BACKGROUND. TheWHO classification of myelodysplastic syndromes (MDS) is based on the evaluation of bone marrow morphology. The two categories of REAB-I and RAEB-II are apparently easy to differentiate on the basis of bone marrow blast percent. However there are no so far data about the differences among cytology, histology and immunophenotypic evaluation of blasts in order to discrimante non-RAEB from RAEB-I and RAEB-II categories.

PATIENTS AND METHODS. The Piemonte MDS Registry was born in 1999 thanks to the cooperation of both Haematology and Internal Medicine departments of our region, with the following aims: a) to follow homogeneous guidelines in diagnosis and treatment of MDS; b) to collect epidemiological and clinical information on a large group of patients; c) to cryopreserve bone marrow cells for molecular biology studies. When obtaining an informed consent, data of patients were prospectively centrally recorded through our web site. A retrospective analysis on differences in diagnosing RAEB, comparing conventional cytology on bone marrow smears (CBM), histochemical evaluation of CD34+ cells on bone marrow trephine biopsy (HBM), and cytofluorimetric count of CD34+ and CD117+ cells (IBM) has been done.

RESULTS. From June 1999 to December 2003, 633 MDS patients were registered from 37 different institutions: 364 (57%) from haematology and/or academic institutions and 269 (43%) from internal medicine departments of community hospitals. Mean age was 72 (range 23–69). The actual diagnostic distribution of cases according to the WHO criteria based on only morphology evaluation of bone marrow smears was: non-RAEB 429 (68%), RAEB-I 134 (21%), and RAEB-II 70 (11%). Information about the quantification of blasts with both CBM and HBM techniques was avilable in 243 cases. An IBM evaluation was also available in 89 out of this 243 cases. A disagreement between CBM and HBM was evident in 65/243 cases (27%), with HBM over-evaluating and under-evaluating WHO class on the basis of blasts count in 54/243 (22%) and 11/243 cases respectively. When comparing CBM and IBM the disagreement was even higher in 29/89 cases (33%), with IBM over-evaluating blast percent in 9 (10%) and under-evaluating it in 20 cases (23%). The disagreement betwen HBM and IBM was maximum with a value of 39%. The role of CBM in predicting a different prognosis of non-RAEB, RAEB-I and RAEB-II was confirmed. However, when comparing the prognostic value of the three different methods of computing bone marrow blasts, IBM was the best in order to define the good prognostic non-RAEB group.

CONCLUSIONS. The distinction among non-RAEB, RAEB-I and RAEB-II is far from beeing highly accurate and reproducible. Important differences are present among CBM, HBM and IBM. While CBM remain the conventional standard system, IBM could offer a tool better and more reproducible than CBM in order to define MDS categories on the basis of blast percentage. A large multicenter study could be useful in order to clarify this point.

Author notes

Corresponding author