A Phase I/II Trial of Vorinostat (SAHA) in Combination with Rituximab-CHOP in Patients with Newly Diagnosed Advanced Stage Diffuse Large B-Cell Lymphoma (DLBCL): SWOG S0806

Background: Loss of major histocompatibility Class II antigens (MHCII) in diffuse large B-cell lymphoma (DLBCL) is associated with a decreased CD8+ tumor-infiltrating T-lymphocyte (TIL) response and poor patient survival. Transcription of the MHCII gene complex is under the control of the master transactivator, CIITA, which in part is regulated by histone acetylation. We tested the hypothesis that combination of histone deacetylase inhibitor vorinostat with standard chemotherapy will enhance MHCII expression and improve patient outcome in DLBCL.

Methods: SWOG S0806 was a phase I/II open label trial of vorinostat given at 400 mg po daily on days 1-9 (subsequently reduced to days 1-5) combined with Rituximab-CHOP (R-CHOP) at standard doses, given on day 3 of a 21-day cycle for 8 cycles. Eligibility criteria included having newly diagnosed advanced stage DLBCL, international prognostic index (IPI) of at least 1, and lack of known CNS involvement or HIV. Primary endpoint of phase I was to establish maximum tolerated dose (MTD) of vorinostat with standard R-CHOP. Primary endpoint of phase II was to estimate 2-year progression free survival (PFS). Translational endpoints included correlation of pre-treatment acetylation status of histones, expression of MHCII genes, and percentage of TIL to PFS; and correlation of cytokine profile to response and outcomes.

Results: Phase I was open in 5 SWOG institutions and enrolled 11 patients. There were only 2 patients who had dose limiting toxicities in the first cycle - grade 3 febrile neutropenia and grade 4 hypokalemia - allowing phase II to proceed with the original vorinostat dosing of 400 mg daily days 1-9, at all SWOG institutions. However, excess rates of febrile neutropenia and sepsis were seen upon further follow up, and the study was amended to reduce the duration of vorinostat to days 1-5. A total of 72 patients were enrolled in phase II, of which 8 were ineligible and 2 withdrew consent prior to treatment. For the remaining 62 patients, median age was 64 years, 92% had stage III/IV disease, 39% B symptoms, 61% elevated LDH, 39% had more than 1 extranodal site of involvement, with IPI breakdown of 13/26/47/13/2%. Notable grade 3-4 non-hematologic toxicities included febrile neutropenia (39%), sepsis (18%), fatigue (15%), hypokalemia (11%), hyponatremia (10%), and small bowel perforation (3%). Grade 3-4 hematologic toxicities included neutropenia (60%), anemia (35%), and thrombocytopenia (35%). There was one death in phase I from sepsis and multi-organ failure at the end of 8 cycles of treatment, but no deaths from toxicity in phase II. Overall response rate was 81% (95% CI: 69-90%). With median follow-up of 24.3 months, estimate of 2-year PFS is 72% (95% CI: 58%, 81%) and of 2-year OS is 85% (95% CI: 74%, 92%). Analysis of the panel of 30 cytokines performed on matched serum specimens of 40 patients showed correlation of baseline elevated IL-2R levels with worsened PFS and OS, and correlation of decrease in Epidermal Growth Factor level with improved PFS and OS. Results of immunohistochemical stains for expression of MHCII genes and percentage of TIL will be reported at the meeting.

Conclusions: The regimen of vorinostat-R-CHOP achieved 2-year PFS estimate of 72%, which is slightly more than 68% expected from R-CHOP alone per IPI adjusted historical rate, but less than an IPI adjusted target of 78% that would be sufficient to warrant further investigation. It also resulted in unexpected excess rates of febrile neutropenia and sepsis. This regimen cannot be recommended for the broad DLBCL population. Current studies are focused on finding biomarkers of response to histone deacetylase inhibitors.