Paper Details

ABSTRACT

Introduction:

Patients with advanced stage cutaneous T-cell lymphoma (CTCL) have poor prognosis. Checkpoint inhibitors and duvelisib independently have shown efficacy in CTCL with response rates of 38% and 34% respectively.  Preclinical data suggests that PI3-kinase inhibition can augment immunotherapy. We conducted a phase I study evaluating the safety and efficacy of nivolumab and duvelisib in stage IIB-IVB CTCL through the Experimental Therapeutics Clinical Trials Network (ETCTN). (NCT04652960)

Methods:

This was a multicenter phase I study with a 3+3 design conducted through the ETCTN. Eligible patients had stage IIB-IVB CTCL, ≥1 prior systemic therapy and adequate organ function. Nivolumab 480mg IV (D1) and duvelisib (25-75mg BID) were given on a 28-day cycle until progression or intolerance.  Duvelisib dosage (D1-28, BID) was escalated (DL1: 25mg, DL2: 50mg, DL3: 75mg) with dose limiting toxicities (DLTs) defined in cycle 1. Due to cumulative toxicity, a modified schedule (Schedule A) with de-escalating doses of duvelisib (D1-14, BID) was adopted (DL3a: 75mg, DL2a: 50mg, DL1a: 25mg) with DLTs defined through cycle 3. DLTs were defined as grade ≥3 non-hematologic toxicities, grade 4 hematologic toxicity, grade ≥3 thrombocytopenia with bleeding, febrile neutropenia or inability to receive cycle 2 due to ongoing grade ≥2 toxicities.  Response evaluations were conducted after cycle 2, cycle 4 and then every 3 cycles using the Global Response Criteria for CTCL (Olsen et al  2011).

Results:

A total of 19 patients (14 mycosis fungoides, 5 Sezary Syndrome) were enrolled from 10/2021-12/2024 with a median age of 66 years (38-75 years) including 10 male patients. All patients had advanced stage disease: IIB (n=5; 26%), IIIA (n=2; 11%); IVA (n=10; 53%); IVB (n=2; 11%).  Patients had a median of 5 prior lines of therapy.

Eleven patients were treated on the original schedule. Duvelisib doses were escalated to 75mg BID without DLT within cycle 1. Repeated immune mediated toxicities in later cycles resulted in holding/discontinuation of therapy for 6/11 patients, including: DL1: one with repeated grade 2 transaminitis requiring holds, DL2: one with grade 3 transaminitis; DL3: one with grade 3 transaminitis, one with grade 4 transaminitis, one with grade 3 pneumonitis. Additionally, one patient at DL3 died of unwitnessed arrest after a recent admission for sepsis.

Therefore, the duvelisib schedule was modified to D1-14 but DLTs through cycles 1-3 were seen at all dose levels. At DL3a, 2/2 patients developed grade≥3 transaminitis. At DL2a, 2/3 patients had DLTs (grade 4 Steven’s Johnsons, transaminitis). At DL1a, 2/3 patients had DLTs (grade 3 pancreatitis, grade 3 cellulitis). Therefore, the toxicity exceeded the prespecified threshold and the study was suspended.

With regards to other toxicities, toxicities (regardless of attribution) that occurred in ≥10% patients included: Transaminitis (67%), hyperbilirubinemia (12%), elevated alkaline phosphatase (22%), rash (44%), infections (44%), edema (22%), weakness (22%), pain (17%), fatigue (34%).  Grade ≥3 adverse events included transaminitis (54%), hyperbilirubinemia (6%), rash (11%), Steven-Johnson syndrome (6%), pancreatitis (6%), acute kidney injury (6%), encephalopathy (6%), pneumonitis (6%), sepsis (11%), cellulitis (6%), CMV reactivation (6%), osteomyelitis (6%), fatigue (6%). One patient with sepsis had death from cardiac arrest.

Of the 19 patients who received treatment, 16 were evaluable for response. Three patients discontinued therapy in cycle 1 due to toxicity (sepsis complicated by cardiac arrest, pancreatitis, and Stevens-Johnson). Overall response rate was 31% (5/16) with 1 complete remission (DL3) and 4 partial remissions (DL3, DL3A, DL2A, DL1A). Of the responders, 4 had mycosis fungoides (1 with complete remission, 3 with partial remission) and 1 with a partial remission had Sezary syndrome. Median number of cycles given was 2. Median duration of response was 20.2 weeks with responses lasting for 2, 9, 38 months and one patient remains in partial remission at 22 months. 

Conclusion:

Duvelisib combined with nivolumab in CTCL resulted in a response rate 31% similar to single agent duvelisib or nivolumab. However, immune mediated therapies were dose limiting.  Further correlative studies are ongoing to understand the mechanisms of immune mediated toxicity.