Phase 1 dose-escalation study of SEA-CD40: a non-fucosylated CD40 agonist, in advanced solid tumors and lymphomas

BACKGROUND

SEA-CD40 is an investigational, non-fucosylated, humanized monoclonal IgG1 antibody that activates CD40, an immune-activating tumor necrosis factor receptor superfamily member. SEA-CD40 exhibits enhanced binding to activating FcγRIIIa, possibly enabling greater immune stimulation than other CD40 agonists. A first-in-human phase 1 trial was conducted to examine safety, pharmacokinetics, and pharmacodynamics of SEA-CD40 monotherapy in patients with advanced solid tumors and lymphoma.

METHODS

SEA-CD40 was administered intravenously to patients with solid tumors or lymphoma in 21-day cycles with standard 3+3 dose escalation at 0.6, 3, 10, 30, 45, and 60 µg/kg. An intensified dosing regimen was also studied. The primary objectives of the study were to evaluate the safety and tolerability and identify the maximum tolerated dose of SEA-CD40. Secondary objectives included evaluation of the pharmacokinetic parameters, antitherapeutic antibodies, pharmacodynamic effects and biomarker response, and antitumor activity.

RESULTS

A total of 67 patients received SEA-CD40 including 56 patients with solid tumors and 11 patients with lymphoma. A manageable safety profile was observed, with predominant adverse events of infusion/hypersensitivity reactions (IHRs) reported in 73% of patients. IHRs were primarily ≤grade 2 with an incidence associated with infusion rate. To mitigate IHRs, a standardized infusion approach was implemented with routine premedication and a slowed infusion rate. SEA-CD40 infusion resulted in potent immune activation, illustrated by dose dependent cytokine induction with associated activation and trafficking of innate and adaptive immune cells. Results suggested that doses of 10-30 µg/kg may result in optimal immune activation. SEA-CD40 monotherapy exhibited evidence of antitumor activity, with a partial response in a patient with basal cell carcinoma and a complete response in a patient with follicular lymphoma.

CONCLUSIONS

SEA-CD40 was tolerable as monotherapy and induced potent dose dependent immune cell activation and trafficking consistent with immune activation. Evidence of monotherapy antitumor activity was observed in patients with solid tumors and lymphoma. Further evaluation of SEA-CD40 is warranted, potentially as a component of a combination regimen.

TRIAL REGISTRATION NUMBER

NCT02376699.

Characterization of cytokine release syndrome (CRS) and neurological events (NEs) in the phase I TRANSCEND NHL 001 trial of lisocabtagene maraleucel (liso-cel) for patients (pts) with relapsed/refractory (R/R) large B-cell lymphoma (LBCL)

e20046

Background: Clinical studies of liso-cel show low incidences of severe CRS and NEs. Pts with high tumor burden and inflammation are at higher risk of CRS and NEs (Borrega, Hemasphere 2019). Here, we characterize the presenting symptoms, timing, and management of CRS and NEs in pts with R/R LBCL treated with liso-cel in TRANSCEND NHL 001 (NCT02631044). A deeper understanding of CRS/NEs after liso-cel treatment may help clinicians identify and manage these toxicities. Methods: Pts with R/R LBCL and ≥2 lines of therapy received liso-cel after lymphodepletion (LD) with fludarabine and cyclophosphamide. Bridging therapy was allowed for pts with PET-positive disease before LD (Abramson, ASH 2019 #241). Investigators were educated on prospectively identifying liso-cel–related NEs and CRS, which were collected and graded per NCI CTCAE v4.03 (NEs) or 2014 Lee criteria (CRS). Timing and severity of individual CRS/NE symptoms and interventions were also collected. Results: The analysis included 269 pts (median age, 63 y): 38% had sum of perpendicular diameters ≥50 cm2 or lactate dehydrogenase ≥500 U/L, and 59% received bridging therapy. CRS and NEs had delayed onset (median, 5 and 9 d, respectively) and a low incidence of grade (Gr) ≥3 events (CRS 2%; NE 10%). CRS occurred before NEs in most pts and was Gr 1/2 at onset in all but 2 pts (Table). The most common CRS symptoms were pyrexia (40%), hypotension (20%), and tachycardia (18%). Overall, 20% of pts received tocilizumab and/or corticosteroids for CRS (10% tocilizumab only; 2% corticosteroids only; 8% both); 3% received vasopressors. The most common NE symptoms were confusional state (11%), tremor (9%), and aphasia (8%); most were low grade. A total of 17% of pts received tocilizumab and/or corticosteroids for NEs (13% corticosteroids only; < 1% tocilizumab only; 3% both); < 1% received vasopressors. Overall, 4% of pts were admitted to the ICU for CRS and/or NEs. Additional analyses on timing of intervention and type/severity of initial symptoms will be presented. Conclusions: In pts with high-risk, aggressive R/R LBCL, liso-cel treatment was associated with a low incidence of severe CRS/NEs, late onset of mostly low-grade events at presentation, and low use of tocilizumab/corticosteroids. Clinical trial information: NCT02631044 . [Table: see text]

Use of a patient-preferences shared-decision-making encounter tool in clinical practice for patients with non-Hodgkin’s lymphoma

e24192

Background: Non-Hodgkin lymphoma (NHL) is the fourth leading cause of cancer in the United States with 77,240 new cases and 19,940 deaths annually. Treatment options are numerous and driven by patient’s molecular profile, risk, preferences/goals, and ability to tolerate treatment. Aligning physician-patient goals of care and integrating patient preferences into a shared-decision making (SDM) model allows patients and providers to select treatment consistent with medical science and personalized to each patient. This project evaluates feasibility of a patient preferences (PP) SDM encounter tool using technology to facilitate SDM at treatment decision (TD) for patients with NHL. Methods: To date, 45 patients with NHL at a TD making point were recruited from two sites to complete a tablet-based PPSDM encounter tool. The tool includes questions about needs, decision making preferences, values and goals of care. Results are reviewed by the provider and used to facilitate SDM in treatment selection during the clinical encounter. Patients also completed measures to assess satisfaction with the TD, patient activation, and perceived achievement of desired role in SDM at 3 weeks and 3 months post TD. Results: Participants are mean age 66 years (range 23-86), 53% male, and 98% white. 47% (n = 45) preferred that their doctor share responsibility with them when deciding which treatment was best for them. 69% said they would like to make the TD together with family and close friends and 69% agreed that their spouse was the most important person in helping make a TD. 51% said spirituality did not play a part in their TD. When asked how they liked to get medical information, 67% said they wanted all the facts, but not the prognosis. 87% said they had identified a medical surrogate to make decisions, yet 31% had an advanced directive on file. 64% agreed their cancer was curable and 84% agreed that a treatment goal was to get rid of all cancer. 73% of providers used the PPSDM results in conversation with the patient and 53% indicated their patient management changed based on the PPSDM results. There was 24% concordance between patient and provider perception of how TD were made. Conclusions: Collecting patient preferences, values, and care goals prior to the clinical visit using technology is feasible in busy clinics. Although most providers used the PPSDM results and over half changed their management plan, there was low concordance between patient and provider perceptions. Final analysis will include 3 week and 3 month measures of patient activation and satisfaction.

MYC+ relapsed and refractory (R/R) diffuse large b-cell lymphoma (DLBCL): Impact of additional hits and outcomes with subsequent therapy

7541

Background: Translocations involving MYC are a hallmark of poor prognosis among patients with newly diagnosed DLBCL. The impact of MYC translocations with or without additional “hits” involving BCL2 or BCL6 in response to salvage therapy and survival in R/R DLCBL is not well defined. Methods: We performed a multicenter retrospective study of 176 patients with R/R DLBCL failing to achieve CR or relapsing within 6 months after completion of upfront chemoimmunotherapy and for whom FISH information on MYC, BCL2 and BCL6 was available. The objectives were to examine the response to salvage therapy, utilization of hematopoietic cell transplantation (HCT) and survival outcomes in MYC- (n = 120), MYC+ single hit (SH, n = 28), and MYC+ double hit (DH, n = 36) R/R DLBCL. Results: Overall response rate to first salvage therapy and utilization of HCT was comparable between the 3 cohorts (Table). 2-year OS was 0% in MYC+ SH, 8.8% in MYC+ DH and 29.9% in MYC- cases (p = 0.001) without difference in OS between SH and DH (P = 0.8). The higher risk of death for MYC+ SH (HR 1.79, 95% C.I. 1.03-3.11, P = 0.03) and MYC+ DH (HR 1.93, 95% C.I. 1.23-3.00, P = 0.004) persisted after adjustment for covariates. For patients who underwent auto-HCT, 2-year OS was 0% in MYC+ SH, 29.3% in MYC+ DH and 55.4% in MYC- cases (p < 0.001) without significant difference between SH and DH (P = 0.8). All 4 MYC+patients who underwent allo-HCT relapsed in < 4 months. Conclusions: MYC+ R/R DLBCL have similar response to salvage therapy than the MYC- counterparts but dismal survival irrespective of additional “hits” and even if HCT can be performed. MYC+ R/R DLBCL represents an unmet medical need and should be prioritized for clinical trials with novel agents and innovative cellular therapies. [Table: see text]

Improving shared decision making in chronic lymphocytic leukemia through multidisciplinary education

154

Background: New treatments for chronic lymphocytic leukemia (CLL) with excellent response rates and varying toxicity profiles have emerged in recent years, creating an opportunity for a patient’s personal preferences to contribute to treatment decisions. Methods: We conducted a prospective, quasi-experimental pre- and post-evaluation of a multi-level educational program and its impact on knowledge of CLL as well as on shared decision-making. We educated patients, patient navigators, and providers (nurses and physicians) using a combination of print, video, in-person lectures, and online case-based education tailored to each learner group. Patients were evaluated for change in patient activation, distress levels, desired role in decision-making, perception of decision-making, satisfaction with oncologist explanation of treatment choice, and knowledge of CLL. Patient navigators and providers were evaluated for change in CLL knowledge and perception of shared decision-making. Results: Forty-one patients and 87 providers participated in the educational program. We observed trends toward improved patient activation with 68% before education vs. 76% after education reporting a high (score of 3 or 4) Patient Activation Measure (PAM) . The percentage of patients desiring and perceiving shared decision-making trended upward from 47% to 67% and 35% to 49%, respectively. There was a trend toward an increase in the percentage of patients understanding that CLL is incurable (80% to 90%) and reporting awareness of signs of progression (64% to 76%). Patients satisfaction with their oncologist’s explanation of therapy increased significantly from 83% to 95% (p = 0.03). Conclusions: This modestly sized project demonstrated trends toward improvements in patient engagement, prognostic awareness, knowledge of signs of progression, and shared decision-making; these promising findings should be tested in larger samples. However, there remains an opportunity for further improvement in shared decision-making.

Whole-exome sequencing (WES) of penile squamous cell carcinoma (PSCC) to identify multiple recurrent mutations

484

Background: Molecular alterations and drivers of PSCC, an orphan malignancy, remain unclear. The Cancer Genome Atlas is not studying PSCC and the Catalogue of Somatic Mutations in Cancer has performed targeted analyses only. We report WES of PSCC tumors from a group of patients (pts). Methods: Freshfrozen macrodissected PSCC tumor tissue and adjacent normal tissue samples were procured from the Cooperative Human Tissue Network. DNA was isolated from tissue sections by phenol chloroform extraction. Exome capture was performed with the Agilent SureSelect clinical research exome kit and whole exome-seq was done on the Illumina HiSeq2500 with paired end 100bp chemistry. Raw sequence data in Fastq format were aligned to human reference genome and quantified, and compared by using a local instance of Galaxy (galaxy.uabgrid.uab.edu). These data were analyzed for mutations (SNPs) analysis, by Partek Genomic Suite/Flow(PGS, Partek, St. Louis, MO) for variance calling against human reference genome (hg19) as referenced to dbSNP; and copy number variants (cnv) by FishingCNV tool together with picard tools/samtools/GATK). We focused on missense mutations and amplifications among ≥ 2 tumor samples but not in normal samples as they may cause upregulation of gene/protein function, which may be therapeutically actionable. Results: PSCC tumors were available from 11 patients and adjacent normal tissue from 3 patients. The 10 most common genes with > 4 missense mutations among ≥ 2 tumor samples overall were the following in decreasing order of frequency: MUC4, HLA-DPA1, MUC16, XIRP2, SSPO, TTN, FCGBP, PABPC3, ALPK2 and MKI67. The top upstream transcriptional regulators were PIH1D3, PRDM5, PTK2, Coup-Tf and NBEAL2. When examining candidate actionable genes, recurrent missense alterations were seen in PIK3C2A and PIK3C2G. Additional analysis will study alterations in functional domains and cnv. Conclusions: WES identified a relatively high mutation burden in PSCC withrecurrent missense mutations in multiple genes, notably including the PI3K gene among potentially actionable genes. Validation of these findings and further study of downstream effects is required.

Multiplatform comprehensive kinase analysis of penile squamous cell carcinoma (PSCC) to identify drivers and potentially actionable therapeutic targets

389

Background: PSCC is an orphan malignancy with poorly understood biology and suboptimal systemic therapy. Given that kinases may be drivers of disease and readily actionable, we performed comprehensive multiplatform analysis of kinases in PSCC tumor tissue and adjacent normal tissue. Methods: We collected fresh frozen tumors from 10 patients with PSCC, of whom 3 had adjacent normal tissue available. After macrodissection to demarcate tumor from normal tissue, the samples underwent analysis of kinases using platforms to assess DNA, RNA and kinase activity. Next Generation Sequencing of 515 kinase genes was performed using the Agilent Kinome capture and run on the Illumina HiSeq2500 at PE100bp. The Nanostring nCounter platform analyzed the expression (RNA) of 519 kinase genes. Global kinase activity of tissue lysates was measured using Pamstation@12 high-content phospho-peptide substrate microarray system (PamGene International). Upstream kinase prediction was performed and network mapping was done with GeneGo MetaCore. Ingenuity pathway analysis data was performed to integrate over-expression at the activity and gene expression level with coexisting missense mutations at DNA level. Results: Top pathways observed to be upregulated in both the kinase activity and gene expression platforms were PTEN Signaling, STAT3 Pathway, GNRH Signaling, IL8 Signaling and B Cell Receptor Signaling. Potentially relevant missense mutations were seen in 176 kinase genes, with the top altered pathways overlapping with gene overexpression being GNRH Signaling, NF-kB Signaling, and STAT3 pathways. ERBB2, ERBB3 and SYK were altered on gene sequencing and also showed elevated kinase activity. Kinases previously implicated in prostate cancer (PAK4, TNK2), breast cancer (ERBB2) and genital dysplasia (ROR2) also demonstrated mutations. Conclusions: Multiplatform comprehensive analysis of kinases at the DNA, RNA and protein activity levels discovered several potential drivers of PSCC and actionable therapeutic targets. Further validation in preclinical and translational trials is warranted to make advances in this rare malignancy with substantial unmet needs.

Integrated comprehensive high-throughput kinomics profiling and whole exome sequencing of penile squamous cell cancer (PSCC)

383

Background: Molecular drivers in penile squamous cell cancer (PSCC), an orphan malignancy, remain unclear. The Cancer Genome Atlas (TCGA) is not studying PSCC and the Catalogue of Somatic Mutations in Cancer (COSMIC) investigators have reported only targeted analyses of PSCC. We report the first integrated analyses of comprehensive kinomics and whole exome sequencing (seq) in tumors from patients (pts) with PSCC . Methods: We performed integrated functional kinomics profiling and comprehensive exome-seq of two frozen tissue samples from men with PSCC with a matched normal tissue procured from the Cooperative Human Tissue Network (CHTN). Kinomic profiling was performed using the PamStation 12 high-content phospho-peptide substrate microarray system (PamGene International). The protein tyrosine kinome and serine/threonine kinome PamChips were used to measure global kinase activity by detecting phosphorylation of various peptides through FITC-labeled antibodies. Upstream kinase prediction was performed using a scoring algorithm that incorporates the phosphonet database (www.phosphonet.ca). Exome capture was performed with the Agilent SureSelect v5 kit and whole exome-seq was done on the Illumina HiSeq2000 with paired end 100bp chemistry. Results: In the single patient, paired kinomics analysis comparing the tumor sample to adjacent normal tissue, the HER family (EGFR, ERBB2, 3 and 4), AXL, TYRO3 and SYK kinases were the most active. When combining the two tumors in an unpaired analysis against the normal sample, the HER (EGFR, ERBB2, 3 and 4), MER, FRK, and FAK, kinases showed increased activity. When comparing whole exome-seq of the two PSCC samples with normal, among the affected genes were CCDC181, ZNF717, MUC4, HGC6.3, NOTCH1, STK11, SIRPB1, SKA3, PDE6B, FAT1, CACNA2D1, USP17L11, MNT, and CEP89. We are evaluating 10 PSCC tumors and matched normal tissue by kinomics and whole exome-seq and will present these complete data and analysis at the conference. Conclusions: In our preliminary analysis of pts that underwent the first reported integrated kinomics and whole exome-seq performed in PSCC, we identified multiple potential therapeutic targets in tumors.

Comprehensive kinase profiling to classify clear cell (cc)-renal cell carcinoma (RCC)

409

Background: Comprehensive high throughput functional kinase activity in localized cc-RCC tumors may assist in devising a classification system and help identify potential therapeutic targets. Methods: Patients (pts) with localized tumors undergoing surgery with minimum follow-up of 18 months underwent kinomics of fresh frozen cc-RCC tumors and matched normal renal tissue using the PamStation12 high-content phospho-peptide substrate microarray system (PamGene). The protein tyrosine kinome (PTK) and serine/threonine kinome (STK) PamChips were used to measure global kinase activity. Advanced network modeling of altered phospho-peptides was performed using MetaCore while upstream kinase prediction scoring was based off of phosphonet (phosphonet.ca). Both unsupervised analyses and supervised analyses of kinomics were done guided by tumor recurrence. Kinomics of tumor and matched normal tissue were compared. Results: Of 41 evaluable pts, the median age was 61 and pathologic stage was 1, 2, 3 and 4 in 19, 1, 20 and 1 pts, respectively. Unsupervised clustering analyses of tumor kinomics showed 3 groups: A (N=12), B (N=16), C (N=13). Potential driver kinases implicated were PFTK1, PKG1 and SRC in groups A, B, and C, respectively. Network modeling of these groups identified many Process Mappings including, but not limited to Inflammation pathways (A), translation initiation (B) and immune response and cell adhesion pathways(C). 5 of 9 pts who progressed were classified as Group C, 1 progressor was in Group B, and 3 were in Group A. Supervised analysis showed decreased CDK1, RSK1-4, ERK1-2, PKG2 and AKT2 kinase activity in those who progressed compared to others. 12 tumors showed increased PIM1 and MAPKAPK3, and decreased JNK2 and CDK1 compared to adjacent normal tissue. Conclusions: Comprehensive kinomics of localized cc-RCC was used to classify tumors based on unsupervised clustering, which appeared to confer differential long-term outcomes while supervised analysis also identified potential pathways related to recurrence. Kinases amplified in renal tumors compared to adjacent normal renal tissue were also identified. Significant heterogeneity of kinases was found with no single dominant kinase across all analyses.