1985) Valor pronostico de la histopatologia medular en la anemia aplasica. Sangre, 30, 982-992. S a w G.F., Sanz, M.A., Vallespi. T., Caiiizo, M.C., Torrabadella. M.. Garcia, S., Irriguible, D. & San Miguel, J.F. (1989) Two regression models and a scoring system for predicting survival and planning treatment in myelodysplastic syndromes A multivariate analysis of prognostic factors in 370 patients. Blood, 74, 395-408. Tricot, G., De Wolf-Peeters, C., Hendrickx, B. & Verwilghen, R.L. (1984) Bone marrow histology in myelodysplastic syndromes. I. Histological findings in myelodysplastic syndromes and comparison with bone marrow smears. British Journal of Haematology, 57, 423-430. Tricot, G., De Wolf-Peeters, C.. Vlientinck. R. & Verwilghen, R.L. ( 1 984) Bone marrow histology in myelodysplastic syndromes. 11. Prognostic value of abnormal localization of immature precursors in MDS. British journal of Haematology. 58, 217-225. Varela, B.L., Chuan. C., Wok J.E. & Bennet. J.M. (1985) Modifications in the classification of primary myelodysplastic syndromes: the addition of a score system. Hematological Oncology, 3, 55-63. Williams, H. (1939) The lymphoid nodules of human bone marrow. American Journal of Pathology, 15, 377-389.

Bone marrow biopsy in myelodysplastic syndromes: morphological characteristics and contribution to the study of prognostic factors.

Ten characteristics of bone marrow (BM) biopsies in paraffin sections, obtained at diagnosis from patients with myelodysplastic syndromes (MDS) classi...
2MB Sizes 0 Downloads 0 Views