In two major clinical trials, researchers found cell therapy dramatically outperformed the standard of care used for the past few decades to counter a common type of lymphoma that’s returned after initial treatment.

The results, presented at the American Society of Hematology’s annual meeting Saturday and Sunday, could support earlier use of the two drugs studied, Gilead’s Yescarta and Bristol Myers Squibb’s Breyanzi.

But the positive findings contrast with disappointing results from a third trial that tested Novartis’ Kymriah, which works similarly to Yescarta and Breyanzi, in the same setting and against the same standard treatment. The divergent results between the three studies are likely to spur debate among doctors over why Kymriah failed and the others succeeded, with potential implications for how cell therapy is used and in which patients.

“It’s an extremely important question,” said Michael Bishop, a hematologist and lymphoma specialist at the University of Chicago, when presenting the Kymriah trial results to reporters Monday. “And if I’m completely honest, I lay awake at night thinking about it.”

Bishop was the study’s lead author and has consulted with Novartis, Gilead and Bristol Myers.

Kymriah, Yescarta and Breyanzi are what’s known as CAR-T therapies, custom made from each individual patient’s own immune cells and engineered to seek out the same target on cancerous cells. All three are approved to treat diffuse large B-cell lymphoma after the first two “lines” of care fail; Breyanzi most recently, in February.

The three Phase 3 trials presented at ASH were meant to test each therapy earlier, when a patient’s lymphoma has become resistant or relapsed following only one line of treatment. Typically, in this second-line setting, doctors turn to chemotherapy and, if patients respond well, stem cell transplants that can cure the disease for some.

But many patients don’t make it to transplantation and, for Yescarta and Breyanzi at least, the trial results suggested CAR-T could be an alternative treatment approach.

In its study, Yescarta extended participants time without an “event” — cancer progression, subsequent treatment or death from any cause — by four times what was observed with chemotherapy and transplantation. Similar results were reported with Breyanzi, which cut the risk of an event by 65%. Rates of treatment response and remission were similar between Yescarta and Breyanzi.

Yet in Novartis’ trial, called BELINDA, Kymriah appeared no better than standard care, missing its main goal. In both groups, participants went a median of three months without an event, which was defined more narrowly as stable or progressive disease at week 12 or death. Rates of overall response and complete response to treatment were identical or nearly so between the two study arms.

Even with the negative results from BELINDA, though, Gilead’s and Bristol Myers’ findings are likely to support second-line use of CAR-T, according to Stephan Grupp, director of the cancer immunotherapy program at the Children’s Hospital of Philadelphia.

“I don’t think it undermines the overall message that cell therapy is better than transplant in this setting,” he said. Grupp helped study Kymriah previously, but was not involved in BELINDA.

Trial researchers had a few theories as to why Kymriah underperformed. Perhaps most notably, patients who were randomized to receive Novartis’ therapy had to wait a long time to receive it — a median of 52 days before CAR-T infusion. By comparison, participants in Gilead’s study of Yescarta were treated a median of 27 days after randomization.

“We think this long infusion time, where you have an extremely progressive disease going on, did not permit the potential benefits of [Kymriah],” said Bishop at the Monday press conference, the contents of which were embargoed until Tuesday morning.

Jeff Legos, Novartis’ global head of oncology and hematology development, said the long time to infusion in BELINDA was half attributable to the complex production process for CAR-T therapies and half to reasons tied to patient scheduling, shipment and disruptions from COVID-19.

Novartis has had issues making Kymriah in the past, but says its manufacturing success rate has improved.

There were also significant differences between how Novartis designed its trial and how Gilead and Bristol Myers did. Gilead, for example, didn’t allow what’s known as “bridging” chemotherapy to help stabilize patients before they get CAR-T therapy. Novartis allowed up to four cycles in BELINDA, while Bristol Myers allowed one course of bridging therapy in its trial.

Researchers at the National Cancer Institute, writing in an editorial published Tuesday in The New England Journal of Medicine, said Gilead’s decision to exclude bridging chemotherapy might bias against the enrollment of patients who have more rapidly progressing disease and need an interim treatment while they wait for their CAR-T therapy.

Novartis also noted imbalances between the study’s treatment and control groups, which may have affected the comparison.

“Concluding that [Yescarta] must be superior to [Kymriah] is the wrong inference from the data,” the NCI researchers wrote. “The selection of suitable patients for the application of CAR T cells is important. What has been learned is that not all patients with ‘relapsed or refractory large B-cell lymphoma’ are the same.”

Yet the BELINDA study might still be the final word on Kymriah in diffuse large B-cell lymphoma, although Novartis is continuing study of it in other cancer types, such as follicular lymphoma. The Swiss drugmaker has also developed a new CAR-T platform called T-Charge that it claims can produce and deliver treatments much faster, potentially taking 50% less time than the current CAR-T turnaround of about two to three weeks.

“Moving forward, because of the emergent data around the T-Charge platform and the ability to have a simplified, more reliable, more robust manufacturing process … it wouldn’t make sense for us to try to repeat the experiment with Kymriah,” said Legos.