‘We cannot paint them all with the same brush’: the need for a better definition of patients with myelodysplastic syndromes for clinical trial design