Medicine's October: Cells, Genes and the Second Number

Imagine a woman in a clinic with a printed slide on her knee and one number circled in red: 42. Her oncologist has just said it, and she asks what every careful patient asks. Forty-two per cent of what? The clinic and the woman are invented. The question is not. It is the right question for one of the most striking medical results of the autumn, and it is the thread I want to follow through this month's laboratories.

Across 2025 and 2026, medicine moved several approaches it had long promised into patients' bodies. Some are cancer vaccines built for each patient, some are cell therapies, and some are gene edits made for a single child. None of these is a cure in the way a headline wants. Each sits at a different stage of evidence, so I take them one at a time, then ask who gets to stand in the queue.

The Number That Needs a Second Number

On 19 August 2026, Merck and Moderna announced that their phase 3 trial of an individualised cancer vaccine had met its primary endpoint. The trial, INTerpath-001, tested patients with completely resected stage IIB to IV melanoma. Those who received the combination of intismeran autogene and Keytruda (pembrolizumab) were compared with those who received Keytruda alone. The combination beat Keytruda alone on recurrence-free survival, the primary measure, and on distant metastasis-free survival, a key secondary measure, according to the companies' topline statement.

Intismeran autogene is made for each patient from the changes found in that person's tumour. A tumour carries mutations that its healthy neighbours do not, and some of those changes produce proteins, called neoantigens, that the immune system might learn to recognise. Think of a sentry who has never seen the intruder's face, now shown a photograph and told what to watch for. That is an analogy of mine, not a claim from any source.

The detailed results were presented at the European Society for Medical Oncology meeting in October 2026. According to the American Association for Cancer Research summary, the vaccine plus pembrolizumab reduced the risk of recurrence or death by 42 per cent compared with pembrolizumab alone, and the risk of distant metastasis or death by 38 per cent. Median follow-up was 24.6 months. The same summary reports that no new safety concerns were observed compared with previously reported studies of the combination, and that an overall survival benefit has not yet been demonstrated.

This is where the patient's question matters. A relative reduction of 42 per cent is a fine number, but it does not tell her how many people in her position will remain well, or for how long. The AACR summary gives no hazard ratio and no absolute rates. It also says that an earlier phase 2 trial of the combination reached five-year follow-up, reported at ASCO 2026, and that this analysis did not definitely demonstrate an overall survival benefit. My view, offered as opinion, is that the absolute difference will decide whether this becomes a standard therapy. The stage is clear: a positive phase 3 result for an investigational vaccine that is not yet approved for melanoma.

There is a second test I would apply to any trial that stops early or reports at an interim look. Does the benefit hold across the whole population studied, and does it survive longer follow-up? Those are the questions the next publication has to answer, and a reader can check for them.

"The first and only combination regimen to demonstrate a clinically meaningful improvement over pembrolizumab alone." That is how Moderna describes its own result.

Teaching a Cell to Work Inside the Body

Most CAR-T therapies approved so far work like a factory visit. A patient's T cells are collected, engineered in a laboratory to carry a receptor that recognises a target on cancer cells, expanded, and infused back. It works, but the process takes time, needs specialist facilities and is expensive. The newer idea, called in vivo CAR-T, asks a different question: what if the cell is engineered inside the patient, by a delivery vehicle that carries the instructions to the right cells?

That is a promise, not yet a proof. Early human studies of this approach are small by any measure, and the questions that matter most are still open: how long responses last, and which patients should receive the treatment. The sentry idea, in other words, has been shown the photograph but has not yet kept watch for years.

In my judgment, a small group with short follow-up can show that a cell can be engineered inside the body. It cannot yet show how long a response lasts or which patients should receive the treatment. The real test will come in larger, longer trials that enrol patients whose disease has already been treated, because those are the patients most clinicians will actually see.

For a reader, the useful habit is to ask three questions of any cell-therapy headline. How many patients were treated? How long were they followed? What happened to the patients who did not respond? A headline that answers the first question and avoids the other two is describing a beginning.

One Child, One Letter, and a Medicine Built to Order

The most intimate story of the year is also one of the oldest in this month's review. In 2025, doctors at the Children's Hospital of Philadelphia (CHOP) treated an infant with severe CPS1 deficiency, a rare inherited condition in which the liver cannot clear ammonia properly. The therapy was a personalised base-editing treatment, made for this child rather than taken from a shelf. The CHOP announcement is the primary account of the case.

Base editing is a gentler tool than the older scissors of CRISPR. Rather than cutting both strands of the DNA double helix, a base editor changes one letter to another, like correcting a single typo in a sentence without tearing the page. That precision is the reason the approach is attractive, and it is also why the safety record has to be read case by case.

In my reading, a medicine built for one child sits awkwardly with rules designed for products made in batches. A factory product can be tested in many patients before it reaches the clinic. A medicine made for one child cannot wait for that kind of evidence, and the regulator has to decide what counts as enough. The question, as I see it, is how a method for making individual edits can be assessed once, rather than judging each child's therapy from scratch. That is my judgment, not a finding of any regulator.

One child is not a trial. The next report from this family of cases should say what happened to the child over years and whether the approach has worked in others. Until that report exists, the honest sentence is the modest one: the case shows that a bespoke gene edit can be designed and given to a patient, and it does not yet show how well it works over time.

A Kidney as a Bridge

Pig organs have been the subject of hope and disappointment for a generation. This year brought a single case that tells the truth about both. On 3 September, The Lancet published a first-in-human study by Riella and colleagues of porcine kidney xenotransplantation as a bridge to a human kidney. The patient was Tim Andrews, aged 66 at the time of surgery. He received a gene-edited pig kidney from eGenesis in January 2025, under an FDA Expanded Access arrangement. The pig kidney kept him off dialysis for 271 days, which the study describes as the longest such period reported in a living person after pig-kidney transplantation. The press release reports the same account.

Then the story turned. After immunosuppression was reduced because of an infection, the graft developed microvascular injury and inflammation that progressed to organ failure. He returned to dialysis. In January 2026 he received a deceased-donor human kidney, which functioned immediately. The authors describe the pig kidney as a bridge, with destination therapy only a possibility. A separate case in 2024 lasted 52 days before the patient died of cardiac causes that the researchers judged unrelated to the transplant.

The study had one patient. Its funding came from Massachusetts General Hospital and eGenesis, and several authors report financial ties to the company. Those are reasons for patience, not for dismissal.

The bridge idea changes what success looks like. A bridge need not last forever; it has to carry a person across until the other side, a human organ, arrives. Imagine a traveller on a riverbank with a single plank across the water. Whether that plank becomes a permanent crossing is the question that any formal trial has to answer.

Formal trials are the next test, and the Andrews case shows why their design matters. In my reading, a trial has to show not only that a pig organ can work for a while, but why it fails when it does, how the immune drugs affect the outcome, and what the patient's options are once the organ is removed. The number of genetic edits made to a pig is a design choice, and the trials will have to show whether that choice helps or only adds complexity.

Who Pays for the Cure

Here the cell and the child meet the price list. Casgevy, a gene therapy for sickle cell disease, shows the question clearly. The question is not whether the science works, but whether a health system can pay for it.

In my reading, that is the African arithmetic in one sentence. The benefit is real for those who can reach it, and the price has to be negotiated before the benefit can be counted in a national budget. A clinic in Nairobi, Kampala or Lagos faces the same arithmetic as a clinic in Philadelphia, only with less money to absorb the gap. A cost-effectiveness study can show that a therapy is worth its price under a given set of assumptions; it cannot make the price smaller, and it cannot decide who pays.

There is also a misinformation problem that any reader in this part of the world will recognise. In my judgment, miracle cancer products circulate on social media faster than any regulator can respond, and a promised cure that arrives before any trial data should be treated as a warning in itself.

Free-cure stories travel faster than trial results, I would argue, because they promise what the careful findings cannot yet promise. The most useful thing a reader can do is ask two questions. What is the absolute benefit, and who will pay for it? Then check the trial registries and the national regulator before believing any headline, including this one.

The Cell That Does Not Know It Is in a Trial

Imagine a single lymphocyte under a scanning electron microscope, a small round soldier with a long memory. Imagine the cell being taught, in a laboratory or inside a patient, to recognise the face of a tumour the way a sentry learns the face of someone who should not be on the wall. The cell does not know it is in a trial. It does not know that its work is described as a 42 per cent reduction in risk, or that a second number is still missing. It knows only the shape of the thing it was taught to watch.

The people in the room do know. The woman with the slide needs the absolute rates before she can decide what the relative number means for her. The family of a child with a rare metabolic disease needs to know what happened to the child over time. The traveller on the plank needs to know whether it will hold. The clinic in Nairobi, Kampala or Lagos needs to know who will pay.

The cell will go on watching for its face. The rest of us have a smaller job: ask for the absolute difference, check whether a trial is open and where, treat the word "cured" as a question, and read the second number when it is published.

Editorial note: This essay expresses the founder’s ideas, developed with AI-assisted research and writing (parallel Claude Haiku 5.5 agents) and checked against the linked sources on 30 October 2026. Convictions, proposals and forecasts are identified in the text; scenes framed as imagined are illustrative. This article is journalism about research, not medical advice; evidence stages are stated in the text.

Cover photograph / illustration: NIAID (National Institute of Allergy and Infectious Diseases), via Flickr (https://www.flickr.com/photos/54591706@N02/53615359955/). CC BY 2.0. Original source. Illustrative image: a colorized scanning electron micrograph of a human T lymphocyte, the immune cell family that cell therapies such as CAR-T engineer. Photograph: NIAID. CC BY 2.0. This image does not document the events described.