Safety and efficacy of parsaclisib in combination with rituximab, bendamustine + rituximab, or ibrutinib in patients with previously treated B-cell lymphoma: analysis of a phase 1 dose-finding study (CITADEL‑112)

Abstract

Parsaclisib, a potent and highly selective phosphoinositide 3-kinase δ inhibitor, has shown clinical activity in relapsed/refractory (R/R) B-cell lymphoma. The phase 1 CITADEL-112 (NCT03424122) study assessed safety and efficacy of parsaclisib in combination with investigator choice standard of care (SOC; rituximab [Treatment A], rituximab plus bendamustine [Treatment B], or ibrutinib [Treatment C]) in 50 patients with R/R B-cell lymphoma. The most common treatment-emergent adverse events included neutropenia (62.5%, 50.0%, and 50.0% of patients in Treatments A, B, and C, respectively); diarrhea (37.5%) and anemia (31.3%) in Treatment A; abdominal pain, asthenia, diarrhea, and nausea (each 33.3%) in Treatment B; and increased alanine and aspartate aminotransferase (each 37.5%) in Treatment C. Objective responses were observed in 13 patients (81.3%) in Treatment A, 10 (55.6%) in Treatment B, and 8 (50.0%) in Treatment C. Parsaclisib combined with SOC therapies had an expected safety profile and promising efficacy in patients with R/R B-cell lymphomas.

Keywords:

• Bendamustine
• ibrutinib
• non-Hodgkin lymphoma
• parsaclisib
• rituximab

Introduction

Non-Hodgkin lymphomas (NHLs) account for 4-5% of all new cancer cases in the United States [1,2]. Approximately 85-90% of NHLs are B-cell–derived [3]; the most common B-cell NHLs include diffuse large B-cell lymphoma (DLBCL), follicular lymphoma (FL), marginal zone lymphoma (MZL), and mantle cell lymphoma (MCL) [1,3]. Rituximab-based therapy or chemoimmunotherapy regimens have remained the backbone of front-line treatment of aggressive NHL subtypes (DLBCL and MCL) and for treatment of advanced indolent disease (FL and MZL) [4]. Rituximab single-agent treatment is indicated for relapsed or refractory (R/R), low-grade or follicular, CD20-positive B-cell NHL, and in combination with first-line chemotherapy for previously untreated follicular, CD20-positive B-cell NHL and as single-agent maintenance therapy in patients achieving a complete or partial response to rituximab in combination with chemotherapy [5]. Bendamustine plus rituximab has been investigated for treatment of relapsed FL and MCL [6], and may be an option for first-line treatment of advanced FL and MZL, less aggressive front-line induction therapy for MCL, or second-line treatment of R/R MCL if not previously given [4]. Many subtypes of lymphoma are not curable with currently available therapies and require further treatment for R/R disease [1].

Different therapeutic approaches have been investigated to improve responses with currently available treatments for patients with B-cell NHLs. However, results have been underwhelming, with immunostimulatory agents demonstrating limited impact on clinical outcomes with rituximab [7,8]. Further understanding of the biology of B-cell lymphomas has led to development of novel therapies with potential to improve outcomes for patients with R/R disease. Potential treatments for R/R B-cell NHLs beyond first- or early-line standard of care (SOC) include targeted therapies such as inhibitors of Bruton’s tyrosine kinase (BTK; e.g. ibrutinib, acalabrutinib, zanubrutinib), phosphoinositide 3-kinase (PI3K) (e.g. idelalisib, duvelisib, copanlisib, umbralisib), enhancer of zeste homolog 2 (e.g. tazemetostat), bispecific antibodies (e.g. CD20/CD3 agents include epcoritamab, glofitamab, mosunetuzumab, odronextamab), and chimeric antigen receptor T-cell therapies (e.g. axicabtagene ciloleucel, brexucabtagene autoleucel, lisocabtagene maraleucel, tisagenlecleucel) [4,9–11].

PI3Kδ inhibitors have shown promise in clinical studies for patients with B-cell NHLs [12–14]. Parsaclisib is a highly selective and potent PI3Kδ inhibitor, designed to minimize hepatotoxicity associated with early-generation inhibitors [15]. Parsaclisib monotherapy demonstrated antitumor activity in a phase 1/2 study in patients with R/R B-cell lymphoma [16]. In a preliminary safety analysis of CITADEL-112 (NCT03424122), an open-label phase 1 study of parsaclisib in combination with investigator choice SOC (rituximab, rituximab plus bendamustine, or ibrutinib) in patients with R/R B-cell lymphoma, manageable tolerability was observed [17]. Here we report final safety and efficacy results from the CITADEL-112 study.

Methods

Study design

CITADEL-112 is a phase 1, open-label, dose-finding study that evaluated safety and tolerability of parsaclisib in combination with investigator choice (rituximab, rituximab plus bendamustine, or ibrutinib) in patients with previously treated B-cell lymphoma (Figure S1). The study was conducted in accordance with the Declaration of Helsinki, Good Clinical Practice guidelines, and all applicable country-specific regulations. The study protocol was approved by institutional review boards or independent ethics committees, and patients provided written informed consent before enrollment.

Using a 3 + 3 design, patients were enrolled into 1 of 3 treatment groups, as assigned by the investigator: patients received parsaclisib 20 mg once daily (QD) orally for 8 weeks followed by 20 mg once weekly (QW) in combination with either: rituximab 375 mg/m² intravenously (IV) QW for 4 doses (days 1, 8, 15, and 22 of cycle 1) with up to an additional 4 QW doses at the discretion of the investigator (days 1, 8, 15, and 22 of cycle 2) (Treatment A); rituximab 375 mg/m² IV on day 1 plus bendamustine 90 mg/m² IV on days 1 and 2 of each 28-day cycle for ≤6 cycles (Treatment B); or ibrutinib 560 mg QD orally (Treatment C). Enrollment of ≥15 patients in each treatment group was necessary to establish safety, and up to 27 patients were required in each treatment group to confirm tolerability and determine the recommended phase 2 dose of parsaclisib.

De-escalation or expansion of parsaclisib dose was carried out within each combination regimen independently and was based on the occurrence of any dose-limiting toxicities (DLTs) during the initial 28-day administration of each combination treatment. Dose interruptions and modifications for rituximab, bendamustine, and ibrutinib were at the discretion of the investigator, as specified by individual product package inserts [5,18,19] and applicable treatment guidelines. Patients continued to receive combination treatment in continuous 28-day cycles until evidence of disease progression, unacceptable toxicity, or consent withdrawal. There was no defined date for end of treatment. All patients were required to receive a standard Pneumocystis jirovecii pneumonia (PJP) prophylaxis regimen while receiving parsaclisib and for ≥2-6 months after the last dose of parsaclisib.

Patient population

Patients older than 18 years of age with histologically confirmed indolent (FL, MZL) or aggressive (DLBCL, MCL) lymphoma were enrolled. Eligible patients had a life expectancy of >3 months, had received ≥1 (≥2 for FL) prior systemic therapies with documented progression or documented failure to achieve complete response (CR) or partial response (PR) after the most recent systemic treatment, and were ineligible for stem cell transplantation. Patients must have had an Eastern Cooperative Oncology Group performance status of 0-2; adequate hematologic, hepatic, and renal function; and willing to undergo an incisional or excisional lymph node or tissue biopsy, or provide a lymph node or tissue biopsy from the most recent available archival tissue.

Key exclusion criteria included evidence of transformed NHL (except FL); NHL subtypes not specified in inclusion criteria; history of primary or metastatic central nervous system lymphoma or leptomeningeal disease; prior treatment with other selective PI3Kδ inhibitors (including idelalisib) or pan-PI3K inhibitors; autologous or allogeneic stem cell transplant within 3 or 6 months before study initiation, respectively; receipt of immunosuppressive therapy, anticancer, or investigational drugs within protocol-defined intervals; or chronic or active infection requiring treatment (including human immunodeficiency virus, hepatitis B or C virus infection).

Patients to be treated with bendamustine (Treatment B) were excluded if they had received bendamustine in the preceding 12 months. Patients treated with bendamustine >12 months before study initiation were eligible if they had not discontinued due to tolerability concerns, had achieved a CR or PR of ≥12 months before relapse/progression, and had experienced progression following an alkylating agent-containing regimen. Patients to be treated with ibrutinib (Treatment C) were excluded if they had received any prior BTK inhibitor; were currently receiving warfarin, warfarin-based anticoagulants, or vitamin K antagonists; or had a known bleeding disorder.

Study assessments and endpoints

The primary endpoint of the study was safety and tolerability of each treatment combination, determined by monitoring the frequency, duration, and severity of adverse events (AEs), including evaluation of DLTs. A DLT was defined as any AE, regardless of investigator attribution of relatedness, that was new or worsening in severity (graded using Common Terminology Criteria for Adverse Events v4.03) (Table S1). For patients with grade ≥3 diarrhea, colonoscopies with biopsies could be considered but were not a protocol requirement. Additional measures of safety included vital signs, physical examinations, 12-lead electrocardiograms, and chemistry and hematologic laboratory assessments.

Secondary endpoints were parsaclisib pharmacokinetic (PK) parameters after oral dose administration. In Treatments A and B, blood samples were collected on day 8 of cycle 1 and day 1 of cycle 2 within 30 min predose, at 1-2 h (±15 min) postdose, and 4-8 h (±15 min) postdose. In Treatment C, blood samples were collected on days 1, 8, and 15 of cycle 1 within 30 min predose, and at 1 h (±10 min), 2 h (±30 min), 4 h (±30 min), and 6-8 h (±30 min) postdose. Parsaclisib plasma concentrations were measured using a validated liquid chromatography with tandem mass spectrometry method with an assay range of 5-5000 nM [20].

Exploratory endpoints included investigator-assessed best objective response, objective response rate (ORR), duration of response (DOR), and progression-free survival (PFS). Response status for each patient was categorized at each tumor assessment visit as complete metabolic response (CMR)/CR, partial metabolic response (PMR)/PR, no metabolic response/stable disease, progressive metabolic disease/progressive disease, or not evaluable (NE)/not assessed, according to response criteria for lymphomas using fluorodeoxyglucose–positron emission tomography (PET) or combined PET–computed tomography scan [21].

Statistical analyses

A total sample size of 81 patients was planned for this study. Up to 3 dose levels of parsaclisib could be evaluated as part of the 3 + 3 design before the recommended dose was expanded to an additional 9 patients. Therefore, 15-27 patients were required for each of the 3 combination regimens. The treatment of 15 patients in each treatment group was expected to provide an approximate 90% chance of observing ≥1 DLT with a true event rate of 15%. All patients who received ≥1 dose of study treatment comprised the safety population (used for all safety analyses) and the full analysis set (used for summary of demographics, baseline characteristics, patient disposition, and efficacy analyses). The PK-evaluable population included all patients who received ≥1 dose of study treatment and provided minimum 1 plasma sample for PK.

As this was an exploratory study, no formal statistical tests were performed. For safety analyses, all treatment-emergent AEs (TEAEs) were tabulated and recorded according to the Medical Dictionary for Regulatory Activities preferred term and system organ class. Descriptive summaries for continuous and categorical data were reported for AE data, including number (percentage) of patients reporting any TEAE, grade ≥3 TEAEs, serious AEs (SAEs), treatment-related TEAEs, fatal TEAEs, or TEAEs leading to either treatment interruption, dose reduction, or discontinuation. For efficacy analyses, best overall response was summarized, and ORR was estimated with 95% confidence interval (CI) calculated using the exact method for binomial distributions. DOR and PFS, and respective medians and 95% CIs, were estimated by the Kaplan-Meier method. Target lesion sizes were measured by the sum of product diameters for multiple lesions, and the percentage change from baseline was calculated.

For PK analyses, standard noncompartmental PK analysis methods were applied to analyze plasma concentration data for Treatment C using WinNonlin v8.3.4 (Certara USA Inc., Princeton, NJ). Descriptive summaries and statistical analyses were prepared in SAS v9.4 (SAS Inc, Cary, NC). Evaluation of the effect of ibrutinib on parsaclisib exposure was performed by analysis of variance (ANOVA) on log-transformed PK parameters by comparison of 95% CIs around the geometric mean ratio after coadministration of parsaclisib with ibrutinib versus parsaclisib alone. Noncompartmental analysis was not performed for Treatments A and B, and PK parameters were not calculated, because only 2 postdose plasma samples were collected; summary statistics on plasma concentrations were prepared.

Results

Patient demographics and disposition

Between July 5, 2018, and June 27, 2022, a total of 50 patients were enrolled and treated with parsaclisib in combination with either rituximab (n = 16, Treatment A), rituximab plus bendamustine (n = 18, Treatment B), or ibrutinib (n = 16, Treatment C). Patients were enrolled at sites in Italy, Spain, and the United States.

Patient baseline characteristics are summarized in Table 1. The median age (range) for each treatment group was 70.5 (53-82) years (Treatment A), 65.0 (46-82) years (Treatment B), and 71.0 (59-82) years (Treatment C). The most common disease subtypes were FL in Treatment A and DLBCL in Treatment B and C. Most patients had received ≥2 prior systemic therapies (87.5%, Treatment A; 61.1%, Treatment B; 68.8%, Treatment C); 8 (50.0%) and 10 patients (62.5%) in Treatments A and C, respectively, had received prior treatment with bendamustine; 4 (25.0%) and 3 patients (16.7%) in Treatments A and B, respectively, had previously received ibrutinib. Of the 7 patients with previous exposure to ibrutinib, 6 had MCL (4 in Treatment A; 2 in Treatment B) and 1 had DLBCL (Treatment B). Most patients enrolled in the study had advanced disease (Ann Arbor Stage III-IV).

Table 1. Summary of patient baseline demographics and clinical characteristics.

Characteristic	Treatment A Parsaclisib + Rituximab (n = 16)	Treatment B Parsaclisib + Rituximab + Bendamustine (n = 18)	Treatment C Parsaclisib + Ibrutinib (n = 16)
Age, median (range), years	70.5 (53-82)	65.0 (46-82)	71.0 (59-82)
Male, n (%)	10 (62.5)	9 (50.0)	11 (68.8)
Disease subtype, n (%)
DLBCL	2 (12.5)	7 (38.9)	6 (37.5)
GCB	2 (12.5)	1 (5.6)	2 (12.5)
Non-GCB	0 (0)	2 (11.1)	0 (0)
ABC	0 (0)	2 (11.1)	1 (6.3)
Double-hit*	0 (0)	1 (5.6)	0 (0)
Unknown	0 (0)	1 (5.6)	3 (18.8)
FL	6 (37.5)	3 (16.7)	3 (18.8)
MCL	5 (31.3)	5 (27.8)	2 (12.5)
MZL	3 (18.8)	3 (16.7)	5 (31.3)
Race, n (%)
White	16 (100.0)	18 (100.0)	15 (93.8)
Other^†	0 (0)	0 (0)	1 (6.3)
Time since initial diagnosis, median (range), years	6.9 (0.7-21.1)	3.4 (0.2-24.2)	2.7 (0.6-15.9)
Number of prior systemic therapy regimens, median (range)	3 (1-4)	2 (1-4)	2 (1-7)
≥2 prior systemic therapy regimens, n (%)	14 (87.5)	11 (61.1)	11 (68.8)
Prior therapies, n (%)
Radiation	4 (25.0)	3 (16.7)	7 (43.8)
Surgery	1 (6.3)	2 (11.1)	2 (12.5)
HSCT	5 (31.3)	4 (22.2)	2 (12.5)
Selected prior therapies
Rituximab	16 (100.0)	17 (94.4)	16 (100.0)
Cyclophosphamide	14 (87.5)	18 (100.0)	13 (81.3)
Vincristine	13 (81.3)	12 (66.7)	9 (56.3)
Doxorubicin	12 (75.0)	13 (72.2)	10 (62.5)
Prednisone	9 (56.3)	14 (77.8)	9 (56.3)
Bendamustine	8 (50.0)	0 (0)	10 (62.5)
Ibrutinib	4 (25.0)	3 (16.7)	0 (0)
ECOG PS, n (%)
0	10 (62.5)	6 (33.3)	9 (56.3)
1	6 (37.5)	11 (61.1)	7 (43.8)
2	0 (0)	1 (5.6)	0 (0)
Ann Arbor stage, n (%)
I / II	0 (0) / 1 (6.3)	0 (0) / 3 (16.7)	0 (0) / 2 (12.5)
III / IV	5 (31.3) / 10 (62.5)	3 (16.7) / 12 (66.7)	4 (25.0) / 10 (62.5)

ABC: activated B-cell type; DLBCL: diffuse large B-cell lymphoma; ECOG PS: Eastern Cooperative Oncology Group performance status; FL: follicular lymphoma; GCB: germinal center B-cell type; HSCT: hematopoietic stem cell transplant; MCL: mantle cell lymphoma; MZL: marginal zone lymphoma.

*MYC and BCL2. ^†Other race type excludes White, Asian, Black, or African American.

The median (range) overall duration (QD + QW) of parsaclisib treatment was 5.1 (0.5-22.9), 9.3 (0.5-38.7), and 9.3 (0.6-37.7) months in Treatments A, B, and C, respectively. The median (range) dose of QD parsaclisib was similar between Treatments A (20.0 [13.3-20.0] mg/day), B (17.9 [2.8-20.0] mg/day), and C (20.0 [7.1-20.0] mg/day). All patients enrolled finished or withdrew from the study; most common reasons for treatment discontinuation were progressive disease (75.0%, 44.4%, and 56.3%) and AEs (12.5%, 16.7%, and 6.3%) in Treatments A, B, and C, respectively.

Safety

Across all treatment groups, all patients experienced ≥1 TEAE (Table 2). One patient receiving parsaclisib in combination with rituximab plus bendamustine experienced a hematological DLT of grade 4 neutropenia, which lasted 14 days.

Table 2. Summary of any-grade TEAEs (≥25%) and corresponding grade ≥3 TEAEs in each treatment group.

Preferred term, n (%)	Any grade	Grade ≥3
Treatment A: parsaclisib + rituximab (n = 16)	16 (100.0)	12 (75.0)
Neutropenia	10 (62.5)	8 (50.0)
Diarrhea	6 (37.5)	3 (18.8)
Anemia	5 (31.3)	0 (0)
Blood creatinine increased	4 (25.0)	0 (0)
Treatment B: parsaclisib + rituximab + bendamustine (n = 18)	18 (100.0)	16 (88.9)
Neutropenia	9 (50.0)	8 (44.4)
Abdominal pain	6 (33.3)	0 (0)
Asthenia	6 (33.3)	0 (0)
Diarrhea	6 (33.3)	1 (5.6)
Nausea	6 (33.3)	0 (0)
Pyrexia	5 (27.8)	0 (0)
Rash	5 (27.8)	1 (5.6)
Treatment C: parsaclisib + ibrutinib (n = 16)	16 (100.0)	10 (62.5)
Neutropenia	8 (50.0)	4 (25.0)
ALT increased	6 (37.5)	2 (12.5)
AST increased	6 (37.5)	2 (12.5)
Anemia	5 (31.3)	0 (0)
Diarrhea	5 (31.3)	2 (12.5)
Pyrexia	5 (31.3)	0 (0)
Stomatitis	5 (31.3)	0 (0)
Thrombocytopenia	5 (31.3)	0 (0)
Blood creatinine increased	4 (25.0)	0 (0)
Cough	4 (25.0)	0 (0)

ALT: alanine aminotransferase; AST: aspartate aminotransferase; TEAE: treatment-emergent adverse event.

The most common any-grade TEAE across all treatment groups was neutropenia, which was reported in 10 (62.5%), 9 (50.0%), and 8 patients (50.0%) in Treatments A, B, and C, respectively (Table 2). In Treatment A, 1 patient had prolonged grade 1 neutropenia for 47 days, 1 patient had grade 2 neutropenia for 57 days, and 1 patient had ongoing neutropenia at data cutoff (grade 4 for 14 days, grade 2 for 14 days, and ongoing grade 3), approximately 2 years from date of onset. In Treatment B, 1 patient had prolonged grade 1 neutropenia for 56 days and 1 patient had grade 4 neutropenia for 34 days. In Treatment C, 1 patient had prolonged neutropenia for 112 days (grade 1 for a total of 93 days and grade 3 for 19 days), 1 patient had grade 2 neutropenia for 32 days, 1 patient had ongoing grade 1 neutropenia, and 1 patient had ongoing grade 3 neutropenia at data cutoff, approximately 3 years from date of onset. Other commonly reported any-grade TEAEs among all treatment groups included diarrhea (6 [37.5%]) and anemia (5 [31.3%]) in Treatment A; abdominal pain, asthenia, diarrhea, and nausea (each 6 [33.3%]) in Treatment B; increased alanine aminotransferase (ALT) and increased aspartate aminotransferase (AST; each 6 [37.5%]), and anemia, diarrhea, pyrexia, stomatitis, and thrombocytopenia (each 5 [31.3%]) in Treatment C (Table 2). Colitis was observed in 3 patients during the study period. In Treatment A, 1 patient had colitis, which resolved in 48 days. In Treatment B, 1 patient had concurrent Clostridium difficile colitis and amebic colitis, which resolved in 18 days, and 1 patient experienced colitis, which resolved in 9 days. COVID-19 infections occurred in a total of 3 patients—in Treatment A, 2 female patients (82 and 71 years old) were reported to have COVID-19 infections on day 428 and day 201 of parsaclisib treatment, respectively; in Treatment C, 1 male patient (64 years old) was reported to have COVID-19 infection on day 471 of parsaclisib treatment. The most common grade ≥3 TEAEs occurring in ≥15% of patients were neutropenia, which was reported in 8 (50.0%), 8 (44.4%), and 4 (25.0%) patients in Treatments A, B, and C, respectively, and diarrhea (3 [18.8%]) in Treatment A (Table 2). TEAEs related to parsaclisib were reported in 14 patients (87.5%) in Treatment A, 13 patients (72.2%) in Treatment B, and all patients in Treatment C (Table S2).

Serious TEAEs were experienced by 6 patients (37.5%) in Treatment A, 6 patients (33.3%) in Treatment B, and 8 patients (50.0%) in Treatment C (Table 3). The only serious TEAEs reported in >1 patient across treatment groups were COVID-19, diarrhea, and pneumonia (n = 2 each) in Treatment A, and atrial fibrillation (n = 2) in Treatment C. Serious TEAEs related to parsaclisib occurred in 3 patients (18.8%) each in Treatments A and C, and 3 patients (16.7%) in Treatment B (Table 3). TEAEs with a fatal outcome occurred in 2 patients in Treatment A (COVID-19 and COVID-19 pneumonia [n = 1 (82 years old); not parsaclisib related per investigator], interstitial lung disease [n = 1 (81 years old); parsaclisib related per investigator]), and 1 patient in Treatment C (64 years old; COVID-19 pneumonia, septic shock, acute kidney injury [parsaclisib and ibrutinib related per investigator], and respiratory failure [parsaclisib related per investigator]).

Table 3. Summary of serious TEAEs and parsaclisib-related serious TEAEs in each treatment group.

Preferred term, n (%)	Serious TEAE	Parsaclisib-related serious TEAE
Treatment A: parsaclisib + rituximab (n = 16)	6 (37.5)	3 (18.8)
COVID-19	2 (12.5)	0 (0)
Diarrhea	2 (12.5)	2 (12.5)
Pneumonia	2 (12.5)	0 (0)
Abdominal sepsis	1 (6.3)	1 (6.3)
COVID-19 pneumonia	1 (6.3)	0 (0)
Interstitial lung disease	1 (6.3)	1 (6.3)
Pneumococcal infection	1 (6.3)	0 (0)
Pneumonia streptococcal	1 (6.3)	0 (0)
Renal hemorrhage	1 (6.3)	0 (0)
Spinal column injury	1 (6.3)	0 (0)
Treatment B: parsaclisib + rituximab + bendamustine (n = 18)	6 (33.3)	3 (16.7)
Amebic colitis	1 (5.6)	0 (0)
Clostridium difficile colitis	1 (5.6)	0 (0)
Colitis	1 (5.6)	1 (5.6)
Device-related infection	1 (5.6)	0 (0)
Gastroenteritis	1 (5.6)	1 (5.6)
Gastrointestinal hemorrhage	1 (5.6)	0 (0)
Pneumonia	1 (5.6)	1 (5.6)
Renal failure	1 (5.6)	1 (5.6)
Toxic skin eruption	1 (5.6)	1 (5.6)
Deep vein thrombosis	1 (5.6)	0 (0)
Treatment C: parsaclisib + ibrutinib (n = 16)	8 (50.0)	3 (18.8)
Atrial fibrillation	2 (12.5)	1 (6.3)
Acute kidney injury	1 (6.3)	1 (6.3)
Alanine aminotransferase increased	1 (6.3)	1 (6.3)
Aspartate aminotransferase increased	1 (6.3)	1 (6.3)
COVID-19 pneumonia	1 (6.3)	1 (6.3)
Dermatitis exfoliative	1 (6.3)	1 (6.3)
Diarrhea	1 (6.3)	0 (0)
Neutropenia	1 (6.3)	1 (6.3)
Pericardial effusion	1 (6.3)	0 (0)
Pneumonia	1 (6.3)	0 (0)
Respiratory failure	1 (6.3)	1 (6.3)
Septic shock	1 (6.3)	1 (6.3)
Squamous cell carcinoma of skin	1 (6.3)	0 (0)

TEAE: treatment-emergent adverse event.

TEAEs leading to discontinuation of parsaclisib were reported in 2 patients in Treatment A (diarrhea, streptococcal pneumonia [each n = 1]), 3 patients in Treatment B (neutropenia, maculopapular rash [each n = 1], diarrhea and renal failure [n = 1]), and 1 patient in Treatment C (thrombocytopenia). TEAEs leading to parsaclisib dose interruption occurred in 12 (75%), 12 (66.7%), and 9 (56.3%) patients in Treatments A, B, and C, respectively. Of these TEAEs, events reported in ≥1 patient across treatment groups were neutropenia (n = 6), diarrhea (n = 4), and COVID-19 (n = 2) in Treatment A; neutropenia (n = 7), asthenia, and rash (each n = 2) in Treatment B; neutropenia (n = 3), pneumonia, ALT increased, and AST increased (each n = 2) in Treatment C. TEAEs leading to parsaclisib dose reduction were reported in 3 (18.8%) patients each in Treatments A and C, and 5 (27.8%) in Treatment B.

Pharmacokinetics

Parsaclisib mean plasma concentration profiles at cycle 1 day 8 (Figure S2A) and cycle 2 day 1 (Figure S2B) were similar after coadministration with rituximab in Treatment A, or rituximab plus bendamustine in Treatment B. The PK exposures of parsaclisib at cycle 1 day 15 (parsaclisib alone) versus cycle 1 day 8 (parsaclisib plus ibrutinib) in Treatment C were also similar (Figure S2C); a summary of parsaclisib PK parameters with Treatment C is included in Table S3. ANOVA of log-transformed parsaclisib PK data was performed to assess a potential effect of ibrutinib on parsaclisib exposure. After coadministration of parsaclisib with ibrutinib versus parsaclisib alone, calculated 95% CIs included unity for maximum and minimum observed plasma concentration, and area under the plasma concentration-time curve from time zero to the last measurable concentration, indicating that ibrutinib had no impact on parsaclisib exposure (Table S3).

Efficacy

Best overall response by treatment group and lymphoma subtype are summarized in Table 4. In total, 3 (18.8%), 6 (33.3%), and 2 (12.5%) patients achieved a CMR/CR in Treatments A, B, and C, and PMR/PR was observed in 10 (62.5%), 4 (22.2%), and 6 (37.5%) patients in Treatments A, B, and C, respectively. Best percentage changes from baseline in target lesions are shown in Figure 1A–C. The median (range) percentage change from baseline in target lesions was −77.8% (−100.00 to 19.0%) for Treatment A, −94.9% (−100.0 to 75.5%) for Treatment B, and −65.9% (−91.2 to 118.6%) for Treatment C.

Figure 1. Waterfall plot of best percentage change in sum of target lesions for combination of parsaclisib with (A) rituximab (Treatment A), (B) rituximab plus bendamustine (Treatment B), and (C) ibrutinib (Treatment C). *Patient with best percentage change >100%.

Table 4. Summary of best overall response and objective response by treatment group.

	DLBCL	FL	MCL	MZL	Total
Treatment A: parsaclisib + rituximab (n = 16)	(n = 2)	(n = 6)	(n = 5)	(n = 3)	(n = 16)
Best overall response, n (%)
CMR/CR	1 (6.3)	0 (0)	1 (6.3)	1 (6.3)	3 (18.8)
PMR/PR	1 (6.3)	6 (37.5)	1 (6.3)	2 (12.5)	10 (62.5)
NMR/SD	0 (0)	0 (0)	1 (6.3)	0 (0)	1 (6.3)
PMD/PD	0 (0)	0 (0)	1 (6.3)	0 (0)	1 (6.3)
NE/NA*	0 (0)	0 (0)	1 (6.3)	0 (0)	1 (6.3)
ORR, n (%) [95% CI]: 13 (81.3) [54.4-96.0]
Treatment B: parsaclisib + rituximab + bendamustine	DLBCL	FL	MCL	MZL	Total
(n = 18)	(n = 7)	(n = 3)	(n = 5)	(n = 3)	(n = 18)
Best overall response, n (%)
CMR/CR	1 (5.6)	2 (11.1)	3 (16.7)	0 (0)	6 (33.3)
PMR/PR	0 (0)	1 (5.6)	0 (0)	3 (16.7)	4 (22.2)
NMR/SD	1 (5.6)	0 (0)	0 (0)	0 (0)	1 (5.6)
PMD/PD	3 (16.7)	0 (0)	1 (5.6)	0 (0)	4 (22.2)
NE/NA*	2 (11.1)	0 (0)	1 (5.6)	0 (0)	3 (16.7)
ORR, n (%) [95% CI]: 10 (55.6) [30.8-78.5]
	DLBCL	FL	MCL	MZL	Total
Treatment C: parsaclisib + ibrutinib (n = 16)	(n = 6)	(n = 3)	(n = 2)	(n = 5)	(n = 16)
Best overall response, n (%)
CMR/CR	0 (0)	1 (6.3)	1 (6.3)	0 (0)	2 (12.5)
PMR/PR	2 (12.5)	2 (12.5)	1 (6.3)	1 (6.3)	6 (37.5)
NMR/SD	0 (0)	0 (0)	0 (0)	4 (25.0)	4 (25.0)
PMD/PD	3 (18.8)	0 (0)	0 (0)	0 (0)	3 (18.8)
NE/NA*	1 (6.3)	0 (0)	0 (0)	0 (0)	1 (6.3)
ORR, n (%) [95% CI]: 8 (50.0) [24.7-75.3]

*No post-baseline response data were available.

CI: confidence interval; CMR: complete metabolic response; CR: complete response; DLBCL: diffuse large B-cell lymphoma; FL: follicular lymphoma; MCL: mantle cell lymphoma; MZL: marginal zone lymphoma; NA: not assessed; NE: not evaluable; NMR: no metabolic response; NR: not reached; ORR: objective response rate; PD: progressive disease; PMD: progressive metabolic disease; PMR: partial metabolic response; PR: partial response; SD: stable disease.

Objective responses were observed in 13 of 16 patients (81.3%) in Treatment A, 10 of 18 patients (55.6%) in Treatment B, and 8 of 16 patients (50.0%) in Treatment C (Table 4), with median DORs for each treatment group reported as 5.2 months (95% CI: 1.7-15.8), not reached (95% CI: 12.1-NE), and 13.4 months (95% CI: 0.7-NE), respectively (Figure 2A–C). Similarly, median PFS was estimated as 5.5 months (95% CI: 3.0-14.6), not reached (95% CI: 2.4-NE), and 13.8 months (95% CI: 1.8-NE) for Treatments A, B,+ and C, respectively (Figure 2).

Figure 2. Kaplan-Meier estimates of DOR and PFS for combination of parsaclisib with (A) rituximab (Treatment A), (B) rituximab plus bendamustine (Treatment B), and (C) ibrutinib (Treatment C). CI: confidence interval; DOR: duration of response; NE: not evaluable; PFS: progression-free survival.

Discussion

Parsaclisib, a potent and selective inhibitor of PI3Kδ, has previously been investigated as monotherapy in a phase 1/2 study in patients with R/R B-cell NHL (NCT02018861) [16], and in phase 2 clinical trials in patients with R/R DLBCL (NCT02998476), FL (NCT03126019), MZL (NCT03144674), and MCL (NCT03235544) [22–25]. The current phase 1 CITADEL-112 study evaluated safety and efficacy of parsaclisib 20 mg QD for 8 weeks followed by 20 mg QW in combination with inves­tigator choice SOC therapies (rituximab, rituximab plus bendamustine, or ibrutinib). A total of 50 patients were enrolled, the majority received ≥2 prior sys­temic therapies, and most had late-stage disease at baseline.

No new safety concerns were identified for any of the combination treatments evaluated in CITADEL-112. A DLT of grade 4 neutropenia lasting 14 days, which exceeded the DLT criteria of ≥7 days, was reported in 1 patient receiving parsaclisib in combination with rituximab plus bendamustine. Neutropenia and diarrhea were the most common TEAEs reported in this study, consistent with previous clinical trials of parsaclisib monotherapy [16,22–25] and studies with other PI3Kδ inhibitors [12,26]. Pneumonia also occurred in all 3 treatment groups, which is not unexpected with PI3K inhibitor therapy. Transaminitis has been reported as a TEAE with early-generation PI3K inhibitors such as idelalisib and duvelisib [12,13,26–29]. In our study, grade ≥3 increased ALT and AST were only reported in the parsaclisib plus ibrutinib treatment group (each 12.5%); no grade ≥3 events of increased ALT or AST were reported in the treatment groups receiving parsaclisib in combination with rituximab, or rituximab plus bendamustine.

Early-generation PI3K inhibitors are associated with AEs, including neutropenia and serious or fatal infections, and immune-mediated AEs such as diarrhea, colitis, hepatitis, pneumonitis, and dermatologic toxicities [12,26,30]. Approaches to balance efficacy and safety to optimize clinical use of PI3K inhibitors have included dose-evaluation studies, and novel dosing schedules such as intermittent dosing [12,26,30]. The phase 1/2 CITADEL-101 study implemented a modified dosing schedule and showed improved long-term tolerability with parsaclisib given 20 mg daily for 9 weeks and weekly thereafter, when compared to continuous daily parsaclisib [16]. Based on data from CITADEL-101, a dosing schedule of parsaclisib 20 mg daily for 8 weeks followed by 20 mg weekly was selected for our study. AEs of special interest included colitis, febrile neutropenia, and PJP infection; colitis of grade 2 occurred in 1 patient each in Treatments A and B, and febrile neutropenia of grade 3 occurred in 1 patient in Treatment B. There were no events of PJP infection in any treatment group.

Parsaclisib PK exposure was investigated in combination with the 3 investigator choice treatment regimens. Bendamustine does not affect the PK of parsaclisib, because parsaclisib concentration profiles were comparable across Treatments A and B. The PK exposures were similar after administration of parsaclisib with ibrutinib versus parsaclisib alone in Treatment C, suggesting no effect of ibrutinib on parsaclisib at the investigational doses.

Most studies reporting efficacy for parsaclisib have investigated monotherapy treatment in patients with R/R B-cell lymphomas; in phase 2 studies, observed ORRs range from 25.5% (95% CI: 14.7-39.0) in patients with DLBCL to 65.2% (95% CI: 42.7-83.6) in patients with FL receiving weekly parsaclisib dosing [22–25]. In a phase 1 study evaluating parsaclisib in combination with bendamustine plus obinutuzumab in patients with R/R FL (CITADEL-102, NCT03039114), investigator-assessed ORR of 76.9% (95% CI: 56.4-91.0) was observed with 65.4% of patients achieving CR/CMR [31]. In the current study, ORRs were 81.3% (95% CI: 54.4-96.0) for combination of parsaclisib plus rituximab, 55.6% (95% CI: 30.8-78.5) for parsaclisib with rituximab plus bendamustine, and 50.0% (95% CI: 24.7-75.3) when parsaclisib was combined with ibrutinib. Together, results from the CITADEL-102 and CITADEL-112 studies suggest that parsaclisib in combination with SOC therapies may provide clinical benefit for patients with R/R B-cell lymphomas. Other PI3K inhibitors (copanlisib, duvelisib, idelalisib) have also been investigated for treatment of R/R NHL in combination with either rituximab [32–34] or bendamustine plus rituximab [32–34]. However, there are currently no PI3K inhibitors recommended or indicated in combination with other therapies for treatment of R/R B-cell lymphomas [4].

CITADEL-112 was an exploratory study with no formal statistical tests performed. Moreover, the number of patients enrolled in each treatment group was low, thus, conclusions that can be drawn from results reported herein are limited.

Recently, due to concerns over reduced overall survival and increased toxicity in patients with NHL or chronic lymphocytic leukemia as well as the feasibility of confirmatory studies, US Food and Drug Administration approval for umbralisib was withdrawn for R/R FL and MZL, and idelalisib approvals for R/R FL and small lymphocytic leukemia and duvelisib approval for R/R FL were removed [30,35–38]. These concerns will need to be addressed in future development of PI3K inhibitors.

Conclusions

Overall, this study showed that parsaclisib combined with SOC therapies had an expected safety profile, and demonstrated promising efficacy in patients with R/R B-cell lymphomas. However, larger studies with increased patient numbers would be necessary to determine the role and confirm the benefit relative to safety of PI3K inhibitors in the treatment landscape for R/R B-cell lymphomas.

Authors’ contributions

J-MS, PA, AK, RC, MT, and AL-G acquired the data. PL, ER, and TL analyzed the data. All authors contributed to interpretation of data, drafting and/or critical review of the manuscript, and provided approval of the final version to be published.

Acknowledgments

The authors wish to thank the patients and their families, the investigators, and the site personnel who participated in this study. This study was sponsored by Incyte Corporation (Wilmington, DE, USA). Medical writing assistance was provided by Matthew Bidgood, Envision Pharma Group (Fairfield, CT, USA) and funded by Incyte Corporation.

Disclosure statement

J-MS reports consultancy, honoraria, and membership on an entity’s board of directors or advisory committees for Celgene, Gilead, Incyte, Kern Pharma, Novartis, and Roche; honoraria from Bristol Myers Squibb, Janssen, and Takeda; consultancy and membership on an entity’s board of directors or advisory committees for BeiGene, Celltrion, Eli Lilly & Company, Miltenyi Biomedicine, and Sandoz. PA reports consultancy, honoraria, and membership on an entity’s board of directors or advisory committees for AbbVie, AstraZeneca, Bristol Myers Squibb, Janssen, and Roche; speakers’ bureau for AbbVie, Bristol Myers Squibb, and Janssen; honoraria from Sandoz; and consultancy and membership on an entity’s board of directors or advisory committees for Incyte. RC reports consultancy, honoraria, and speakers’ bureau for AbbVie, AstraZeneca, Janssen, and Roche; consultancy for BeiGene, Eli Lilly & Company, GenMab, Incyte, and Takeda; consultancy and speakers’ bureau for Kite; consultancy and honoraria from Bristol Myers Squibb; research funding from Pfizer; and honoraria from Celgene and Gilead. PL, ER, and TL report employment and stock ownership for Incyte. AL-G reports membership on an entity’s board of directors or advisory committees for Celgene, Gilead Sciences, Incyte, and Roche; and research funding from Roche. AK and MT have no disclosures to report.

Data availability statement

Incyte Corporation (Wilmington, DE, USA) is committed to data sharing that advances science and medicine while protecting patient privacy. Qualified external scientific researchers may request anonymized datasets owned by Incyte Corporation for the purpose of conducting legitimate scientific research. Researchers may request anonymized datasets from any interventional study (except phase 1 studies) for which the product and indication have been approved on or after January 1, 2020 in at least 1 major market (eg, United States, European Union, Japan). Data will be available for request after the primary publication or 2 years after the study has ended. Information on Incyte Corporation’s clinical trial data sharing policy and instructions for submitting clinical trial data requests are available at: https://www.incyte.com/Portals/0/Assets/Compliance%20and%20Transparency/clinical-trialdata-sharing.pdf?ver=2020-05-21-132838-960.

Additional information

Funding

This study was supported by Incyte Corporation (Wilmington, DE, USA).