Biomarkers may help identify patients at increased risk of neurotoxicity from chimeric antigen receptor (CAR) T-cell therapy.

New potential biomarkers and a novel algorithm could help identify patients at increased risk of developing severe neurotoxicity after receiving CD19 CAR T-cell therapy.

So far, two CAR T-cell products have received FDA approval for patients who have limited or no therapeutic options, and “these therapies will become more widely used,” he continued. “Understanding how to minimize the risk of neurotoxicity will be important as we move forward. We need to devise assays to determine who will develop severe toxicity as a trigger to allow us to intervene and ‘cool off’ CAR T-cells and limit neurotoxicity.”

Turtle and colleagues sought to provide a detailed clinical, radiological, and pathological characterization of neurotoxicity arising from CD19 CAR T-cell therapy. They used data from 133 adult patients with relapsed/refractory CD19 B-cell acute lymphoblastic leukemia (ALL), non-Hodgkin lymphoma, or chronic lymphocytic leukemia who were treated with lymphodepletion chemotherapy followed by infusion of JCAR014, a type of CD19 CAR T-cell therapy developed at the cancer center.

Within 28 days of treatment, 53 patients (40%) developed grade 1 or higher neurologic adverse events. Of these, 28 patients (21%) had grade 3 or higher neurotoxicity; alterations in neurologic status completely resolved in a majority of cases. Four of the 133 patients (3%) developed fatal neurotoxicity.

Patients with an early onset of CRS were at increased risk of subsequently developing severe neurotoxicity, the team noted. Tocilizumab, an interleukin-6R antagonist approved by the FDA to treat CRS, ameliorated CRS-related fever and hypotension in most patients, but the role in preventing or treating neurotoxicity was less clear.

Patients who had neurotoxicity were mostly younger and had B-cell ALL, a higher tumor burden, and more CD19-positive cells in the bone marrow as compared with those who did not develop this side effect. Data also revealed that those with severe neurotoxicity had endothelial activation, which could contribute to the capillary leak, blood coagulation abnormalities, and disruption of the blood-brain barrier observed in patients with severe CRS and neurotoxicity.

“This is the most comprehensive description of what neurotoxicity looks like when clinicians come across it in the clinic,” said Turtle. “Others have found that high cytokine levels increase the risk of neurotoxicity, but ours is the first association with endothelial activation. We need to work out whether this is due to focal endothelial activation in neurotoxicity or just a systemic manifestation.”

Turtle and colleagues also developed a predictive classification tree algorithm based on the side effects — including fever, high serum IL-6, and MCP-1 — to identify patients within the first 36 hours after CAR T-cell infusion who are at increased risk for severe neurotoxicity. This algorithm predicted severe neurotoxicity with 100% sensitivity and 94% specificity, the team reported. Eight patients were misclassified; one of the eight patients did not subsequently develop grade 2-3 neurotoxicity and/or grade 2 or higher CRS.

Using algorithms may help identify patients who have an increased risk of severe neurotoxicity and who could benefit from early intervention: “If a patient develops seizure 24 to 36 hours after receiving CAR T-cell therapy, then it’s time to worry about neurotoxicity, if there is no other obvious cause,” said Turtle. “CRS may be occurring aggressively and early. Watch the patient closely and, if need be, intervene therapeutically.”

He cautioned that the cytokine thresholds and seizures outlined in the study apply only to the CD19 CAR T-cell product at Fred Hutchinson — i.e., other CAR T-cell products may lead to different toxicity rates and clinical kinetics. “These biomarkers can’t be applied to other CAR T-cells, and other biomarkers for other CAR T-cells will have to be validated. Ultimately, when an algorithm triggers therapy, clinicians can give steroids or cytokine-directed therapy to mitigate toxicity to ‘cool off’ cytokine secretions.”

The next step in this research, he said, is to look at early intervention in high-risk patients and determine whether the effects of shutting down CAR T-cell response affects efficacy. Turtle’s team and others are currently working on these trials.

Turtle said that most toxicity appeared early in treatment during dose escalation, and the risk of toxicity of CAR T-cell therapy has dropped over the past few years with more experience with this gene-directed immunotherapy. “The potential for severe toxicity still exists with CAR T-cell therapy, but moving forward this will become less frequent. We can optimize CAR T-cell regimens to reduce toxicity. If we can predict how many patients will develop severe toxicity, we may be able to intervene at early time points,” he said.

Enormous improvements have been made in the last few years in strategies to minimize the risk of toxicity, he added. “Because CAR T-cell therapy is so new, we are still learning how to improve the delivery and reduce the side effects.”