Blood tests may predict which patients with lymphoma will have poor outcomes after CAR-T
Click Here to Manage Email Alerts
Two blood biomarkers — pre-lymphodepletion C-reactive protein and ferritin — may help predict which patients with lymphoma are at risk for poor outcomes after chimeric antigen receptor T-cell therapy.
Investigators from Roswell Park Comprehensive Cancer Center and Moffitt Cancer Center analyzed data from 136 patients with relapsed or refractory diffuse large B-cell lymphoma.
All patients had received at least two prior lines of therapy. They subsequently received axicabtagene ciloleucel (Yescarta, Kite Pharma), an anti-CD19 CAR-T.
Almost all (93%) experienced cytokine release syndrome and 61% developed immune effector cell-associated neurotoxicity syndrome (ICANS).
Researchers determined patients who had baseline serum blood levels of c-reactive protein (CRP) of at least 4 mg/dL and baseline ferritin of 400 ng/mL or higher faced greater risk for poor outcomes, including shorter PFS, shorter OS and higher rates of severe toxicities.
“This allows us to identify those patients at the highest risk and the lowest risk for poor outcomes,” researcher Marco L. Davila, MD, PhD, senior vice president and associate director for translational research at Roswell Park, told Healio. “For example, I can use the ‘low-risk’ categorization to help me identify patients who I would like to give CD-19 CAR T cells as an outpatient. Because patients in this category are much less likely to have severe toxicities or any kind of major morbidity or mortality, I could feel comfortable managing these patients with CAR-T in an outpatient setting. That’s very helpful. It also helps us begin to identify the 50% of patients who don’t respond.”
Healio spoke with Davila about the need to more accurately predict outcomes after CAR-T, the implications of his team’s findings and next steps in research.
Healio: Why did you conduct this study?
Davila: One goal was to understand how to identify those patients who aren’t going to have a durable response to CAR T-cell therapy so we can offer them potentially innovative clinical trials geared toward improving the probability of response. The second goal was more geared toward toxicity. Patients treated with some of these agents can have severe toxicities, such as cytokine release syndrome, neurologic toxicity, or cytopenia and infections that can cause significant morbidity. We want to be able to identify those patients earlier so we can offer them prophylactic agents to minimize risk for severe toxicity.
Healio: Can you describe the approach you developed and the potential it has demonstrated so far?
Davila: These patients were treated with standard-of-care CD-19 CAR T-cell products. We looked at various patient demographic and general laboratory measurements to correlate them with outcomes.
This is somewhat of a continuation of a study we published in Clinical Cancer Research, which showed pretreatment levels of [interleukin-6 (IL-6)] predicted poor outcomes in terms of likelihood to respond to CAR T-cell treatment or toxicities. IL-6 is a specialized test and sometimes it takes days to get the results. It’s not helpful in guiding the management of the patient in front of you. We started to look at laboratory values that could be a substitute.
We screened a large number of patients treated at Moffitt, and we saw that CRP and ferritin were commonly elevated among patients who had high levels of IL-6. We created a hierarchy of risk, classifying patients who had CRP greater than 4 mg/dL and ferritin greater than 400 ng/mL as high-risk. Patients who did not meet either of those criteria were classified as low-risk, and patients who met only one of those criteria were classified as intermediate risk.
When we broke patients down into those categories, we saw that they correlated very nicely with IL-6.
Next, we used those risk categories to look at outcomes for our patients. Risk categories correlated with outcomes, as PFS and OS were highest among patients in the lowest-risk group.
Our main concern was that this was a single-center study, and our findings could be impacted somewhat by the center selection and treatment strategy. We wanted to validate this, so we reached out to collaborators.
One was Kite/Gilead. We asked them to evaluate the patients who were part of ZUMA-1, the registrational trial that led to the approval of axicabtagene ciloleucel. This same risk category correlated with outcomes for patients treated on ZUMA-1.
We also collaborated with colleagues from Europe. Again, PFS and OS correlated strongly with the CRP and ferritin categories. We felt that this validated our findings against a multicenter, phase 2 registration trial cohort and a multinational European cohort in which other CD-19 CAR T-cell products were used.
Healio: What are the next steps in research?
Davila: The important biological concept is that these laboratory tests are done before the patient is given conditioning chemotherapy or CAR T cells. This suggests the patient’s disease or their inflammatory/immune status is impacting the response to CAR T cells. We’re trying to understand what it is about these patients with high levels of CRP or ferritin. Is something about their lymphoma or their inflammatory status driving these poor responses? We want to really understand mechanistically what is behind this high risk. Is it a mutation in the lymphoma? Is it in the microbiome? We also want to know whether there is any way we can rationally intervene with another therapy to improve their response.
Healio: Is there anything else you’d like to mention?
Davila: This was a single-center study, and we validated our findings with multicenter data. It is important for other CAR-T investigators to look at these risk factors to see if they correlate as well with their patient experience. Also, it might be beneficial for them to consider how they might use this information in ways that my colleagues and I might not have conceived. We’re hoping that this starts to be adapted within the general field of CAR T-cell therapy clinicians. That could generate information or applications that we might not have anticipated.
References:
- Faramand RG et al. Blood Cancer Discov. 2024; doi: 10.1158/2643-3230.BCD-23-0056.
- Faramand RG, et al. Clin Cancer Res. 2020;doi:10.1158/1078-0432.CCR/2-1434.
For more information:
Marco L. Davila, MD, PhD, can be reached at marco.davila@roswellpark.org.
Collapse
Healio Interviews
Click Here to Manage Email Alerts