Citations

Introduction: Bruton tyrosine kinase (BTK) inhibitors are a class of drugs that inhibit B-cell receptor (BCR) and are increasingly used in B-cell lymphoproliferative neoplasms, including chronic lymphocytic leukemia (CLL), mantle cell lymphoma (MCL), and Waldenstrom's macroglobulinemia. Ibrutinib, a first-generation BTK inhibitor has been associated with increased risk of cardiovascular adverse events (AEs), including atrial fibrillation (AF), hypertension (HTN) and bleeding. These unique AEs are thought to be due to off-target effects. Acalabrutinib, a second-generation BTK inhibitor is characterized by less off-target effects, and is thought to be associated with a decreased risk of cardiovascular and other AEs. However, a head to head comparison of ibrutinib and acalabrutinib has not been conducted. Herein, we conducted a systematic review and network meta-analysis of AEs from prospective clinical trials of ibrutinib and acalabrutinib in B-cell lymphoproliferative disorders to compare their safety profile. Methods: We searched PubMed, Embase, Scopus, and Web of Science from database inception through November 15th 2019. Only full-text articles were included. Other inclusion criteria included prospective trials (single arm or randomized) with ibrutinib, ibrutinib plus anti-CD20 antibody, or acalabrutinib as investigational agents. Trials investigating BTK inhibitor plus chemotherapy were excluded. When updated results of prospective trials were available, data were extracted from the most recent publication with the longest follow-up. Reports of 17 AEs of interest, including number of events (any grade and grade 3 or higher) were documented. Rate of discontinuation was investigated. Results: Twenty-seven prospective clinical trials, 12 multicenter single-arm, 9 multicenter randomized, 5 single center single-arm, and 1 single center randomized, were included. Data from 29 study arms including 3207 patients were analyzed in 3 groups - ibrutinib, ibrutinib plus anti-CD20 antibody, and acalabrutinib with augmented Bayesian network meta-analysis and meta-regression implemented in R including packages gemtc and rjags. The most common any grade AEs (>20%) with ibrutinib were diarrhea (46%, 95% CI 36-55%), myalgias/arthralgias (37%, 95%CI 28-46%), fatigue (33%, 95% CI 24-42%), cough (26%, 95% 17-36%), anemia (23%, 95% 15-30%), thrombocytopenia (22%, 95% 15-30%), and pyrexia (21%, 95% 13-30%). The most common any grade AEs with acalabrutinib were headache (37%, 95%CI 26-48%), diarrhea (30%, 95% 20-41%), peripheral edema (21%, 95% 15-28%), fatigue (20%, 95% 11-29%), and myalgias/arthralgias (16%, 95% 8-24%). The most common any grade cardiovascular AEs with ibrutinib were bleeding/bruising (32%, 95% 23-41%), HTN (23%, 95% 15-32%), AF (9%, 95% 3-15%). The most common any grade cardiovascular AEs with acalabrutinib were bleeding/bruising (41%, 95% CI 30-52%), and HTN (6%, 95% 1-11%). The rate of AEs with ibrutinib compared to ibrutinib plus anti-CD20 antibody were similar so the data was pooled. Of all AEs of interest, there was a significant difference in any grade AEs favoring ibrutinib for headache (12% vs. 37%), and infections (35% vs 57%). There was a significant difference in any grade AEs favoring acalabrutinib for myalgias/arthralgias (16% vs. 37%), anemia (6% vs. 23%), thrombocytopenia (5% vs. 22%), and HTN (6% vs. 23%). After adjusting for median follow-up and age, there was no significant difference in rates of bleeding/bruising and any grade infections between ibrutinib and acalabrutinib. However, there was a significant difference favoring acalabrutinib for any grade HTN (OR 0.26, 95% CI 0.17-0.40) p<0.0001, grade 3 HTN (OR 0.15, 95% 0.08-0.27) p<0.0001, any grade AF (OR 0.35, 95% 0.18-0.66), p=0.0012, grade 3 AF (OR 0.04, 95% 0.01-0.25) p=0.0009, and grade 3 infections (OR 0.62, 95% 0.46-0.85),p=0.003. Conclusions: Acalabrutinib appears to have an overall improved safety profile compared to ibrutinib. This is particularly true for anemia, thrombocytopenia, and cardiovascular AEs, including AF and HTN. Figure 1 Disclosures No relevant conflicts of interest to declare.

70 Background: The prevalence of health insurance literacy (HIL) and financial literacy (FL) in cancer patients is not well described. Further, a better understanding of the association between these constructs and the domains of financial hardship (FH) can guide potential interventions. Methods: We examined the prevalence of HIL and FL and FH in cancer patients on active treatment at our institution. HIL and FL were assessed using Health Insurance Literacy Measure (HILM) and 5 questions from the National Financial Capability Study respectively. COST FACIT and National Health Interview Survey (NHIS) questions were used to examine overall FH and its 3 domains (material, behavioral and psychological hardship). Pearson correlation coefficient was used to describe the relation between the COST score and FH domains assessed by NHIS questions. Logistic regression analysis was performed to determine the associations of overall FH (measured by COST FACIT) with FL and HIL. Results: Among 256 patients approached, 202 completed questionnaires. Median age was 67 (range 22 to 91) years. 51% were male, 93% Whites and 5% were Hispanic. 69% had a solid tumor with 37% stage IV. Median number of lines of treatment was 4 (range 1 to 18). 21% were currently employed, 59% had ≥college degree, 41% had private insurance and 42% had monthly income≥$7000. 50% participants answered at least 4/5 FL questions correctly, indicating high FL. Being male, White, having insurance, higher education and income ≥$7000 were associated with higher FL. 72% of participants had high HIL, and they were more likely to be Non-Hispanic, married and have received more lines of treatment. FL significantly correlated with HIL (r = 0.29, p < .001). 65% patients endorsed FH based on NHIS questions (10% behavioral, 20% material, and 62% psychological hardship). 48% patients had FH based on a COST score < 28 (Median score), which correlated with material (r = -0.59, p < .001), psychological (r = -0.55, p < .001) and behavioral FH (r = -0.42, p < .001). The association between FH and HIL was not significant (OR 0.78; p = 0.08), though higher FL was associated with lower FH (OR 0.69; p = 0.02). Conclusions: Despite high FL and HIL (as compared to national standards), the prevalence of FH in our relatively affluent patient cohort was high, driven mainly by psychological hardship. We did not find strong association between the HIL and FH unlike in the reported literature. Our findings behoove us to develop tailored interventions to address unique domains of the FH in a diverse population.

2057 Background: To date, a comprehensive evaluation of core limitations in clinical trials leading to anti-cancer drug approvals by the US Food and Drug Administration (FDA) has not been undertaken.The aim of this analysis was to assess the percentage of clinical trials with core limitations, defined as lack of randomization, lack of overall survival data, inappropriate use of crossover, and use of sub-optimal control arms that led to FDA approvals from 2014 to 2019. Methods: This observational analysis included all approved anti-cancer drug indications by the FDA from July 2014 through July 2019. All indications were investigated and each clinical trial evaluated for design, enrollment period, primary endpoints, and presence of core limitations. The standard of care therapy was determined by evaluating the literature and published guidelines 1-year prior to start of clinical trial enrollment. Crossover was examined and evaluated for optimal use. We then calculated the percentage of approvals based on clinical trials with any or all core limitations. Results: A total of 187 anti-cancer approvals were evaluated. The number of anti-cancer drug approvals doubled over time with 68 in first half of study period (June 2014 to December 2016) to 119 in second half of study period (January 2017 to July 2019). Of those, 125 (67%) were based on a clinical trial with at least one core limitation. 64 (34%) approvals were based on a single-arm clinical trial. Of the remaining 123 approvals based on randomized trials, 60 (32%) had a core limitation. Of all randomized trials, 37 (30%) lacked overall survival benefit, 31 (25%) had a sub-optimal control, and 17 (14%) used crossover inappropriately. Conclusions: The majority of cancer drugs are approved based on clinical trials with core limitations. Efforts to minimize core limitations at the time of clinical trial design are essential to ensure that new anti-cancer drugs being marketed truly improve patient outcomes over current standards.

AbstractOver the last 5 years, there have been more than 140 new drug approvals in the field of Oncology alone, all based on newly published clinical trials. These approvals have led to an ongoing change in clinical practice, offering new therapeutic options for patients. Therefore, it is important for healthcare providers to be able to appraise a clinical trial and determine its validity, understand its results, and be able to apply such results to their patients. In this guide, we provide a simplified approach tailored to practicing clinicians and trainees. The same concepts and principles apply to other medical specialties.

Abstract Introduction: Primary central nervous system lymphoma (PCNSL) is a rare intracranial lymphoma that accounts for less than 1% of all non-Hodgkins lymphomas and 3% of all brain tumors. Histopathologically, approximately 90% of PCNSL cases are categorized as a diffuse large B-Cell lymphoma (DLBCL). DLBCL malignancies are subdivided by Cell of Origin (COO), with the vast majority of PCNSL categorized as non-germinal center B cell (non-GCB) by immunohistochemistry. Gene expression profiling (GEP), however, has shown that immunohistochemically defined non-GCB resolves into two distinct subtypes, namely activated B-cell (ABC) and unclassified (UNC) subtypes. Using the Lymph2Cx molecular COO subtyping assay, we have found that 91% of PCNSL are ABC (unpublished data). Unlike systemic-DLBCL, PCNSL is largely confined to and rarely metastasizes outside of the immune privileged central nervous system. Despite this, PCNSL is one of the most aggressive forms of DLBCL. Given the immune privileged milieu in which PCNSL arises, we hypothesized that this milieu elicits a transcriptional profile that contributes to the enhanced aggressive nature of PCNSL compared to systemic-DLBCL. To investigate this hypothesis, this study assessed the gene expression differences between ABC-PCNSL and ABC-systemic-DLBCL, in order to identify novel players in the pathogenesis of ABC-PCNSL. Methods: A total of 35 HIV negative samples, with proven ABC-subtype COO as per the GEP Lymph2Cx assay, were employed; including 10 ABC systemic-DLBCL and 25 primary ABC PCNSL cases with no concurrent or prior history of systemic DLBCL. Samples were reviewed by a hematopathologist to confirm diagnoses and determine tumor content. Samples with <60% tumor content were macro-dissected before nucleic acid extraction, which was performed using the Qiagen AllPrep DNA/RNA FFPE Kit. Extracted DNA and RNA were quantified using the Qubit HS-kit and NanoDrop respectively. Digital gene expression technology was used to perform the PanCancer Pathways panel (NanoString, Seattle, WA). Differential gene expression analysis was performed using the NanoString specific statistical method NanoStringDiff. Identified gene sets were analyzed using the online Gene Set Enrichment Analysis (GSEA) Molecular Signatures Database (MSigDB). Results: Of the 739 cancer related genes targeted by the PanCancer panel, 256 were found to be significantly differentially expressed in the ABC-PCNSL cohort compared to the ABC-DLBCL cohort (p<0.05). Fifty six genes were upregulated and 200 were downregulated. With a 4.9 fold change, the most significantly overexpressed gene was FGF1 (p=4.7E-11). FGF1 encodes a primary ligand for the fibroblast growth factor receptors (FGFR) -1, -2, -3 and -4; of which, FGFR2 (p=1.0E-7) and FGFR3 (p=0.003) were also significantly overexpressed. Moreover, MSigDB identified the FGF signaling pathway as enriched in the upregulated gene set (5 genes, p=7.4E-9, FDR=6.6E-7). FGFRs are a family of receptors that activate known mitogenic signaling pathways including MAPK signaling, which MSigDB identified as the most enriched pathway in the upregulated gene set (14 genes, p=1.65E-19, FDR=2.2E-16). MSigDB analysis of the 200 down regulated genes revealed that 5 of the top 20 enriched signaling pathways were immune related and included Signaling by interleukins (26 genes, p=2.9E-38, FDR=3.2E-36), Immune cytokine signaling (31 genes, p=1.1 E-34, FDR=1.1E-32), chemokine signaling (28 genes, p=1.5E-34, FDR=1.4E-32), T-cell receptor signaling (24 genes, p=2.4E-34, FDR=1.9E-32) and Toll-like receptor signaling (23 genes, p=4.1E-33, FDR=3.0E-31). Conclusions: We show, for the first time, that ABC-PCNSL and ABC-systemic-DLBCL possess significantly different transcriptional profiles despite identical, molecularly determined, COO status. A principle difference between these DLBCL malignancies is their anatomical location related immune privilege status which is reflected as reduced immune related signaling in the CNS-DLBCL cohort and may have important mitogenic signaling implications. Indeed, the results suggest that the enhanced aggressive nature of PCNSL compared to systemic-DLBCL is mediated, at least in part, by enhanced FGF signaling; a pathway with known roles in cell survival and proliferation. Disclosures Rimsza: NanoString: Other: Inventor on the patent for the Lymph2Cx assay.

Abstract Objectives: Primary central nervous system lymphoma (PCNSL) is a rare aggressive B-cell lymphoma, particularly in HIV-negative individuals, that represents a clinical challenge due to its location and lack of comprehensive molecular and biologic description. Histopathologic features are that of diffuse large B-cell lymphoma with expression of pan-B-cell markers as well as cell of origin (COO) germinal center B cell (GCB) and post-germinal center B cell (non-GCB) markers. Previous studies using immunohistochemistry (IHC) suggest that the majority of PCNSL cases are non-GCB. However, gene expression profiling has revealed non-GCB to be comprised of two distinct subtypes, namely activated B-cell (ABC) and unclassified (UNC) subtypes that are indistinguishable by IHC. To date COO testing using the highly accurate Lymph2Cx gene expression profiling assay has not been reported in PCNSL. Methods: IRB approval was obtained and HIV negative patients diagnosed with PCNSL, who had given informed consent and for whom archived tumor tissue was available for testing were identified. COO testing was performed using the Lymph2Cx NanoString assay on RNA extracted from formalin-fixed, paraffin embed tissues using our established laboratory protocols. Clinical data including patient demographics, lines of treatment and survival outcomes were collected and correlated with each other and Lymph2Cx COO results. Results: Thirty-two HIV-negative patients diagnosed between January 2005 and June 2015 were included. Median age was 61 years (30-82) and 53% were male. Radiographic information was available for 18/32 patients. Eleven (61%) had a single brain lesion at diagnosis, while 7 (39%) had >1 brain lesion. Lines of systemic therapy were 1 (91%) and 2 (9%). All patients received methotrexate-based induction therapy (44% received methotrexate, rituximab and temozolomide, 16% received single-agent methotrexate, 31% received methotrexate and rituximab, and 9% received the modified Bonn regimen; methotrexate and cytarabine-based induction). A total of 10 patients (31%) received high-dose chemotherapy and autologous stem cell transplant (ASCT) for consolidation. Of the 10 patients that underwent consolidation therapy, 9 underwent ASCT after first line induction, and 1 underwent ASCT after second line therapy. None of the patients received whole-brain radiation therapy. At a median follow-up of 29 months (range of 2-107) median event-free survival (EFS) was 16.3 months (95% CI, 8.8-45.7), and median overall survival (OS) was 41.2 months (95% CI, 28.5-NE). COO testing using the Lymph2Cx assay revealed that 91% (29/32) were ABC, 9% (3/32) were GCB, and none were UNC. Histopathology reports described COO using the Hans algorithm in 11 of the 32 cases. Of the 3 determined to be GCB on Lymph2Cx, 1 was denoted GCB by the Hans algorithm and 2 were not stained to determine COO. Of the 29 determined to be ABC by Lymph2Cx, 9 were denoted non-GCB and 1 was denoted GCB by the Hans algorithm, and 19 were not stained to determine COO. Conclusions: This series of HIV-negative patients with PCNSL showed median survival consistent with previous studies. In this first series using the Lymph2Cx assay, we confirmed that over 90% of PCNSLs are of ABC subtype, which concurs with previous reports that PCNSL tumors are predominately non-GCB by the Hans algorithm. These findings provide biological rationale forthat pharmacologic interventions targeting B-cell receptor signaling to be explored in clinical trials in the majority of patients with PCNSL. Figure. Figure. Disclosures Rimsza: NanoString: Other: Inventor on the patent for the Lymph2Cx assay.