Interobserver concordance of assessments of dysplasia and blast counts for the diagnosis of patients with cytopenia: From the Japanese central review study

Abstract

The diagnosis of myelodysplastic syndromes (MDS) is based on morphology and cytogenetics. However, limited information is currently available on the interobserver concordance of the assessment of dysplastic lineages (<10% or ≥10% in bone marrow (BM)). The revised International Prognostic Scoring System (IPSS-R) described a new threshold (2%) for BM blasts. However, the interobserver concordance of the categories (0–≤2% and >2–<5%) has limited data. The purpose of the present study was to investigate the assessment of dysplastic lineages and IPSS-R reproducibility. Our study was divided into two Steps. In each Step, the microscopic examinations were performed separately by two morphologists. Regarding the category of BM blasts ≤2% and >2–<5%, interobserver agreement was more than ‘moderate’ in all pairs (kappa test: 0.43–0.90). Regarding dysgranulopoiesis (dysG) and dyserythropoiesis (dysE) in BM, interobserver agreement was more than ‘moderate’ in all pairs (kappa test, dysG: 0.45–0.96, dysE: 0.45–0.81). Regarding the category of dysmegakaryopoiesis (dysMgk) in BM, interobserver agreement was more than moderate in 4 out of 5 pairs (kappa test: 0.58–1.00), and was fair for one pair (kappa test: 0.37). We consider that high interobserver concordance may be possible for the BM blast cell count (≤2% or >2–<5%) and dysplasia (<10% or ≥10%) of each lineage.