For emerging therapeutic platforms, progress rarely happens in a straight line. The closing weeks of August 2026 offered a particularly clear example. Within days, Novartis and Bristol Myers Squibb paused autoimmune CAR-T programs following safety events. At almost the same time, Merck and Moderna reported a positive Phase 3 result for their personalized mRNA cancer therapy in melanoma. Then, just over a week later, BioNTech and Genentech terminated a trial of a competing individualized mRNA approach in colorectal cancer. Taken together, the developments offer a reminder of an important distinction in drug development: a setback for one program does not necessarily mean a setback for an entire technology.
CAR-T’s expansion beyond cancer hits a hurdle
CAR-T therapies are already established treatments for some blood cancers, but developers are increasingly investigating whether the same approach could also be used in autoimmune diseases. That effort suffered a significant setback at the end of August. Novartis paused eight studies of rapcabtagene autoleucel, or rap-cel, across autoimmune and neurological conditions after three patients died following severe immune reactions.
The response has since extended beyond the company itself. In September, Brazil’s health regulator Anvisa suspended three rap-cel studies being conducted in the country while the safety concerns are assessed.
Bristol Myers Squibb also voluntarily paused enrollment in autoimmune trials of its own CAR-T candidate, zola-cel, after identifying what the company described as transient and reversible inflammatory events. BMS has said it is reviewing the clinical data and aims to resume enrollment once that assessment is complete.
Importantly, the two situations are not identical. And neither implies that CAR-T as a whole has suddenly become unviable. Novartis’ cancer studies of rap-cel have continued, while other CAR-T programs across the industry remain active.
What the events do raise is a more specific question on whether moving CAR-T into different patient populations changes the balance between its potential benefits and its known risks.
One mRNA win, one mRNA setback
The contrasting fortunes of individualized mRNA cancer therapies make the same point from another direction.
On August 19, Merck and Moderna announced that intismeran autogene, combined with Keytruda, met the main objectives of a Phase 3 trial in patients with surgically removed melanoma. The companies said the combination reduced the risk of the cancer returning or spreading compared with Keytruda alone and plan to discuss regulatory submissions with health authorities.
It represents an important milestone for the idea of designing cancer treatments around the characteristics of an individual patient’s tumor.
Nine days later, however, BioNTech and Genentech announced that they would terminate a Phase 2 colorectal cancer trial of their individualized mRNA therapy, autogene cevumeran.
The study had already failed an earlier assessment of whether it was likely to achieve its efficacy goal. A later review identified an imbalance in overall survival between the treatment and control groups and concluded that continuing the trial was unlikely to alter its outcome.
There is an important qualification. The monitoring committee did not identify a new safety signal associated with the therapy. At the time of writing, BioNTech also continues to investigate autogene cevumeran in other settings.
The platform is only part of the story
It is tempting to interpret clinical news at the level of a technology, concluding that CAR-T is struggling, or mRNA cancer vaccines are succeeding. However, the reality is less tidy.
The recent results show how much can depend on the individual drug, disease, treatment setting and patient population. The same broad technology can produce encouraging results in one trial and disappointment in another without either outcome providing a definitive verdict on the platform itself.
For companies, investors and business development teams assessing emerging modalities, that distinction matters. Platform potential can help identify where innovation is heading, but ultimately the evidence still has to be assessed program by program.
The last few weeks have provided unusually visible examples of that principle, and a reminder that, in drug development, technologies rarely move forward or backward as one.