Pembrolizumab for Front-Line Treatment of Extranodal NK/T-Cell Lymphoma

Introduction: Pembrolizumab is highly active in relapse or refractory extranodal NK/T-cell lymphoma (ENKTL) and therefore has potential to improve efficacy and tolerability of front-line treatment. Early-stage (ES) ENKTL is treated with combined modality therapy while advanced-stage (AS) ENKTL with an asparaginase-based chemotherapy regimen followed by consolidation with autologous or allogeneic stem cell transplant. Despite aggressive approaches, outcomes for pts with AS disease remain poor. We conducted a pilot study of front-line treatment with pembrolizumab to determine its single-agent activity and potential to impact outcomes for untreated pts.

Methods: This pilot study evaluated front-line treatment with pembrolizumab for ENKTL. All pts initially received 4 cycles of single agent pembrolizumab and were assessed with PET following 2 and 4 cycles. Pts with ES disease (by Ann Arbor) in complete response (CR) after pembrolizumab subsequently received involved site radiation therapy (ISRT) of 45 Gy followed by 8 more cycles of pembrolizumab maintenance. Pts with AS disease with at least partial response (PR) after pembrolizumab subsequently receive 3 cycles of modified SMILE (steroids, methotrexate, ifosfamide, peg-asparaginase, etoposide) followed by 8 more cycles of pembrolizumab maintenance. The primary endpoint was CR rate after 4 cycles of pembrolizumab. Secondary endpoints included progression-free survival (PFS), overall survival (OS) and assessment for markers of response or resistance to pembrolizumab. Progression events were defined as progression, death, or initiation of non-protocol therapy. Assuming a CR rate of at least 40%, we planned to enroll 19 pts to yield a sufficiently narrow 95% confidence interval of 0.44 to provide rationale for a larger phase II study.

Results: Among 19 pts, 15 (79%) had ES disease and 4 (21%) had AS disease. Median age was 53 (range 19-83), 58% were male, 53% white, 32% Asian, and 16% unknown. Immune related toxicities included adrenal insufficiency (1pt, G2), pancreatitis (1pt, G2), hypothyroidism (1pt, G1), and hepatitis (7pts, G1). Treatment related toxicities were all grade 1 or 2 with the exception of 1 episode of febrile neutropenia (G3, related to SMILE) and 1 episode of grade 3 lipase elevation (asymptomatic).

After 4 cycles of pembrolizumab, 8 of 19 (42%) pts achieved CR (7 ES and 1 AS). Among 15 pts with ES disease, 7 (47%) achieved CR, 1 (7%) PR, 4 (27%) stable disease (SD), and 3 (20%) had progression of disease (POD). Among 4 pts with AS disease, 1 (25%) achieved CR and 3 (75%) POD.

After a median follow-up of 18.4 months, overall PFS and OS were 37% and 69%, respectively. For pts with ES disease, PFS and OS were 40% and 74%; for pts with AS disease, PFS and OS were 25% and 50%.

Among the 8 pts with CR after pembrolizumab, 1 pt with ES disease experienced unrelated sudden death following 4 cycles of pembrolizumab, ISRT, and 1 cycle of pembrolizumab maintenance. For the other 7 pts in CR after 4 cycles of pembrolizumab, the median follow-up is 36 months (range 16-62 months) and all remain in remission.

Conclusion: In untreated ENKTL, complete response rate to 4 doses of pembrolizumab is 42%. Pts achieving CR after pembrolizumab alone have the potential to achieve long-term remissions with pembrolizumab and ISRT (if ES disease) or pembrolizumab and SMILE (if AS disease). Understanding which pts are likely to achieve CR to pembrolizumab is critical for applying this treatment program going forward. Studies aimed at identifying predictors for CR in pts with ENKTL are ongoing.