Predicting CAR T-Cell Treatment Failure in Patients With Aggressive B-Cell Lymphomas

Nearly one-half of patients with relapsed/refractory diffuse large B-cell lymphoma (DLBCL) who are treated with chimeric antigen receptor (CAR) T-cell therapy experience a relapse within one month of infusion, according to research published in Blood Advances. Study authors also established a set of risk factors for early progression, including number of extranodal sites and total metabolic tumor volume (TMTV).

“Axicabtagene ciloleucel and tisagenlecleucel have been approved … for patients having received ≥2 prior lines of therapy for most aggressive B-cell [non-Hodgkin lymphomas]. Data from pivotal trials suggest durable complete response (CR) rates around 30% to 40%, particularly for individuals with early CR,” the authors, led by Laetitia Vercellino, MD, from Hôpital Saint-Louis in Paris, France, explained. “Little is known regarding the factors associated with nonresponse to CAR T-cell therapy. The identification of such factors may help to better select patients who will respond.”

To help identify these factors, researchers retrospectively collected medical chart data at the time of therapy decision (TD) and time of treatment (TT) in 116 patients who were treated consecutively at five French lymphoma centers between June 2018 and January 2020. Patients received either tisagenlecleucel (n=67) or axicabtagene ciloleucel (n=49).

The most significant prognostic markers were related to the lymphoma, including extranodal disease and total metabolic tumor volume.

Patient characteristics analyzed included individual components of the International Prognostic Index (IPI), C-reactive protein (CRP), ferritin, albumin, and lymphocyte count. The type of bridging, time for CAR T-cell manufacturing, and survival data were also recorded. Imaging evaluation was performed at TT, and TMTV and tumor bulk were measured at TT, after the last bridging treatment (if any), and before starting lymphodepletion therapy. Response was evaluated at one month post-infusion, followed by every three months for the first year, every six months for the second year, and then annually.

At TD, median age was 60.7 years, and most patients (90%) presented with a good performance status and disseminated disease (stage III/IV; 77%. The authors noted that 28% presented with more than two extranodal sites. Approximately half of patients (n=55; 47%) had elevated lactate dehydrogenase (LDH).

DLBCL was the most common diagnosis (80%), followed by primary mediastinal B-cell lymphoma (5%), and transformed follicular lymphoma (14%). Patients were heavily pretreated, with 30% having received more than four lines of therapy.

At a median follow-up of 8.2 months, 55 patients (47%) experienced a disease relapse. All but one relapse occurred between day four and four months after infusion, at a median of 30.5 days post-infusion. Twenty-seven relapses (49%) were considered early progressors within one month.

Overall, 29 patients died, including 27 after disease progression/relapse. The estimated six- and 12-month survival rates were 79% and 67%, respectively.

“The most significant prognostic markers analyzed were related to the lymphoma,” the researchers wrote, including the extranodal disease and the metabolic activity of the lymphoma.

At TD, risk factors for relapse were:

  • Eastern Cooperative Oncology Group performance status ≥2
  • elevated LDH
  • extranodal sites ≥2
  • high IPI
  • high revised IPI

At TT, risk factors for relapse were:

  • elevated LDH
  • low lymphocyte count and albumin
  • high levels of ferritin and CRP
  • bulky mass (>5 cm)
  • high TMTV

Looking at TMTV, a cutoff value of 80 mL correctly predicted 21 of the 27 early relapses, while the optimal cutoff for discriminating survival was 55 mL, with 24 of 29 deaths correctly predicted.

Multivariate analyses confirmed that elevated LDH and greater extranodal involvement at TD and the same factors at TT (increased CRP, ≥2 extranodal sites, and TMTV >80 mL) predicted relapse.

Next, when the authors categorized patients into three prognostic groups based on these parameters, they were able to differentiate groups better than the IPI, “making it a better suited risk classifier for CAR T-cell strategies.”

However, they noted that the type of CAR T-cell treatment used may have affected outcomes, a factor that was not accounted for in this analysis. “We could not imagine that the function of the CAR T cells, their expansion after infusion, and their persistence would not have major influences on the response to treatment,” the authors wrote. “This should be analyzed in future studies, focusing on a comparison with an adequate sample size and causal inference methods of analyses to correct for potential confounding by indication bias.”

Study authors report no relevant conflicts of interest.


Vercellino L, Di Blasi R, Kanoun S, et al. Predictive factors of early progression after CAR T-cell therapy in relapsed/refractory diffuse large B-cell lymphoma. Blood Adv. 2020;4(22):5607-5615.