Mutations in TP53 and RAS are among the most feared findings in a blood-cancer diagnosis. They mark disease that tends to resist standard therapy and relapse quickly, and for decades they were considered nearly undruggable — TP53 is a damaged tumor suppressor, not an enzyme you can easily block, and mutant RAS proteins have frustrated direct inhibition. A Phase 1 trial that updated on the U.S. registry this week tries to turn those same mutations from a liability into a target, by training a patient's own immune cells to recognize the very thing that makes the cancer dangerous.
The study, NCT06904066, comes from the National Cancer Institute and enrolls patients with any of nine hematologic malignancies, including acute myeloid leukemia and myelodysplastic syndrome. The therapeutic concept rests on a piece of immunology: when a gene like TP53 or RAS carries a missense mutation, the altered protein can be chopped into short peptides that look subtly foreign to the immune system. Those altered peptides are called neoepitopes, and they are, in principle, a flag that distinguishes cancer cells from healthy ones. The registry lays out the rationale directly.
"Missense mutations in TP53 and RAS result in immunogenic peptides (neoepitopes) that can be presented by human leukocyte antigens (HLA) to initiate an immune response."— ClinicalTrials.gov, source
The strategy builds on years of work at the NCI Surgery Branch, which has previously identified T-cell receptors — the molecular sensors T cells use to spot threats — that selectively recognize p53 or Ras neoepitopes. This trial proposes to evaluate seven such receptors: three targeting p53 neoepitopes and four targeting Ras neoepitopes. The manufacturing is a bespoke, per-patient process. A participant first has a bone marrow biopsy to confirm the diagnosis and the specific TP53 and RAS mutations, plus a skin biopsy. Their T cells are then collected by apheresis, transduced with a retroviral vector carrying the matched receptor, and grown into a population of neoepitope-specific T cells before being infused back. Because the receptor must match both the patient's mutation and their HLA type, this is personalized therapy in the most literal sense — the product is built around each individual's tumor genetics.
What surrounds the cell infusion
The regimen around the engineered cells follows the established template for adoptive T-cell therapy. Before infusion, patients receive a short course of cyclophosphamide and fludarabine over three days — a lymphodepleting conditioning regimen that clears space for the transferred cells to expand. After the cells go in, high-dose aldesleukin (interleukin-2) is given over the following days to support the T cells' survival and proliferation. The trial also uses a sequencing panel to detect the TP53 or RAS mutations that determine eligibility, underscoring how tightly the diagnostic and the therapeutic are coupled here.
The choice to evaluate seven receptors rather than one is itself a statement about the difficulty of the target. Because each TCR recognizes a specific mutant peptide presented on a specific HLA type, no single receptor can serve every patient — the field is, in effect, assembling a library of receptors so that more patients can be matched to a product. That panel approach is what allows a planned enrollment of up to 86 participants across nine hematologic malignancies; the breadth comes from the number of receptors on offer, not from any one of them being broadly applicable. It also raises the bar for interpretation, since safety and any activity must ultimately be assessed receptor by receptor rather than for the program as a whole. A retroviral vector, rather than a more modern editing approach, is used to install the receptor into the patient's T cells — a mature, well-characterized technique for stable gene transfer that the NCI Surgery Branch has long deployed in its adoptive-cell programs. That continuity matters: the manufacturing and conditioning playbook here is borrowed from years of solid-tumor adoptive-cell work, and applying it to blood cancers with defined driver mutations is the genuinely new variable the study is testing.
Reading the endpoints
The single primary outcome is safety — adverse events graded by the standard CTCAE criteria, by type, grade, and frequency. That is the appropriate and honest framing for a first-in-human-style study of seven distinct engineered receptors in a heavily pretreated population. Adoptive cell therapies carry their own characteristic risks, from cytokine-driven toxicities to the effects of the conditioning regimen, and a Phase 1 study exists first to characterize those risks and establish that the approach can be delivered safely. Any antitumor activity observed would be encouraging but secondary to the safety question this trial is built to answer, and the sequential, non-randomized, open-label design reflects that priority.
From a landscape perspective, this trial is a marker of a broader shift in cell therapy. The first wave of engineered T cells, the approved CAR-T products, target proteins on the surface of cancer cells. Targeting intracellular neoepitopes from driver mutations like TP53 and RAS is a fundamentally different and more difficult game, because it depends on the cancer cell properly presenting those altered peptides on its HLA molecules — and it opens, in principle, a route to attacking mutations that have defied direct drugging. The disciplined read of NCT06904066 is that the NCI is testing whether T cells armed with receptors against p53 and Ras neoepitopes can be manufactured and infused safely in patients with difficult blood cancers. Whether they meaningfully control those cancers is the question the field is watching, and one that the safety-focused Phase 1 data will only begin to inform.