Abstract 2238Poster Board II-215Chronic graft-versus-host disease is the most common late complication occurring following allogeneic transplantation, and is the principal cause of late morbidity and non-relapse mortality. Corticosteroids remain the cornerstone of primary treatment for cGvHD with about half of patients responding. Patients who fail to respond to initial corticosteroid therapy, however, have a poor outcome with no standard approach uniformly accepted. Pentostatin is a purine nucleoside analogue that irreversibly inhibits adenosine deaminase, with accumulation of dATP, causing apoptosis of lymphocytes, and holds promise for the management of acute and cGvHD. We conducted a multi-institutional, prospective phase II trial to evaluate the response rate (RR) of pentostatin in patients with steroid-refractory cGvHD. In a two-stage design with both types I and II errors set at 0.1, we hypothesized that pentostatin would produce a response rate (complete response + partial response) of 40% or more, with less than 20% being unaccpetable. Patients were eligible if they were >18 years old, had a Karnofsky performance status (KPS) of >30%, histologically-proven extensive stage cGvHD that was refractory to corticosteroids defined as progression of disease despite at least 2 weeks of treatment with 1 mg/kg/day of prednisone or equivalent, no response or minor response to at least 4 weeks of 1 mg/kg/day prednisone (or equivalent), or who achieved only a partial response after 8 weeks of corticosteroid therapy. Response was assessed using the Hopkins scoring system as previously reported (Jacobsohn et al. J Clin Oncol, 25:4255-61, 2007). Pentostatin was administered at a dose of 4 mg/m2 IV every 2 weeks for a maximum of 12 doses. From March 2004 to March 2007, 38 patients with median age 46.5 (27-66) years and median KPS 70% (40%-90%) were treated with pentostatin. cGvHD followed transplantation from matched related donors in 29 and unrelated donors in 9 patients. Thirty-three patients had progressive onset cGvHD, while onset was quiescent in 2, and de novo in 3 patients. The median time from onset of cGvHD to start of study treatment was 416 (26-2,813) days, and the median number of prior lines of therapy for cGvHD was 3 (range, 1-6). cGvHD involved skin (n=35) with scleroderma present in 21 patients, oral mucosa (n=24), liver (n=9), gut (n=4), and lung (n=2). Thirty-five patients were evaluable for response; 2 patients died after 6 and 7 days of treatment, respectively, due to sepsis, and data was incomplete in 1 patient who committed suicide after 3 cycles. Of 35 patients evaluable for response, there were 2 complete responses (CR), 5 partial responses (PR) for a total response rate of 20%. If minor responses (in 7 patients) are also considered, then 14 of 35 (40%) had an objective improvement in cGvHD. Six patients had stable disease, 1 a mixed response, and 14 patients progressed on therapy. Of 34 patients evaluable for toxicity, grade 3-4 hematological toxicity included neutropenia (n=4), thrombocytopenia (n=3), and anemia (n=1). Grade 3-4 clinically significant non-hematological toxicity that was at least possibly related to pentostatin included fatigue (n=3), renal failure (n=3), anorexia (n=2), infection (n=9), CNS hemorrhage (n=1), and rash (n=1). Thirteen patients died on study; causes of death were infection (n=6), cGvHD progression (n=3), and unrelated causes (n=3). The 1-year and 2-year progression-free survival (for GvHD) for all patients treated on study were 32% (95% CI: 18%-46%) and 24% (95% CI: 12%-38%), respectively. The 1-year and 2-year overall survival was 53% (95% CI: 38%-60%) and 50% (95% CI: 33%-65%), respectively. We conclude that pentostatin is an active agent in patients with steroid-refractory chronic GvHD but should be investigated further in cGvHD patients in an earlier course of their disease where a greater impact on long-term outcome may be observed. Disclosures:No relevant conflicts of interest to declare.