Non-invasive classification of MPN subtypes

Results from a prospective, multicenter study evaluating non-invasive multiparametric characterization of myeloproliferative neoplasms (MPN) in 128 patients with newly diagnosed (ND) essential thrombocythemia (ET; n = 44), prefibrotic myelofibrosis (preMF; n = 53), primary overt myelofibrosis (PMF; n = 21), and unclassifiable MPN (MPN-U; n = 10) were published in Leukemia by Mosnier et al.

Key data: The variant allele frequencies (VAFs) of JAK2 and CALR driver mutations were 9.5% and 19% in ET, 15.3% and 33.3% in preMF, and 37.9% and 44.7% in PMF, respectively. Mutational complexity increased progressively from ET to preMF and PMF, with high-risk myelofibrosismutations more frequent in PMF (48%) than in ET (4.5%; p < 0.001) and preMF (7.5%; p < 0.001). Significant differential expression of inflammatory mediators (interleukin [IL]-1RA, IL-1β, IL-6, S100A8/S100A9, and CCL4), immune activation markers (CD25 and CXCL10), and matrix regulators (YKL-40 and TIMP1) was observed between PMF and ET/preMF. Spleen size increased from ET (9.9 cm) to preMF (11 cm) and PMF (15.1 cm; PMF vs ET, p < 0.001; PMF vs preMF, p < 0.001). A two-step Lasso regression model reliably identified PMF (area under the curve [AUC], 0.909; 95% confidence interval [CI], 0.804–0.994), with a sensitivity and specificity of 76.2% and 95.5%, respectively, and demonstrated reasonable discrimination between ET and preMF (AUC, 0.762; 95% CI, 0.656–0.847), with a sensitivity and specificity of 70.9% and 74.1%, respectively.

Key learning: Non-invasive multiparametric assessment, integrating genomic, cytokine, and clinical data, effectively discriminates PMF from ET and preMF, and may help to distinguish preMF from ET, supporting the hypothesis of a pathophysiological continuum across MPN subtypes. External validation in larger multicenter cohorts is required.