Oral azacitidine compared with standard therapy in patients with relapsed or refractory follicular helper T-cell lymphoma (ORACLE): an open-label randomised, phase 3 study

Background Follicular helper T-cell lymphomas (TFHL) harbour frequent alterations in genes that regulate DNA methylation. Preliminary reports suggest that treatment with 5-azacitidine has clinical activity in patients with relapsed or refractory TFHL. We aimed to compare the oral form of azacitidine with investigator's choice standard therapy (ICT; ie, gemcitabine, bendamustine, or romidepsin) in patients with relapsed or refractory TFHL. Methods Patients older than 18 years with relapsed or refractory TFHL (angioimmunoblastic T-cell lymphoma, follicular lymphoma, or nodal T-cell lymphoma with phenotype, ie, positive with two or more markers among CD10, BCL6, CXCL13, PD1, or ICOS) based on the 2017 WHO classification of haematological neoplasms, with an Eastern Cooperative Oncology Group performance status score of 0–3, were recruited in university hospitals from five European countries and from Japan. Patients were randomly assigned 1:1 to treatment with either azacitidine given at a dose of 300 mg once a day (200 mg in Japanese patients) for 14 days in a 28-day cycle or gemcitabine, bendamustine, or romidepsin according to the investigator's choice. Random assignment was stratified by the number of previous lines of therapy and by the presence of previous or concomitant myeloid malignancy. The primary endpoint was investigator-assessed progression-free survival, presented in the intention-to-treat population. This Article is the final analysis of this trial, registered at ClinicalTrials.gov (Europe NCT03593018 and Japan NCT03703375). Findings 86 patients (median age 69 years [IQR 62–76], 50 patients were male, 36 were female) were enrolled between Nov 9, 2018, to Feb 22, 2021; 42 in the azacitidine group and 44 in the ICT group. With a median follow-up of 27·4 months (IQR 20·2–32·9), the median progression-free survival was 5·6 months (95% CI 2·7 –8·1) in the azacitidine group versus 2·8 months (1·9–4·8) in the ICT group (hazard ratio of 0·63 (95% CI 0·38–1·07); 1-sided p=0·042). Grade 3–4 adverse events were reported in 32 (76%) of 42 patients in the azacitidine group versus 42 (98%) of 43 patients in the ICT group. The most adverse grade 3 or worse adverse events were haematological (28 [67%] of 42 patients vs 40 [93%] of 43 patients), infection (8 [19%] and 14 [33%]), and gastrointestinal (5 [12%] vs 1 [2%] for azacitidine and ICT, respectively). There were two treatment-related deaths in the azacitidine group (one endocarditis and one candidiasis) and three in the ICT group (one heart failure, one COVID-19, and one cause unknown). Interpretation Although the pre-specified primary outcome of the trial was not met, the favourable safety profile suggests that azacitidine could add to the treatment options in these difficult to treat diseases especially in combination with other drugs. Trials with combination are in preparation in a platform trial. Funding Bristol-Myers Squibb. Translation For the French translation of the abstract see Supplementary Materials section.