CAR T-Cell Therapy for Lymphoma: Assessing Long-Term Durability Julie M. Vose, MD, MBA
Relevant Disclosures Research Funding: Kite Pharma/Gilead, JUNO/Celgene, Novartis Honorarium/Ad Boards: Novartis, Kite Pharma/Gilead, JUNO/Celgene, Legend, Janssen
What does AutoSCT achieve in r/r DLBCL in the rituximab era? CORAL Study 3-year PFS 29% (A) Progression-free survival (PFS) according to prior rituximab (ITT) (B) PFS according to time to failure from diagnosis (ITT) 3-year PFS 29% Gisselbrecht et al. J Clin Oncol 2010 28:27;4184-4190.
What does AutoSCT achieve in r/r DLBCL in the rituximab era *? 100 Relapsed or Refractory DLBCL 50 Transplant-Ineligible 50 Transplant Eligible Potential Deaths from Lymphoma 25 Respond to Salvage Therapy and ASCT 10 Patients Cured * Estimates based on Gisselbrecht et al. J Clin Oncol 2010 28:27;4184-4190. * Assumes all patients received rituximab as part of primary therapy.
What can CD19-directed CAR T-Cells achieve in r/r DLBCL in the rituximab era?
KTE-C19 axicabtagene ciloleucel (axi-cel) CD19-directed CAR T-Cells CTL019 tisagenlecleucel (CTL019) JCAR017 lisocabtagene maraleucel (liso-cel) Kite Pharma Novartis Juno Therapeutics scfv = anti-cd19 scfv = anti-cd19 scfv = anti-cd19 CD28-CD3ζ 4-1BB-CD3ζ 4-1BB-CD3ζ FDA approved FDA approved investigational
CD19-directed CAR T-Cells: What is the Response Rate in r/r DLBCL?
JULIET Trial: Eligibility and endpoints tisagenlecleucel (CTL019) N = 111; Median follow-up, 14 mo (max, 23 mo) Key eligibility criteria 18 years of age Central confirmation of histology 2 prior lines of therapy for DLBCL PD after or ineligible for auto-sct No prior anti-cd19 therapy No active CNS involvement Endpoints Primary endpoint: best overall response rate (ORR: CR + PR) Lugano criteria used for response assessment by IRC 1 Secondary endpoints: DOR, OS, safety auto-sct, autologous stem cell transplant; CNS, central nervous system; CR, complete response; DLBCL, diffuse large B-cell lymphoma; DOR, duration of response; IRC, Independent Review Committee; ORR, overall response rate; OS, overall survival; PD, progressive disease; PR, partial response. 1. Cheson BD, et al. J Clin Oncol. 2014;32(27):3059-3068.
JULIET: Patient characteristics Patients (N = 111) Age, median (range), years 56 (22-76) 65 years, % 23 ECOG performance status 0/1, % 55/45 Central histology review Diffuse large B-cell lymphoma, % 79 Transformed follicular lymphoma, % 19 Double/triple hits in CMYC/BCL2/BCL6 genes, % 17 Cell of origin Germinal/Nongerminal center B-cell type, % 57/41 # of prior lines of antineoplastic therapy, % 2/3 / 4-6 44/31 / 21 IPI 2 at study entry, % 72 Refractory/relapsed to last therapy, % 55/45 Prior auto-sct, % 49 Bridging chemotherapy, n 102 Lymphodepleting chemotherapy, n 103 * from Borchmann et al. EHA 2018.
JULIET: Response rates Best ORR w/in 3 months of infusion, 52% (95% CI, 41%-62%): 40% CR, 12% PR *Borchmann et al. EHA 2018.
ZUMA-1 Trial: Eligibility and endpoints axicabtagene ciloleucel (KTE-C19) Key eligibility criteria ZUMA-1 phase 2 portion Cohort 1: patients with refractory DLBCL (n = 77) Cohort 2: patients with refractory PMBCL or transformed FL (n = 24) Key inclusion criteria No response to last CT or relapsed within 12 mos of ASCT Prior treatment with anthracycline and anti-cd20 monoclonal antibody Secondary Endpoints Assess TTR for patients with both objective response and CR Assess PR and CR at Month 3 as PFS prognostic factor TTR, time to response Locke, et al. ASCO 2018.
ZUMA-1: Patient Characteristics Characteristic Overall N=101 Median age, yrs (range) 58 (23-76) Male, n (%) 68 (67) ECOG PS 1, n (%) 59 (58) Disease stage III/IV, n (%) 86 (85) IPI score 3-4, n (%) 46 (46) 3 prior therapies, n (%) 70 (69) Median SPD of index lesions, mm 2 (range) 3721 (171-23, 297) Refractory to 2 lines of therapy, n (%) 77 (76) Best response as PD to last therapy, n (%) 67 (66) Relapse post-asct, n (%) 21 (21) Locke, et al. ASCO 2018.
ZUMA-1: Response rates Neelapu, et al. ASH 2017 Phase 2 (Primary Analysis) N = 101 Median follow-up, mo 8.7 ORR CR Best objective response, % 82 54 Ongoing, % 44 39 Neelapu, et al. ASH 2017.
TRANSCEND NHL 001 Trial: Eligibility and endpoints lisocabtagene maraleucel (liso-cel; JCAR017) N = 73*; Median follow-up, 8 mo (CORE cohort) Key eligibility criteria* DLBCL after 2 lines of therapy: DLBCL, NOS (de novo or transformed FL) High-grade B-cell lymphoma (double/triple hit) Prior SCT allowed Secondary CNS involvement allowed ECOG 0-2 No minimum lymphocyte count requirement for apheresis Response rates - ORR, CR, PR Endpoints DOR, OS, safety *CORE cohort SCT, stem cell transplant; CNS, central nervous system; CR, complete response; DLBCL, diffuse large B-cell lymphoma; DOR, duration of response; ORR, overall response rate; OS, overall survival; PR, partial response.
TRANSCEND: Response rates ORR* at 3 months from infusion, 59% (95% CI, 47%-70%): 45% CR, 14% PR *Includes 2 dose levels in CORE cohort *Abramson, et al. ASCO 2018.
CD19-directed CAR T-Cells: Are Responses Durable?
JULIET: Tisagenlecleucel Response Duration by Best ORR w/in 3 months of infusion (JULIET) Median DOR not reached at 14-mo median follow-up 12-mo relapse-free survival rate CR = 78.5% (95% CI, 60%-89%) CR + PR = 65% (95% CI, 49%-78%) * from Borchmann et al. EHA 2018.
ZUMA-1: Axicabtagene Ciloleucel Response Duration by Best Objective Response (ZUMA-1) More than one half of patients with PR progressed by Month 3 12 mo PFS for CR or PR at 3 mos: CR = 79% (95% CI, 63-88); PR = 78% (95% CI, 36-94) Neelapu SS, et al. NEJM. 2017;377:2531; Locke FL, et al. ASCO 2018.
TRANSCEND: lisocabtagene maraleucel Response Duration (TRANSCEND) Median DOR not reached at 8-mo median follow-up * from Abramson et al. ASCO 2018.
CD19-directed CAR T-Cell Therapy Summary CTL019 * tisagenlecleucel KTE-C19 ** axicabtagene ciloleucel JCAR017 *** lisocabtagene maraleucel Disease state r/r DLBCL r/r tfl r/r DLBCL r/r tfl/pmbcl r/r DLBCL r/r tfl Response evaluable pts, n 89 22 77 24 53 20 Follow-up, median 14 months 15.4 months 8 months Efficacy n = 111 n = 101 N = 73 ORR / CR 52% / 40% [w/in 3 mo] 82% / 54% [best] 59% / 45% [at 3 mo] % PFS for CR @ 12 mos 78.5% 79% [88% 3 mo-cr in CR @ 6 mo] DOR (CR + PR; median) not reached 11.1 months not reached DOR (CR; median) not reached not reached not reached Safety n = 111 n = 101 n = 73 CRS 22% grade 3/4* 13% grade > 3** 1% grade > 3** Neurotoxicity 12% grade 3/4 28% grade > 3 15% grade 3/4 * Borchmann et al. EHA 2018; ** Locke, et al. ASCO 2018; Neelapau, et al. NEJM. 2017; *** Abramson et al. ASCO 2018. * Penn scale; ** Lee scale
Inflammatory state Factors That Influence Treatment Success and Failure to CAR T Cell Therapy: A Hypothesized Model Manufacturing starting material Product fitness: T cell expansion capability T cell polyfunctionality Number of specialized T cells infused Conditioning and concomitant medications Tumor burden Tumor immune microenvironment CAR functional avidity Tumor cell biology PK/PD profile CAR, chimeric antigen receptor; PD, pharmacodynamics; PK, pharmacokinetics. Clinical efficacy and toxicity Adapted from Locke et al SITC 2018 #P212 22
In ZUMA-1 CAR T Cell Expansion Associated With Response Objective Response Rate (ORR) CAR expansion was significantly associated with response (P <.001), with an AUC Day0-28 that was 5.4 times as high among patients with a response as among those who did not have a response CAR AUC Day0-28 is defined as cumulative levels of CAR+ cells/μl of blood over the first 28 days post axi-cel AUC fold change is shown for patients with vs. without response Locke et al ASCO 2017 #3023 23
Post-Infusion CAR Peak Levels Pre-Infusion Polyfunctional Strength Normalized CAR Peak + PSI CAR Peak Levels In Vivo and Polyfunctional Strength Index Associated With Objective Response CAR Peak P =.0326 a PSI P =.0119 a CAR Peak + PSI P =.0046 a NR R NR a Mann-Whitney U Test. CAR, chimeric antigen receptor; NR, nonresponder; PSI, polyfunctional strength index; R, responder. R NR R Rossi et al AACR 2017 #2990 24
Association Between Immunosign 21 Score Measured Before CAR T Cell Treatment and Clinical Outcome a A high Immunosign 21 score was associated with objective response at a minimum follow-up of 9 months (P =.012) In a sensitivity analysis, which included the delayed responder, the association between a high Immunosign 21 score and objective response had a P =.053 a This analysis was performed on samples from 25 patients treated with axi-cel with a minimum follow-up of 9 months. One patient subsequently converted from a nonresponder to a responder at 12-month follow-up. b Cutoff was arbitrarily defined as the 25th percentile of the observed scores among samples. Rossi et al AACR 2018 #LB-016 25
Differences in Expression of Immunosign 21 Genes Measured Before CAR T Cell Treatment in Responders vs Nonresponders a a This analysis was performed on samples from 25 patients treated with axi-cel with a minimum follow-up of 9 months. One patient subsequently converted from a nonresponder to a responder at 12-month follow-up. Rossi et al AACR 2018 #LB-016 26
Treatment with Axi-Cel Results in Rapid and Dramatic Changes in the Tumor Immune Microenvironment Checkpoints IFN related Effectors Proliferative Chemokines Galon et al, ASCO 2017 PD-L1 CTLA4 Top transcripts from a pre-specified 43 immune gene panel upregulated in tumor 7-21 days after treatment. IDO1 and other genes not in the 43 panel are pending. LAG3 TNFRSF18 ICOS IRF1 STAT1 STAT4 IFNg CD8A GNLY GZMA GZMM GZMB GZMH IL15 CXCL9 CCL2 CCL5 Galon et al ASCO 2017 #3025 27
Analysis of B Cell and Immune-Related Molecules at Progression Identifies Relapse with CD19+ or CD19- Tumor cells Progression Biopsies N=21 CD19 (n=21) CD19+ 14/21 (67) CD19-7/21 (33) PD-L1 (n=19) PD-L1+ 9/14 (64) PD-L1-4/14 (29) PD-L1 N/E 1/14 (7) PD-L1+ 4/7 (57) PD-L1-2/7 (29) PD-L1 N/E 1/7 (14) Post-progression tumor biopsies (21 evaluable patients) - 33% were CD19- - 62% were PD-L1+ At Baseline, 94% (16/17) of evaluable patients were CD19+ Example CD19+ relapse Example CD19- relapse Baseline and post-progression samples not obtained from the same lesions. PD-L1, programmed death ligand 1. 28 CD19 RNA splice variants identified In DLBCL relapse biopsies Adapted from Neelapu et al ASH 2018 #578 28
Conclusions In ZUMA-1 ~20% of treated patients were primary refractory to anti-cd19 CAR T cell therapy Mechanisms of primary treatment related failure ascribable to: o Product T cell fitness o CAR T cell function in product o Immune exclusionary tumor microenvironment In ZUMA-1 ~35% of treated patients experienced a secondary treatmentrelated failure Mechanisms of secondary treatment-related failure ascribable to: o High tumor burden o Rapid upregulation of immune checkpoints o CD19 target antigen loss 29
Peak Cytokine Levels Correlate with Response to Anti-CD19 CAR T Therapy Kochenderfer JN et al. J Clin Oncol. 2017;35(16):1803-1813. 30
JCAR017 (lisocabtagene maraleucel; liso-cel): CD19-Targeted CAR T-cell Tumor cell Patient s PBMCs Tumor antigen Epitope Immunomagnetic selection Lentivirus transduction Transmembrane domain scfv Spacer Expansion Formulated at specified composition of CD4 + and CD8 + CAR + T cells Administered at precise doses of CD4 + and CD8 + CAR + T cells CAR + CD8 + T cell Intracellular costimulatory domain Signaling sequence CD8 + (targets tumor) CD4 + (targets tumor, supports persistence) CAR + CD4 + Other PBMC Cell Types EF1p CD19 scfv murine monoclonal FMC63 Transmembrane domain Spacer Signaling domain Transduction marker VL linker VH CD28tm 4-1BB CD3ζ T2A huegfrt 3 LTR Siddiqi, ASH 2017
Units Tumor Burden, Baseline Markers of Inflammation, and Inflammatory Cytokines May Trend Lower in Patients with Durable Response 10000 1000 100 10 1 No Response at 3 Months Response (CR/PR) at 3 Months 0.1 a Similar result seen when analyzed at DL2 alone. Pre-LD Parameter a Ferritin and D-dimer measured in μg/l, CRP and SAA-1 measured in mg/l; all cytokines measured in pg/ml p < 0.05 for all parameters except SPD (p = 0.12) Data as of October 9, 2017 Siddiqi, ASH 2017
Estimated Probabilities Preliminary Logistic Modeling Data Suggest a Therapeutic Window Exists that Could Limit Toxicity and Optimize Efficacy Low expansion Potential tumor-mediated suppression Potential product and/or patient characteristics Identify at risk population and investigate strategies to enhance expansion High expansion High tumor burden and inflammatory cytokines Potential product and/or patient characteristics Identify at risk population and investigate strategies to limit expansion 1.0 Target Expansion 0.8 0.6 0.4 Any CRS Any NT Gr 3-4 NT ORR M3 Response 0.2 0 CRS, cytokine release syndrome; NT, neurotoxicity; ORR, overall response rate; M3, month 3 Increasing Maximum CAR T Cell Expansion Siddiqi, ASH 2017 Data as of October 9, 2017
Conclusions CD 19+ CAR-T cell for DLBCL CR rates 30-50% If patients in CR at 6 months post-car-t many maintain response Long term responders some understanding: Lower tumor volume Lower CRP and ferritin less inflammation Cytokine profile in responders Hi PD-1 may decrease the response rates and/or increase early relapse. Combinations during or after CAR-T may help to decrease relapse rates