We propose to test whether the new refined Minnesota risk score, which represents a new tool for acute Graft-versus-Host-Disease (aGVHD) grading, may be useful to predict the final outcome of patients with aGVHD after haploidentical stem cell transplantation (Haplo-SCT) with post-transplant cyclophosphamide (PT-Cy). Hundred consecutive patients with grade 2-4 aGVHD were included. Twenty-two percent of the patients had high-risk (HR) aGVHD and had a lower chance to respond at day 28: 41% of non-responders (NR) were in the HR vs 13% in the standard-risk (SR) group (p = 0.003). By multivariate analysis, grade 3-4 aGVHD according to the traditional Keystone classification was the main independent predictor of non-response to front-line treatment at day 28, while HR aGVHD by the new refined Minnesota score remained the main independent variable associated with adverse NRM and OS. The new Minnesota refined risk score is a useful tool to predict the outcome of patients with aGVHD after Haplo-SCT with PT-Cy. Due to the few patients exchanging between categories in the two classifications, it is not possible to discriminate which system better predicts the outcome of patients with aGVHD in the setting of Haplo-SCT. Extending these preliminary observations to a larger cohort is warranted.