As FDA puts CAR-T safety under microscope, researchers dig into adverse events record of BCMA therapies

BCMA-targeted therapies are transforming care for multiple myeloma patients. But these immunotherapies also come with various potentially dangerous side effects. A new study now looks at those risks from the FDA Adverse Event Reporting System (FAERS), hoping to help inform doctors in their treatment decisions.

After going through about 4,400 reported adverse reactions for three FDA-approved BCMA agents from the 2019 to the second quarter of 2023, researchers found that Bristol Myers Squibb’s CAR-T therapy Abecma had the highest proportion of cytokine release syndrome (CRS) in its total reported events and that Johnson & Johnson and Legend Biotech’s CAR-T Carvykti reported the most Parkinson’s-like cases, whereas J&J’s bispecific antibody Tecvayli had a “notably higher” rate of infections.

Because these three therapies are all approved for late-line treatment of myeloma, “clinicians will need to carefully consider their distinct toxicity profiles when choosing and sequencing treatments,” the study’s authors said in an abstract for the 65th American Society of Hematology annual meeting.

Before getting into the details, it’s necessary to name a couple important limitations of the study. First, the rates calculated in the study only reflect how a specific type of side effect contributes to a drug’s overall adverse events. It’s different from the rate describing the frequency of an adverse event among all treated patients. Second, FAERS is a voluntary reporting system that doesn’t capture all adverse events happening in the real world. And the cases reported are not validated. 

Rahul Banerjee, M.D. (Fred Hutch)

The nature of FAERS “makes it very tricky to really understand what’s happening” with these therapies, Rahul Banerjee, M.D., a multiple myeloma expert at the Fred Hutch Cancer Center who’s not involved in the study, said in an interview with Fierce Pharma. 

“FAERS is helpful for rare, rare side effects,” Banerjee said. As for other more common problems such as CRS, which has become “bread-and-butter CAR-T management,” he wouldn’t draw too many conclusions from FAERS.

As David Miklos, M.D., Ph.D., who leads the Blood and Bone Marrow Transplantation and Cellular Therapy program at Stanford University noted during a Q&A after the data presentation, there may be a more authoritative resource for recording adverse events from the CAR-Ts. The Center for International Blood & Marrow Transplant Research collects information from the FDA-mandated drug safety programs known as Risk Evaluation and Mitigation Strategy (REMS) for the CAR-Ts, he pointed out. Because of their risky safety profile, all commercial CAR-Ts are only offered through REMS in the U.S.

Common problems

Of all the adverse events, 874 were associated with Carvykti, 2,132 with Abecma and 1,363 with Tecvayli. The drugs got their FDA approvals in February 2022, March 2021 and October 2022, respectively.

When it comes to CRS, a well-understood toxicity that’s spelled out in black-box warnings of all three drugs’ labels, Abecma had 344 reports (16.1% of all the drug’s adverse cases). CRS accounted for 10.3% of all adverse reports for Carvykti. Tecvayli had CRS making up 9.5% of its cases.

For both CAR-Ts, CRS was the most common reason for deaths that weren’t caused by cancer relapse, followed by immune effector cell-associated neurotoxicity syndrome (ICANS), another known problem of CAR-T therapy. In the FAERS study, Carvykti had the lowest proportion of non-relapse deaths.

For Tecvayli, COVID-19, sepsis and pneumonia—which are all infection-related complications—had the most frequent non-relapse deaths. Infections in general appeared to be a bigger problem among Tecvayli takers, as the proportion of infectious events in the J&J bispecific’s total reports was four times that of the two CAR-Ts combined.

But Banerjee said he would want to see a time-variant analysis to show how these adverse events evolve over time. For example, he pointed out an increasing use of intravenous immunoglobulin (IVIG) to control infections.

“I do think as a field, as we get more familiar with these drugs, we’re learning more and more about how to prevent these side effects,” Banerjee said.

Managing side effects

Of the three BCMA therapies included in the study, Carvykti has shown the best efficacy. But some doctors flagged Parkinsonism as a concern that they’re closely monitoring for the J&J/Legend product. In this study, researchers found seven cases of Parkinsonism from Carvykti and four from Abecma. Plus, unexpected facial paralysis—a neurological disorder called Bell’s palsy—was reported almost exclusively from Carvykti, with 13 cases, versus none for Abecma and one for Tecvayli. 

Since the Parkinson’s-like problem emerged, J&J and Legend have identified certain risk factors and developed a mitigation strategy that’s now being deployed in Carvykti’s clinical trials and commercial launch, Legend CEO Ying Huang, Ph.D., said in an interview with Fierce Pharma. 

Ying Huang, Ph.D. (Legend Biotech)

If a patient has a very high tumor burden at baseline, that could correlate to fast and high peak expansion of the CAR-T cells, which corresponds to this delayed neurotoxicity, Huang explained. To solve that problem, J&J and Legend suggest doctors first treat patients with bridging therapy to reduce that tumor burden.

In addition, higher grade of CRS also appeared to correspond to Parkinsonism, according to Huang. So, doctors are advised to proactively treat CRS to prevent it from worsening. Besides, doctors could do a test as simple as a handwriting assessment to look for early signs of neurotoxicity. When discovered early, these neurotoxicities typically resolve after a standard four-week course of oral corticosteroid, Huang said.

Overall, non-ICANS neurotoxicity was reported most often in Abecma’s adverse events reports at 10.1%. This group of events accounted for 6.4% of Carvykti’s adverse events.

Both Carvykti and Abecma are under FDA review to potentially expand into earlier lines of treatment. But BMS said the agency plans to hold an advisory committee meeting to discuss Abecma’s KarMMa-3 trial with a focus on overall survival data, raising suspicions that the drug’s life extension benefit might be markedly worse than the 51% tumor progression improvement. Investigators will share the overall survival data at ASH on Monday.

Discrepancies between overall survival and progression-free survival could largely be explained by non-relapse deaths. Again, Banerjee argued that non-relapse mortality should be examined together with the timing of the event and in the context of efficacy.

FDA’s secondary cancer investigation

The FAERS study presentation didn’t include any information on secondary cancer. But it comes as the FDA recently launched an investigation into potentially serious risks of developing secondary cancer after treatment with existing CAR-Ts, including four CD19-directed lymphoma products. 

Almost simultaneously, an article published in ASH’s journal Blood described a 51-year-old male patient in Carvykti’s CARTITUDE-4 trial who developed CAR-positive T-cell lymphoma after Carvykti treatment.

Legend’s Huang stressed that the FDA didn’t say anything about a causal link between CAR-T and secondary cancer. He also pointed to an up to 0.5% background incidence rate of T-cell lymphoma among multiple myeloma patients, as well as other treatments that a patient would typically go through that are also associated with an increased risk of secondary cancer.

The FDA has identified 19 total T cell lymphoma cases from FAERS and clincal trials. Based on the tens of thousands of patients that have been treated, Huang argued “there’s overwhelming positive benefit-risk here.”

Banerjee also argued that CAR-Ts’ benefits are “very real” against the risk of theoretical cancer, especially in heavily pretreated patients. Even if all those 19 cases were directly related to insertion mutations caused by CAR-Ts’ viral vectors, the benefit-risk calculation is still favorable, he said.

Besides, Banerjee said he wouldn’t be surprised to one day see CAR-T recipients have a higher lifetime incidence of certain cancers versus non-recipients simply because CAR-Ts are helping them live longer. In other words, patients who didn’t get CAR-Ts might have died from their primary cancers before any possibility of developing new cancers later in life.

As for that Carvykti case, Banerjee acknowledged that its CAR-positive nature and the fact it emerged just five months post-infusion definitely looked bad. But he argued it was still not a smoking gun.

The case had other cancer-driving mutations in the cells that might predispose the patient to T cell malignancies, he noted. And the CAR inserted into a region of an entirely different gene.

One way Banerjee said could confirm a causal link requires CAR insertion into an important gene with the cells becoming cancerous, which didn’t happen in the Carvykti case. Alternatively, Banerjee said he would also be convinced if there were very high vector copy numbers with chaotic CAR insertions everywhere, a phenomenon that had happened to earlier generations of candidates but not in any existing therapies.

Banerjee also flagged the secondary cancer risks linked to other myeloma treatments.

“I’ve actually never had a patient turn down [BMS’ Revlimid] or transplant purely for the second cancer risk, because everyone gets this, like, look, the cancer at hand is always a bigger problem than the theoretical cancer that may be years from now.”