Inter-Observer Discordance With The Diagnosis Of Myelodysplastic Syndromes With Less Than 5% Bone Marrow Blasts: Unilineage Vs Multilineage Dysplasia and Reproducibility Of The Threshold Of 2% Blasts

Previous studies have shown the reproducibility of the 2008 WHO classification in myelodysplastic syndromes (MDS), especially in the categories refractory cytopenia with multilineage dysplasia (RCMD) and refractory anemia with excess blasts (RAEB) type 1 and type 2. However, there is few data regarding the reproducibility of the diagnosis of MDS with unilineage dysplasia. Recently, two new morphological categories have been included in the R-IPSS, distinguishing the BM blast cell count between 0 and 2%, and >2 - <5%. This distinction is critical for establishing prognosis, but the reproducibility of this threshold is still not demonstrated.

1-To explore the reliability of the 2008 WHO classification of MDS in cases without excess of marrow blasts (<5%), regarding unilineage vs multilineage dysplasia. 2- To study if the threshold of 2% marrow blasts is reproducible among different observers.

110 samples from patients previously diagnosed with MDS according to 2008 WHO classification were included. The design of the study recommended include preferentially patients with <5% BM blasts, encouraging the inclusion of cases with unilineage dysplasia, but the proportion of samples from each category was not predefined. Specimens were collected from 11 hospitals and were evaluated by 11 morphologists. Each observer evaluated 20 samples, and each sample was analyzed independently by two morphologists. The second observer was blinded to the clinical and laboratory data, except for the peripheral blood (PB) counts. The interobserver concordance was evaluated using the Cohen kappa test.

Finally 108 samples were considered suitable by the second observers. Regarding 2008 WHO categories, 7 cases had refractory cytopenia with unilineage dysplasia (RCUD), 15 refractory anaemia with ring sideroblasts (RARS), 73 RCMD (including 33 with RS), 9 MDS with del(5q), 2 RAEB-1 and 2 MDS -unclassified (MDS-U). The MDS–U included the only case of the series with 1% blasts in the PB and one case with unilineage dysplasia and pancytopenia. Discordance was observed in 36/108 cases (33%). There was very good correlation in identifying the threshold of 15% RS (kappa test=0.755). The inter-observer concordance was very good regarding MDS with del(5q) (8/9, 88%), and RCMD (56/73 cases, 77%). One of the two cases of RAEB-1 was recognized by the second observer, and the only case with 1% PB blasts was also identified. Despite the bias of selecting cases with<5% BM blasts, there were 5 cases of RCMD classified as RAEB-1 by the second observers. The new morphological categories using the threshold of 2% BM blasts were not reproducible (discordance rate 31/108 cases, k test=0.279).

Diagnosis of MDS with unilineage dysplasia (RCUD, or RARS, or MDS-U) was assessed in 33 patients, by either of the two observers. However, they only found concordance in 3/15 cases originally diagnosed with RARS (20%) and in 3/7 with RCUD (42%).

Amongst MDS WHO 2008 categories, inter-observer discordance seems to be high regarding cases with unilineage dysplasia.The proposed 2% cutoff BM blasts as settled by the R-IPSS may be not easy to reproduce by morphology in real practice. MDS-U category could be probably discarded in upcoming classifications, as different groups like ours proves it to be marginal.

No relevant conflicts of interest to declare.