Early software-assisted response predicts survival in a prospective cohort of 258 newly diagnosed cGvHD patients.

Two hundred and fifty-eight patients with newly diagnosed cGVHD have been enrolled in a multicenter study (NCT02991846). Primary endpoint was failure free survival (FFS); secondary endpoints were ORR at 6 and 12 months after best available therapy (BAT) and OS. Response evaluation was supported by a software (Crosy) collecting cGvHD master data. Software-assisted (SA) response and clinician-based (CB) evaluation, according to NIH criteria, have been performed in parallel at 6 and 12 months. Among 258 patients, 200 needed systemic therapy (ST); 167 were evaluable for response; ORR according to SA was 62% (CR:19%; PR: 43%), and 65% (CR: 23%; PR: 42%) according to the CB evaluation; overall agreement was 66%. In multivariate analysis, only 6-month SA response significantly predicted OS, being more effective in capturing Mixed Responses, compared to CB evaluation. After 12 months 23% of patients were able to stop ST: significantly more patients who maintained ST1 stopped ST (39%) than those who requested a ST1 change (9,5%); (p < 0.001). The 2-year FFS of patients receiving ST was 67%; 2-year OS was 85%: 32 patients (18%) died (30 due to cGVHD progression or causes coexisting with cGvHD). SA evaluation of response in cGVHD is reliable and predicts FFS.