Parikh, Sameer A.

Abstract Background Chronic lymphocytic leukemia (CLL) can effect renal function in a variety of ways including direct infiltration of the kidney, ureteral obstruction by lymphadenopathy, and treatment related tumor lysis syndrome (uric acid nephropathy). Rarely, CLL has also been reported to be associated with light chain nephropathy, renal amyloidosis, membranoproliferative glomerulonephritis (MPGN), granulomatous interstitial nephritis (GIN), and minimal change disease (MCD). Nearly all the data on the effects of CLL on renal function is at the case report level. We systematically evaluated the prevalence of renal insufficiency at diagnosis as well the incidence of acquired renal insufficiency during follow-up in a large cohort of patients with newly diagnosed CLL to more accurately define the effects of CLL on the kidney and its impact on clinical outcomes. Methods Between January 1995 -February 2013, previously untreated CLL patients seen in the Division of Hematology at Mayo Clinic at diagnosis (<12 months) and who had baseline assessment of serum creatinine were included in this analysis. Patients with serum creatinine (Cr) ≥1.5 mg/dL at baseline were classified as having renal insufficiency at diagnosis. Patients who initially had baseline creatinine <1.5 mg/dL but who developed a Cr≥1.5 mg/dL during the course of their disease were considered to have acquired renal insufficiency. Results Existing renal insufficiency at the time of CLL diagnosis: Of 2047 patients who met the eligibility criteria, 153 (7.5%) patients had renal insufficiency (Cr≥1.5 mg/dL) at the time of CLL diagnosis including 15 (0.7%) with a Cr≥3 mg/dL. Renal insufficiency was also more common among men (9.3% vs. 3.9%; p<0.00001), those with advanced stage disease (Rai 0=7.0%; Rai I-II=6.4%, Rai III-IV=20.2%; p<0.0001), and CD49d positive patients (6.8% vs. 3.8%; p<0.038). Patients with renal insufficiency at diagnosis were also older (median age 72.2 vs. 63.9; p<0.0001). No difference in the prevalence of renal insufficiency at diagnosis was observed based on cytogenetic abnormalities detected by FISH or CD38, ZAP-70 or IGHV gene mutation status. Although renal insufficiency at diagnosis was strongly associated with OS on univariate analysis (p<0.001), no association was observed between renal insufficiency and TTT or OS on multi-variate analysis adjusting for age, sex, and Rai stage. Acquired renal insufficiency during CLL disease course: Among the 1894 patients with normal renal function at diagnosis, 304 (16.1%) acquired renal insufficiency (Cr≥1.5 mg/dL) during the course of their CLL disease course including 43 (2.3%) with peak Cr≥3 mg/dL. In addition to age (older) and male sex, a number of CLL disease characteristics were associated with a higher likelihood of acquired renal insufficiency including: IGHV UM (OR=2.0; p=0.0001), unfavorable FISH (del17p- or 11q-; OR=2.0; p=0.001), and being CD49d+ (OR=1.8; p=0.002), ZAP-70+ (OR=1.6; p=0.004), or CD38+ (OR=1.4; p=0.0.032),. Shorter TTT (p<0.001) and OS (P<0.001) was observed among patients with initially normal creatinine who acquired renal insufficiency (Figure 1A and 1B). On MV analysis adjusting for age, sex, and stage at diagnosis, acquired renal insufficiency remained an independent predictor of TTT (OR=1.77; p=0.001) and OS (OR=2.67; p<0.001). Renal insufficiency and therapy selection After median follow-up of 4.5 years (range 0-18.0), 620 of 2047 (30.3%) patients have progressed to require treatment. Patients with renal insufficiency prior to treatment were less likely to receive purine nucleoside analogue based therapy and more likely to receive single agent alkylator based treatment. Conclusions Approximately 1 in every 13 patients (7.5%) with CLL has renal insufficiency at the time of diagnosis and an additional 16.1% acquire renal insufficiency during the course of the disease. The risk of developing renal insufficiency is associated with a variety of CLL B-cell characteristics and is associated with TTT and OS. Data on causes of acquired renal insufficiency is being abstracted and will be presented at the meeting. Disclosures: Shanafelt: Genentech: Research Funding; Glaxo-Smith-Kline: Research Funding; Cephalon: Research Funding; Hospira: Research Funding; Celgene: Research Funding; Polyphenon E International: Research Funding. Off Label Use: MK2206 in a phase 1 trial of CLL.

Abstract Introduction Approximately 2-8% chronic lymphocytic leukemia (CLL) patients develop diffuse large B-cell lymphoma (DLBCL, so-called Richter syndrome [RS]) over long-term follow-up. The outcomes of patients with RS are variable and poorly understood. Methods CLL patients seen at Mayo Clinic from January 1995 to April 2013 who developed biopsy-proven DLBCL during follow-up were included in this analysis. Clinical characteristics, therapy and overall survival (OS) of all patients with RS were analyzed. Results One hundred and twenty patients with biopsy-proven RS were identified in the Mayo Clinic database. The median age at diagnosis of RS was 67 years (range, 29-88 years), 87 (73%) were male, and 83 (69%) had received prior CLL therapy. The median duration between diagnosis of CLL and development of RS was 3.8 years (range, 0-34.5 years). Stereotyped B-cell receptors were found in 9/30 (30%) RS patients compared to 219/1294 (17%) non-RS patients (OR=2.1, p=0.07). IGHV4-39 gene usage was present in 2/26 (8%) RS patients compared to 57/1354 (4%) non-RS patients (p=0.3). Of the 30 RS biopsy samples in which Epstein Barr virus (EBV) in-situ hybridization studies were performed, EBV was detected in 7 (23%) patients. After a median follow-up of 5.2 years, 86 (72%) patients have died. Causes of death were as follows: progressive DLBCL (n=46, 53%), progressive CLL (n=2, 2%), infection (n=8, 9%), unknown (n=20, 23%) and other (n=10, 12%). The median OS of the entire cohort was 0.9 years (95% CI, 0.5-1.3 years). Tsimberidou and colleagues proposed a prognosis score (RS score) based on the following variables at the time of RS diagnosis: Eastern Co-operative Group performance status of 2-4, lactate dehydrogenase >1.5 x normal, platelets<100 x 109/L, tumour size >5 cm and >1 prior therapy for CLL. Each parameter was assigned one point, and patients were then assigned to one of four risk groups based on their calculated RS score: 0 or 1, low risk; 2, low-intermediate risk; 3, high-intermediate risk; 4 or 5, high risk. The OS according to the RS score for our group of patients is shown in Figure 1. Median survival in the low, intermediate-1, intermediate-2 and high-risk groups were 2.4 years, 0.7 years, 0.3 years and 0.2 years, respectively (p<0.001). Fifteen (12%) patients received cyclophosphamide, doxorubicin, vincristine and prednisone (CHOP) as initial therapy for RS, and 68 (57%) received rituximab in addition to CHOP (R-CHOP). There was no difference in survival among patients who received CHOP compared to R-CHOP (median, 1.8 years. vs. 1.2 years, p=0.5). Twenty-four patients (20%) received non-CHOP-based chemotherapy, and treatment was unknown in 4 (3%) patients. Nine patients (7%) died before definitive therapy could be instituted. Thirteen (11%) patients were able to proceed to stem cell transplantation (SCT): 10 (8%) autologous SCT, 2 (2%) allogeneic SCT and 1 (1%) autologous SCT followed by an allogeneic SCT. Although RS patients who underwent SCT were younger compared to those who did not undergo SCT (median, 59 years. vs. 69 years, p=0.001), no differences in prior CLL therapy or receipt of R-CHOP therapy for DLBCL was noted between the two groups. The median survival of patients who proceeded to SCT was superior compared to those who did not (5.2 years vs. 0.7 years; p=0.005). Conclusion Stereotyped B-cell receptors but not VH4-39 had a suggested evidence of an increased risk of RS in our cohort. Although 90% RS patients received intensive chemotherapy, only 11% patients were able to proceed to SCT. Median survival of RS patients remains dismal at 1 year, and newer therapies are needed to improve outcomes. Disclosures: Shanafelt: Genentech: Research Funding; Glaxo-Smith-Kline: Research Funding; Cephalon: Research Funding; Hospira: Research Funding; Celgene: Research Funding; Polyphenon E International: Research Funding.

Abstract Abstract 1040 Introduction: Hemophagocytic syndrome (HPS) is a rare disorder characterized by fever, splenomegaly, cytopenias, hypofibrinogenemia, hyperferritinemia and hemophagocytosis in the bone marrow or lymph nodes. It is caused by pathologic activation of the immune system, and manifests clinically with signs and symptoms related to extreme inflammation. Primary HPS or familial hemophagocytic lymphohistiocytosis (HLH) occurs predominantly in children (<18 years) due to inherited defects in natural killer and cytotoxic T-cell function. Secondary HPS generally occurs in adults (>18 years) and is precipitated by autoimmune disorders, malignancy and a variety of infections. The results of HLH-94 trial, the largest prospective study in pediatric patients with HLH, reported an estimated 5-year survival of 54%. There is a paucity of information about HPS in adult patients. Methods: Records of all adult (≥ 18 years) patients with a diagnosis of “hemophagocytosis“ or “hemophagocytic syndrome” presenting from January 1996 through December 2011 at Mayo Clinic were reviewed. Our hematopathology database was also analyzed for patients bearing a diagnosis of HPS. Out of a total of 127 patients identified in the database, 50 adult patients were diagnosed to have HPS as per the diagnostic criteria of the HLH-1994 trial (for patients seen prior to 2004), and HLH-2004 trial (for patients seen after 2004). This study was approved by the Institutional Review Board at Mayo Clinic. Results: Baseline characteristics of all patients are shown in Table 1. The median time to diagnosis from the onset of symptoms was 38 days (4–158 days). Secondary HPS was diagnosed in 46 (92%) patients. The causes of secondary HPS are shown in Table 2. Twenty (40%) patients were treated with etoposide containing regimen, 26 (52%) were treated with therapeutic agents directed specifically to the underlying precipitating disease, and 4 (8%) patients did not receive any treatment. No patient in our study underwent a stem cell transplant (SCT). After a median follow-up of 43 months, 36 (72%) patients have died. The causes of death were infection (47%), malignancy (42%), bleeding (8%) and unknown (3%). The median overall survival (OS) for the entire cohort was 1.5 months (95% CI, 0.7–4.5 months). The median OS for patients with malignancy- related HPS was 0.8 months (95% CI, 0.4–2.4 months) compared to 11 months (95% CI, 0.7-not reached) for patients with non-malignancy related HPS (p-value:0.006; Figure). On univariableanalysis, factors associated with a worse survival were serum albumin <2 g/dL (hazard ratio, 8.3), malignancy-related HPS (hazard ratio, 7.3) and serum creatinine >1.3 mg/dL (hazard ratio, 3.8). Conclusion: In this large series of adult HPS patients treated at a single tertiary care center, most cases are secondary, and are associated with a worse survival compared to historical outcomes reported for children enrolled on the HLH-94 trial. This is likely related to a marked delay in referral to our institution, and high early mortality precluding consideration of SCT. Adult patients with malignancy- related HPS in this study had a markedly inferior outcome compared to patients with non-malignancy related HPS. Disclosures: No relevant conflicts of interest to declare.

Abstract Introduction T-cell prolymphocytic leukemia (T-PLL) is a rare mature T-Cell malignancy typically associated with aggressive clinical course. Leukemic T-cells demonstrate post-thymic T-cell phenotypes (Tdt-, CD1a-, CD5+, CD2+ and CD7+) and commonly are CD4+/CD8- T-cells, but co-expression of CD4+/CD8+ or CD8+/CD4- have also been observed. Rearrangement of chromosome 14 involving TCL1 locus is the cornerstone for T-PLL diagnosis which leads to overexpression of oncogene TCL1. As a binding partner for TCL1, AKT expression correlates with worse prognosis. Despite advances in identification of novel mutations, therapeutic option is limited with most patients having very short survival. Herein, we performed a detailed analysis of specific T-cell subsets affected in T-PLL and a comprehensive genomic analysis by whole exome sequencing (WES) and RNA sequencing. We also explored the preclinical efficacy of targeting AKT activation by AKT inhibitor MK2206 on leukemic T-cells isolated from T-PLL patients. Methods T-PLL leukemic cells were isolated from blood or marrow samples of T-PLL patients (n=9) and were tested for involved T-cell subsets by multicolor flow cytometric analysis. Leukemic T-cells were treated with escalating doses of MK2206 (0 to 10 µM) for about 80 hours and were assessed for their apoptosis levels using Annexin V flow cytometric analysis. Whole exome sequencing (WES) and RNA sequencing were conducted using genomic DNA and RNA isolated from purified leukemic T cells from T-PLL patients. Results Phenotypic analysis of T-cell subsets in T-PLL patients showed that leukemic T-cells in 4 untreated T-PLL patients are consistent with naïve T-cell phenotype (CD45RA+, CD45RO-, CCR7+). Leukemic T-cells in 3 of the untreated patients and in 2 of the treated patients have central memory T (CD45RA-, CD45RO+, CCR7+) and effector memory T phenotypes (CD45RA-, CD45RO+,CCR7-) respectively. MK2206 treatment was able to induce dose-dependent apoptosis on the isolated PBMC (containing > 90% leukemic T-cells) of T-PLL patients (n=4) (IC50: 5 µM). FISH analysis found a rearrangement of TCL1 locus on chromosome 14 in all T-PLL cases (n=9). Three cases have also been detected to have del (11q) involving ATM locus and one patient has both del(11q) and del(17p). Subsequent WES (n=7) and RNA sequencing (n=6) analysis revealed recently reported mutations in ATM (frame shift mutation and early stop mutation W579*) in 2 cases, JAK3 (M511I) in 2 cases and STAT5B (T628S and N642H) in 2 cases. Importantly, novel somatic mutations in gene IKZF1 (N159S), NTRK1 (R33W), AP2A2 (P514L), HDAC8 (I115R), RARB (G90W) and TNIP2 (K104Q) were detected by WES. Mutations in EML4 (L548W and F304S) and VAV3 (C282Y and splice site mutation) were also identified in 2 different T-PLL cases. RNA sequencing revealed several fusion transcripts resulting in early stop of several different genes including PTPRT, L3MBTL1, UCKL1 in one T-PLL case. Conclusion T-PLL leukemic T-cells are consistent with either naïve or central memory T-cell subsets in untreated patients. The AKT inhibitor MK-2206 was capable of inducing apoptosis and could be a potential therapeutic agent for T-PLL patients. In addition, we detected known mutations in ATM and JAK-STAT pathways. Novel mutations in genes involving DNA binding and chromatin remodeling (IKZF1, HDAC8) or kinase signal pathway (NTRK1, TNIP2, VAV3, EML4) were uncovered. These results suggest that further therapeutic approaches could be developed to target DNA binding factors or JAK-STAT or AKT-TCL1 signal pathway with an ultimate goal to improve survival of T-PLL patients. Disclosures Off Label Use: Off label use of MK2206: MK-2206 is an AKT inhibitor obtained from Selleck Chemicals and is used to treat leukemic cells in vitro to test the leukemic sensitivity to AKT inhibition. .

SummaryIbrutinib‐related data in Waldenström macroglobulinaemia (WM) remain sparse, particularly outside of trials. We report on 80 patients [previously treated, n = 67 (84%), treatment‐naïve, n = 13 (16%)] with WM, evaluated consecutively at Mayo Clinic, who received ibrutinib off‐study after its approval in 2015 for WM. Overall response rate (ORR) was 91%; major‐response rate (MRR) was 78%. The median time to first response and best response was 2·9 [95% confidence interval (CI): 2–4] and 5·7 (95% CI: 4–12) months, respectively. The median follow‐up was 19 (95% CI: 14–21) months; 18‐month progression‐free survival (PFS) was 82%. The median time on therapy was 12·5 (95% CI: 9·3–16·7) months, and the median duration‐of‐response was 32 (range: 23–32) months. Twenty‐five patients (31%) had discontinued therapy at last follow‐up (68% due to treatment‐related toxicities) and 18% of patients required dose reduction. Fatigue (12%) and atrial‐fibrillation (11%) were common non‐haematological toxicities. IgM rebound occurred in 36% of patients who abruptly discontinued ibrutinib. Following ibrutinib discontinuation, 84% of patients received subsequent treatment, achieving an ORR of 57% and MRR of 50%. The median PFS from commencement of subsequent salvage therapy was 18 months. Ibrutinib therapy, outside of clinical trials, is effective in WM, but is associated with toxicities and challenges, including IgM rebound and a high drug discontinuation rate for reasons other than disease progression.

Abstract CLL is a clinically heterogeneous disease with wide ranging disease course. A novel CLL-IPI based on Rai/Binet stage, IGHV-mutation status, TP53 mutation/deletion, B2M level, and age was developed to stratify patients into 4 risk groups, with a c-statistic of 0.75. Next-generation sequencing has identified ~60 genes recurrently mutated in CLL, some of which are associated with poor overall survival, whereas the clinical effect of most genes is still unknown. Herein, we examine whether somatic mutations in these putative driver genes are associated with time to first treatment (TTT), and whether they add prognostic value beyond CLL-IPI. Based on the 2008 International Workshop CLL criteria, we identified 100 CLL and 96 high-count monoclonal B-cell lymphocytosis (MBL) newly diagnosed from the Mayo Clinic CLL biobank. Pre-treatment peripheral blood mononuclear cells were collected &lt;2 years of diagnosis and tumor DNA was extracted from sorted CD5+/CD19+. We sequenced the coding regions of 61 recurrently mutated CLL driver genes using a custom SureSelect panel, with 24 samples per flow cell in Illumina HiSeq 4000. The average coverage depth was &gt;1000X. Somatic mutations were called using MuTect2 in tumor-only mode. To remove germline variants, variants were eliminated based on minor allele frequencies &gt;0.01%, identified in 1000 Genomes Project, ExAC and/or ESP6500 databases, unless present in known mutation hotspots or COSMIC. After filtering, high/moderate impact mutations were analyzed using Cox regression, to estimate hazard ratios (HR) and 95% confidence intervals (CI) to test for associations with TTT. Among 196 patients the most commonly mutated genes were TP53 (11%), ATM (10%), SF3B1 (10%), NOTCH1 (9%), CHD2 (8%), and BIRC3 (7%). The median follow-up was 8.7 years, and 73 patients were subsequently treated. ATM (HR=3.27, CI:1.8-6.1, P=0.0002) and NOTCH1 (HR=2.41, CI:1.3-4.6, P=0.008) were associated with TTT. When evaluating the total number of mutated genes, we found 32%, 29%, and 39% patients had ≥2, 1, or 0 genes mutated, respectively, and this was associated with shorter TTT (HR=1.74, CI:1.3-2.4, P=0.0005) adjusting for sex and CLL-IPI with a c-statistic=0.8 (CI: 0.75-0.84). When stratified by CLL-IPI, the association held for low (N=99, HR=1.88, CI:1.1-3.4, P=0.03) and intermediate risk (N=54, HR=1.87, CI:1.1-3.2, P=0.03) but not high/very high risk (N=35, HR=1.07, CI:0.6-1.9, P=0.83). We demonstrated that the total number of CLL putative driver genes with high or moderate impact mutations provided prognostic information in newly diagnosed CLL/MBL beyond CLL-IPI. Moreover, even among those with low or intermediate CLL-IPI risk, the total number of somatic mutations separated those patients who progressed. Sequencing the CLL driver genes at time of diagnosis could be a potential biomarker for outcome prediction. Citation Format: Geffen Kleinstern, Daniel R. O’Brien, Brian F. Kabat, Kari G. Chaffee, Aaron D. Norman, Timothy G. Call, Sameer A. Parikh, Jose F. Leis, Wei Ding, James R. Cerhan, Neil E. Kay, Susan L. Slager, Esteban Braggio. Somatic mutations within chronic lymphocytic leukemia (CLL) putative driver genes are associated with outcomes beyond the CLL international prognostic index (CLL-IPI) [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 4466.

Background:We have previously reported that ∼50% of CLL patients (pts) experience rapid flare of disease when they discontinue ibrutinib therapy due to progression of disease (CLL or Richter's transformation [RT]) (Hampel et al, JCO suppl, 2018). This lead to the clinical practice of overlapping venetoclax with ibrutinib for combination therapy in this setting to prevent ibrutinib withdrawal.Aims:To evaluate the role of continuing ibrutinib beyond the start of salvage venetoclax in pts who have progression of disease on ibrutinib monotherapy, both in preventing ibrutinib withdrawal associated with ibrutinib discontinuation and in disease response.Methods:We identified all CLL pts at Mayo Clinic (3/2017–11/2018) who had disease progression on ibrutinib therapy, where venetoclax was added as subsequent treatment while ibrutinib was continued. CLL‐specific characteristics, treatment timeline and toxicities, and disease course following combination therapy were ascertained. All pts started venetoclax according to the weekly ramp‐up per the package instructions.Results:Eighteen pts treated with the addition of venetoclax to ibrutinib therapy were identified; indications were CLL progression (n = 13) and RT (n = 5). BTK or PLCG2 mutations were present in 8/13 pts. Median duration of follow‐up was 9.6 months after the start of combination therapy.The target venetoclax dose was reached by 5 weeks in 13 pts; reasons for delaying or stopping dose escalation included diarrhea (n = 2), sepsis (n = 2), and neutropenia (n = 1). No patient experienced a disease flare during venetoclax ramp‐up.The median duration of combination ibrutinib and venetoclax therapy was 3 months (range, 0–14 months). Of 18 pts who started combination therapy with ibrutinib and venetoclax, ibrutinib was discontinued in 10 pts. Reasons included hematopoietic stem cell transplant (n = 2), planned stop after adequate disease control (n = 4), financial reasons (n = 1), toxicity (neutropenia n = 1, immune thrombocytopenia/bleeding risk n = 1), and death (n = 1). Only 1 patient experienced ibrutinib withdrawal, which occurred following planned discontinuation after clinical complete response (CR) with a 41 day overlap and toleration of venetoclax 400 mg daily; management with re‐initiation of ibrutinib was effective in regaining disease control. Of 9 pts who stopped ibrutinib, 4 pts had a slow taper over 2–8 weeks, and 2 pts also received corticosteroids at the time of ibrutinib discontinuation.Among the other 8 pts who did not discontinue ibrutinib, 6 pts remain on the combination with venetoclax (median duration 6.5 months, range 3–14 months), and 2 pts continued on ibrutinib following venetoclax discontinuation due to disease progression (both pts received further combination of ibrutinib with multi‐agent chemotherapy).The overall response rate (ORR) to combination therapy with ibrutinib and venetoclax was 81% (CR/clinical CR n = 9, partial response n = 4, stable disease n = 1, and primary refractory disease n = 2). The duration of combination therapy was <4 weeks in 2 pts, precluding response assessment. The estimated 12‐month event‐free survival was 54% (95% CI 33–87%), and the estimated 12‐month overall survival from start of combination therapy was 79% (95% CI 60–100%); see Figure 1.Summary/Conclusion:Among CLL pts who develop progression of disease on ibrutinib monotherapy, continued ibrutinib therapy with addition of venetoclax was feasible, ameliorated ibrutinib withdrawal, and was associated with high ORR.image

Background:The expression of the receptor tyrosine kinase (RTK) Axl on cancer cells has been associated with poor prognosis and implicated in disease progression. Inhibition of Axl‐RTK with TP‐0903, a high affinity Axl‐RTK inhibitor induces robust apoptosis of several tumors, including chronic lymphocytic leukemia (CLL). Axl‐RTK inhibition has also been found to induce nonspecific immunomodulatory effects in syngeneic tumor models.Aims:We aimed to investigate the impact and downstream effects of Axl‐RTK inhibition on naïve T cells and chimeric antigen receptor T cells (CART), using CD19 directed CART cells (CART19) as a model.Methods:We utilized either PMA/Ionomycin or the CD19+ mantle cell lymphoma cell line, JeKo as a stimulant for naïve T cells and CART19, respectively. Cells were analyzed for cytokine profile, T cell phenotype, and kinetics after the stimulation in the presence or the absence of the Axl‐RTK inhibitor TP‐0903. Regarding in vivo experiment, a relapse JeKo xenograft model was used. Mice were treated with vehicle alone, TP‐0903 alone, CART19 alone, or TP‐0903 + CART19. Three weeks after the treatment, mice were re‐challenged with JeKo as a CART19 relapse model.Results:Stimulation of naïve T cells in the presence of the Axl‐RTK inhibitor TP‐0903 resulted in a significant reduction of Th2 cytokines and downregulation of inhibitory receptors. The differential expansion of either effector T cells or Tregs in the presence of TP‐0903 resulted in a preferential reduction of Tregs (Fig 1A). Axl‐RTK inhibition of CART19 led to Th1 polarization upon stimulation with the JeKo, as evident by the selective reduction of Th2 cytokine. Additionally, this resulted in a reduction of myeloid attracting cytokines, such as CXCL1, IL‐8, IL‐6, and IL‐10, cytokines implicated in the development cytokine release syndrome and neurotoxicity after CART cell therapy and in adaptive resistance to immunotherapy (Fig 1B). TP‐0903 also significantly downregulated inhibitory receptors on activated CARTs (Fig 1C). In our relapsed in vivo model, mice treated with CART19 + TP‐0903 rejected the JeKo challenge while mice received CART19 alone redeveloped disease, suggesting that Axl‐RTK inhibition enhanced CART cell persistence. Peripheral blood analysis of mice after the rechallenge showed a higher number of circulating T cells and lower level of MCP‐1. Western blotting was performed to investigate the downstream effect of Axl‐RTK inhibition on CART19 and it revealed a relatively selective inhibition of LCK on stimulated CART19. Transcriptome analysis of activated CARTs treated with TP‐0903 revealed >100 genes that were differentially expressed compared to untreated cells. Among these, Th cell and Treg polarization genes were significantly altered in activated CARTs treated with TP‐0903.imageFinally, we validated our findings in phase I clinical trial of TP‐0903 (NCT02729298). T cells isolated from 3 cancer patients showed a significant reduction of Tregs and Th1 polarization. This will be further investigated in a planned phase I clinical trial of TP‐0903 in relapsed/refractory CLL (NCT03572634).Summary/Conclusion:We showed for the first time that Axl‐RTK inhibition polarizes T cells into a Th1 functional phenotype, downregulates inhibitory receptors, and reduces myeloid associated cytokine production by CART19. The combination therapy yielded synergistic cytotoxicity against JeKo in vitro and in a CART19 relapse model. These findings encourage further study of Axl‐RTK inhibition with TP‐0903 as an enhancer of T cell immunotherapies.

PURPOSE Time to event is an important aspect of clinical decision making. This is particularly true when diseases have highly heterogeneous presentations and prognoses, as in chronic lymphocytic lymphoma (CLL). Although machine learning methods can readily learn complex nonlinear relationships, many methods are criticized as inadequate because of limited interpretability. We propose using unsupervised clustering of the continuous output of machine learning models to provide discrete risk stratification for predicting time to first treatment in a cohort of patients with CLL. PATIENTS AND METHODS A total of 737 treatment-naïve patients with CLL diagnosed at Mayo Clinic were included in this study. We compared predictive abilities for two survival models (Cox proportional hazards and random survival forest) and four classification methods (logistic regression, support vector machines, random forest, and gradient boosting machine). Probability of treatment was then stratified. RESULTS Machine learning methods did not yield significantly more accurate predictions of time to first treatment. However, automated risk stratification provided by clustering was able to better differentiate patients who were at risk for treatment within 1 year than models developed using standard survival analysis techniques. CONCLUSION Clustering the posterior probabilities of machine learning models provides a way to better interpret machine learning models.