Blinatumomab Nonresponse and High-Disease Burden Are Associated With Inferior Outcomes After CD19-CAR for B-ALL

Abstract

Purpose: CD19-targeted chimeric antigen receptor T cells (CD19-CAR) and blinatumomab effectively induce remission in relapsed or refractory B-cell acute lymphoblastic leukemia (ALL) but are also associated with CD19 antigen modulation. There are limited data regarding the impact of prior blinatumomab exposure on subsequent CD19-CAR outcomes.

Patients and methods: We conducted a multicenter, retrospective review of children and young adults with relapsed or refractory ALL who received CD19-CAR between 2012 and 2019. Primary objectives addressed 6-month relapse-free survival (RFS) and event-free survival (EFS), stratified by blinatumomab use. Secondary objectives included comparison of longer-term survival outcomes, complete remission rates, CD19 modulation, and identification of factors associated with EFS.

Results: Of 420 patients (median age, 12.7 years; interquartile range, 7.1-17.5) treated with commercial tisagenlecleucel or one of three investigational CD19-CAR constructs, 77 (18.3%) received prior blinatumomab. Blinatumomab-exposed patients more frequently harbored KMT2A rearrangements and underwent a prior stem-cell transplant than blinatumomab-naïve patients. Among patients evaluable for CD19-CAR response (n = 412), blinatumomab nonresponders had lower complete remission rates to CD19-CAR (20 of 31, 64.5%) than blinatumomab responders (39 of 42, 92.9%) or blinatumomab-naive patients (317 of 339, 93.5%), P < .0001. Following CD19-CAR, blinatumomab nonresponders had worse 6-month EFS (27.3%; 95% CI, 13.6 to 43.0) compared with blinatumomab responders (66.9%; 95% CI, 50.6 to 78.9; P < .0001) or blinatumomab-naïve patients (72.6%; 95% CI, 67.5 to 77; P < .0001) and worse RFS. High-disease burden independently associated with inferior EFS. CD19-dim or partial expression (preinfusion) was more frequently seen in blinatumomab-exposed patients (13.3% v 6.5%; P = .06) and associated with lower EFS and RFS.

Conclusion: With the largest series to date in pediatric CD19-CAR, and, to our knowledge, the first to study the impact of sequential CD19 targeting, we demonstrate that blinatumomab nonresponse and high-disease burden were independently associated with worse RFS and EFS, identifying important indicators of long-term outcomes following CD19-CAR.

FIG 1.

Characteristics of blinatumomab-naive and blinatumomab-exposed patients. (A) Distribution of full patient cohort (N = 420) by CD19-CAR construct. (B) Median and interquartile range of days from most recent blinatumomab exposure to CAR infusion among blinatumomab-exposed patients (n = 77). (C) Outcomes of patients by CAR construct. (D) CR rate to blinatumomab on the basis of pre-blinatumomab disease burden (data only available on 70 patients, missing data for seven patients). Among the low-disease burden patients (n = 30), 11 were in an MRD-negative CR at the time of blinatumomab administration. (E) Flowchart of blinatumomab-exposed patients (n = 77), including best response to blinatumomab, status of post-blinatumomab interim HSCT, and response to CAR. (F) CR rates to CD19-CAR among blinatumomab-exposed patients who were evaluable for response (n = 73), stratified by response to prior blinatumomab. (G) CIR, stratified by blinatumomab-naïve (solid line), blinatumomab-exposed prior responders (Blina-CR, dotted line), and blinatumomab-exposed prior nonresponders (Blina-No CR, dashed line). alloHSCT, allogeneic hematopoietic stem-cell transplant; Blina, blinatumomab; CAR, chimeric antigen receptor; CIR, cumulative incidence of relapse; CR, complete remission; MRD, minimal residual disease; NE, not evaluable; NR, no response.

FIG 2.

EFS, RFS, and OS. (A-C Kaplan-Meier survival curves for all patients. (A) EFS, defined as the time from CD19-CAR infusion to one of the following events: no response, relapse, or death. (B) RFS, defined as the time from the first response to relapse or death. (C) OS, (D) EFS, (E) RFS, and (F) OS, stratified by blinatumomab-naïve patients (No Blina—teal) versus blinatumomab-exposed and achieved CR to Blina (Blina-CR—blue) versus blinatumomab-exposed and did not achieve a CR to Blina (Blina-No CR—red). P values for EFS curve: .59 (No Blina v Blina-CR); .01 (Blina-CR v Blina-No CR); .0001 (No Blina v Blina-No CR). P values for RFS curve: .70 (No Blina v Blina-CR); .0004 (Blina-CR v Blina-No CR); < .0001 (No Blina v Blina-No CR). P values for OS curve: .97 (No Blina v Blina-CR); < .0001 (Blina-CR v Blina-No CR); < .0001 (No Blina v Blina-No CR). (G) EFS, (H) RFS, and (I) OS, stratified by high-disease burden (≥ 5% bone marrow blasts—blue [high]) versus low-disease burden (< 5% bone marrow blasts—red [low]). (J) EFS, (K) RFS, and (L) OS, restricted to blinatumomab-exposed patients only, stratified by blinatumomab exposure, blinatumomab response, and disease burden (high-disease burden [≥ 5% bone marrow blasts] v low-disease burden [< 5% bone marrow blasts]). Blina-CR-low—red; Blina-CR-high—blue; Blina-No CR-low—brown; Blina-No CR-high—teal. P value shown represents the global P value. Blina, blinatumomab; CAR, chimeric antigen receptor; CR, complete remission; EFS, event-free survival; HD, high disease; high, high-disease burden; LD, low disease; low, low-disease burden; OS, overall survival; RFS, relapse-free survival.

FIG 3.

Factors associated with EFS after CD19-CAR. EFS, stratified by (A) CD19-CAR construct (41BB v CD28), (B) primary refractory disease versus relapsed disease, (C) active extramedullary disease at CAR infusion versus no extramedullary disease, and (D) active peripheral blasts at CAR infusion versus no blasts. (E) Multivariable comparisons of EFS. ^aAssociated with worse EFS with CD19/CD28 CAR. Active EM, active extramedullary disease at infusion; Active PB, detection of peripheral blasts at infusion. Blina, blinatumomab, CAR, chimeric antigen receptor; EFS, event-free survival.

FIG 4.

CD19 modulation. Top: Patients with CD19-positive expression pre-CAR. (A) Outcomes following CD19-CAR for blinatumomab-naïve patients and subsequent CD19 immunophenotype at relapse. (B) Outcomes following CD19-CAR for blinatumomab-exposed patients and subsequent CD19 immunophenotype at relapse. Middle: Patients with CD19 dim expression pre-CD19-CAR. (C) Outcomes following CD19-CAR for blinatumomab-naïve patients and subsequent CD19 immunophenotype at relapse. (D) Outcomes following CD19-CAR for blinatumomab-exposed patients and subsequent CD19 immunophenotype at relapse. (Outcomes for patients with CD19 partial [n = 3] and CD19 status unknown [n = 6] are not shown. Among the CD19 partial population, one patient had received prior blinatumomab, and all three remain in an ongoing CR. Among patients with CD19 unknown expression, two had prior blinatumomab, three remain in an ongoing remission, three relapsed with CD19-positive disease [one had prior blinatumomab], one relapsed with CD19-negative disease [n = 1], and one relapsed with unknown CD19 expression. One patient was considered nonevaluable for response but emerged with CD19-negative relapse following initial disease assessment.) Bottom: CD19 expression at relapse. (E) CD19 expression at relapse, stratified by disease burden. (F) CD19 expression at relapse, stratified by time point of relapse post-CAR infusion. CAR, chimeric antigen receptor; CR, complete remission.