Hypoplastic MDS: Validation of the New World Health Organization 2022 Proposal

Introduction: Myelodysplastic syndromes (MDS) are a heterogeneous group of hematologic neoplasms that is characterized by ineffective hematopoiesis. The morphological diagnosis is essential and reveals a subset of patients with hypoplastic marrow, referred to as hypoplastic MDS (MDS-h) that is more recently being proposed as a distinct entity in the World Health Organization (WHO) 2022 proposal for myeloid neoplasms. MDS-h shares several features with aplastic anemia (AA) and paroxysmal nocturnal hemoglobinuria (PNH), thus, it needs a careful diagnostic approach, but also, the natural history of the disease is distinct than those with non-hypoplastic MDS (MDS-nh) warranting consideration for further subclassification.

Here, we aim to study the characteristics and molecular profiling of MDS-h as opposed to the (MDS-nh), in addition to treatment and disease outcomes, and overall survival.

Methods: We conducted a retrospective analysis of all patients diagnosed with MDS at the H. Lee Moffitt Cancer Center & Research Institute from our MDS database. Patients were divided into two groups: those with MDS-h, defined as bone marrow cellularity of less than 30%, and those with MDS-nh, with cellularity greater than 30%. We further subdivided patients based on the marrow blasts (<5% vs >=5%). Patients with MDS/MPN were excluded.

Results: A total of 4037 patients with MDS were identified, of those, 403 (10%) were MDS-h. The median age at diagnosis was 65 and 68 for MDS-h and MDS-nh respectively (P= 0.03).

In females, MDS-h was observed more compared to MDS-nh (41% vs 35%, P= 0.008). There was no difference in terms of IPSS-R, LGL clones, PNH clones, and history of autoimmune diseases. MDS-RS was more observed in MDS-nh.

Interestingly, therapy-related MDS (t-MDS) was more observed in MDS-h (29% vs 20%, P <0.005). Patients with MDS-h had more thrombocytopenia, neutropenia and were more likely to be platelet transfusion dependent (31% vs 23% P=0.001).

The following somatic mutations were found more frequently in MDS-nh compared to MDS-h: SF3B1 (P<0.005), SRSF2 (P<0.005), TET2 (P=0.004), and ASXL1 (P=0.02).

There was a better response to ATG/CSA among MDS-h compared to MDS-nh (46% vs 21% P=0.019). On the contrary patients with MDS-nh were more likely to respond to lenalidomide when compared to patients with MDS-h (46% vs 22% P=0.001). There was no statistically significant difference in response to hypomethylating agents.

The median overall survival (OS) for MDS-h was 41.6 months compared to 35 months in patients with MDS-nh (P=0.05). The HR for OS was 0.88 (P =0.057) adjusted for IPSS-R. There was no difference in the rate of AML transformation.

Then, we further categorized patients based on the marrow blasts. Among MDS-h, 43% (170/400) had myeloblasts >=5% (MDS-EB), The median OS was 53 months for MDS-h with < 5% blasts (n=231) compared to 27 months for those with >= 5% blasts (n=170) (P < 0.005).

Among all patients with MDS-EB, those with MDS-h (n= 170) had median OS of 27 month compared to 19 months for MDS-EB with MDS-nh (n=1465) (P=0.02).

Conclusions: In summary, we conclude that MDS-h accounted for 10% of MDS cases. The OS for MDS-h is better than MDS-nh. Additionally, MDS-h tended to have lower frequency of splicing and epigenetic somatic mutations. Among MDS-h, patients with marrow myeloblasts >=5% clearly had inferior outcomes compared to those MDS-h with blasts <5%. Therefore, our findings support the new WHO 2022 proposal for MDS-h as a distinct entity with restriction to patients with marrow cellularity of <30% and marrow myeloblasts <5%.