Fit patients with mantle cell lymphoma aged 18-65 years are usually given cytarabine and rituximab-based induction regimens followed by autologous haematopoetic stem-cell transplantation (HSCT). We investigated whether post-autologous HSCT maintenance with lenalidomide improves progression-free survival in this population. This open-label, randomised, multicentre, phase 3 trial was done at 49 haematology and oncology units in Italy and Portugal. Eligible patients had Ann Arbor stage III or IV treatment-naive mantle cell lymphoma (or stage II plus bulky disease [≥5 cm] or B symptoms), and had evidence of cyclin D1 overexpression or the translocation t(11;14)(q13;q32). Patients were aged 18-59 years with Eastern Cooperative Oncology Group (ECOG) performance status 0-3, or aged 60-65 years with ECOG 0-2. After an optional prephase with vincristine and steroids (intravenous vincristine 1·4 mg/m2 on day 1, oral prednisone 100 mg [total dose] on days 1-5), patients were given three courses of R-CHOP (21-day cycle, intravenous rituximab 375 mg/m2 on day 1; intravenous doxorubicin 50 mg/m2, vincristine 1·4 mg/m2, and cyclophosphamide 750 mg/m2 on day 2; oral prednisone 100 mg/m2 on day 2-6). Patients then received one cycle of high-dose CTX (intravenous cyclophosphamide 4 g/m2 on day 1, intravenous rituximab 375 mg/m2 on day 4). After restaging, patients received two cycles of R-HD-cytarabine (high-dose intravenous cytarabine 2 g/m2 every 12 h on days 1-3, intravenous rituximab 375 mg/m2 on days 4 and 10). Patients with complete remission or partial remission proceeded to autologous HSCT and responding patients (complete remission or partial remission) with haematological recovery were randomly assigned (1:1) to receive 24 courses of oral lenalidomide maintenance (15 mg per day for patients with platelets >100 × 109 cells per L or 10 mg per day for platelets 60-100 × 109 cells per L, days 1-21 every 28 days) for 24 months, or observation. The primary endpoint was progression-free survival, measured in the randomised population. This study is registered with EudraCT (2009-012807-25) and ClinicalTrials.gov (NCT02354313). Between May 4, 2010, and Aug 24, 2015, 303 patients were screened for inclusion and 300 patients were enrolled (median age 57 years, IQR 51-62; 235 [78%] male). 95 patients were excluded before randomisation, mostly due to disease progression, adverse events, and inadequate recovery. 104 patients were randomly assigned to the lenalidomide maintenance group and 101 patients to the observation group. 11 (11%) of 104 patients assigned to lenalidomide did not start treatment (3 withdrew, 6 adverse events or protocol breach, 2 lost to follow-up). At a median follow-up of 38 months after randomisation (IQR 24-50), 3-year progression-free survival was 80% (95% CI 70-87) in the lenalidomide group versus 64% (53-73) in the observation group (log-rank test p=0·012; hazard ratio 0·51, 95% CI 0·30-0·87). 41 (39%) of 104 patients discontinued lenalidomide for reasons including death or progression. Treatment-related deaths were recorded in two (2%) of 93 patients in the lenalidomide group (1 pneumonia, 1 thrombotic thrombocytopenic purpura), and one (1%) of 101 in the observation group (pneumonia). 59 (63%) of 93 patients in the lenalidomide group had grade 3-4 haematological adverse events versus 12 (12%) of 101 patients in the observation group (p<0·0001). 29 (31%) of 93 patients in the lenalidomide group and eight (8%) of 101 patients in the observation group had grade 3-4 non-haematological adverse events (p<0·0001), of which infections were the most common.Serious adverse events were reported in 22 (24%) of 93 patients in the lenalidomide group and five (5%) of 101 patients in the observation group. Pneumonia and other infections were the most common serious adverse events. Despite non-negligibile toxicity, lenalidomide after autologous HSCT improved progression-free survival in patients with mantle cell lymphoma, highlighting the role of maintenance in mantle cell lymphoma. Fondazione Italiana Linfomi and Celgene.