443,451 cells from 62 tumours, and a cancer that keeps changing shape
What happened
Researchers in the United States have built the most detailed picture yet of how pancreatic cancer changes over time. A team led from NYU Grossman School of Medicine and its Perlmutter Cancer Center read the gene activity of 443,451 individual cells. Those cells came from 62 tumours, taken from 56 patients. Six of the patients gave samples both before and after chemotherapy, so the team could see what the drugs left behind. The study was published in the journal Nature Genetics on 13 October 2026. NYU Langone Health set out the findings in a press release that was reported more widely on 25 October. Most earlier studies ground up whole lumps of tumour and measured the average. This one looked at cells one at a time. That change of method is the reason the team could see something new: a single tumour is not one cancer. It is a crowd of slightly different cancers living side by side. The researchers found between 2 and 11 distinct groups of cancer cells, called clones, inside each tumour. Some clones were the gentler “classical” type. Others were the more aggressive “basal” type. Crucially, cells appeared able to switch from one state to the other as the tumour picked up new genetic changes. Two genes stood out. One is called MYC. It was busier in some of the most aggressive cell groups. When the team blocked MYC in laboratory experiments, basal cells drifted back towards the calmer classical state, and became easier for immune cells to see. The other finding involved two immune “switches” called TIGIT and PD1. Some clones had high TIGIT activity and low PD1 activity. That pattern may help explain a long-standing puzzle, which the next section sets out.
Why this matters
Pancreatic cancer is one of the hardest cancers to treat. The NYU team notes that nearly 85% of patients die within five years of diagnosis. Part of the problem is that treatments which work well elsewhere have not worked here. Immunotherapy drugs that target PD1 have transformed care for some skin and lung cancers. In pancreatic cancer they have largely disappointed, and nobody could fully explain why. This study offers a possible reason. If the cancer cells are leaning on TIGIT rather than PD1 to hide from the immune system, then a drug aimed at PD1 is knocking on the wrong door. That is useful in a very practical way. Aristotelis Tsirigos, one of the senior investigators, points out that trials of drugs aimed at MYC and at TIGIT are already under way for pancreatic and other cancers. So this is not a discovery that needs twenty years of groundwork before anyone can test it. The tools exist. There is a second reason the work matters, and it is about how doctors think. If a tumour is a shifting crowd rather than a fixed thing, then a single biopsy taken on a single day may not describe it well. Treatment that works for a while, then stops working, starts to look less like bad luck and more like evolution happening inside one person.
What the evidence actually says
It is worth being clear about what this study did and did not show. It showed structure. The 443,451-cell map is real data from real tumours, and the clone counts, the classical and basal mix, and the TIGIT and PD1 patterns come from that data. The MYC result is different. Blocking MYC pushed aggressive cells back towards a calmer state in laboratory experiments. That happened in the lab, not in patients. Nobody has yet shown that blocking MYC helps a person with pancreatic cancer live longer or feel better. The immunotherapy explanation is also a proposal, not a proven fact. The team suggests that high TIGIT activity may be why PD1 drugs have struggled. Testing that properly means running a trial, not reading a map. The numbers are modest, too. Fifty-six patients is a serious piece of work for this kind of deep analysis, but it is a small group. Only six patients gave matched samples from before and after chemotherapy. The authors themselves call for larger studies. Despoina Kalfakakou, who led the work, put the contribution carefully: “Our study provides the most granular view to date of pancreatic ductal adenocarcinoma, revealing how ongoing genetic reshaping drives tumour evolution over the course of disease.” She added that “targeting the mechanisms underlying tumour evolution, including MYC, may help restrain the cancer’s rapid progression.” Note the word “may”. That is the honest state of play.
Practical advice
Nothing in this study changes what any patient should do tomorrow. But a few things follow from it. If you or someone close to you is being treated for pancreatic cancer, it is reasonable to ask the specialist team whether any clinical trial is open and suitable. Trials of drugs aimed at MYC and TIGIT are running. Your team will know what is available locally and whether it fits your situation. Asking costs nothing. Do not change or stop any treatment because of a news story, including this one. The findings are about biology, not about a licensed medicine you can request. If you have symptoms that are new, persistent and unexplained, see your GP rather than waiting to see if they settle. That is standard advice for any possible cancer, and it is the single thing most within your control. If you are supporting someone through treatment, keep a simple written note of which drugs they are on, the doses, and when things changed. Specialist teams change over time, and a clear record helps every clinician who picks up the file. Finally, be patient with the science. A study that explains why something has failed is genuinely valuable, even though it does not hand anyone a new prescription this week.
What to know
Researchers read 443,451 single cells from 62 pancreatic tumours in 56 patients, and found each tumour held between 2 and 11 distinct groups of cancer cells. Cells could shift between gentler and more aggressive states as new genetic changes appeared. The gene MYC was busier in some of the most aggressive groups, and blocking it in the laboratory calmed those cells down. Some groups also leaned on an immune switch called TIGIT rather than PD1, which may help explain why PD1 immunotherapy has worked poorly in this cancer. The important limits: the MYC work was done in the laboratory, the immunotherapy explanation is a proposal that still needs testing, and the patient group was small. Trials of MYC and TIGIT drugs are already running, so the ideas can be tested without starting from scratch. Sources: NYU Langone Health / NYU Grossman School of Medicine, “Scientists discover how pancreatic cancer rewrites its DNA to escape treatment”, ScienceDaily, 25 October 2026, https://www.sciencedaily.com/releases/2026/10/261024221849.htm; Nature Genetics, “Clonal heterogeneity in human pancreatic ductal adenocarcinoma and its impact on tumor progression”, 13 October 2026, https://doi.org/10.1038/s41588-026-02779-5 This article is for general information and does not replace advice from a doctor, pharmacist or other qualified healthcare professional.
Related articles
48 hours from 25 November, and the first consultant strike since 2023
On Monday 26 October 2026, the British Medical Association (BMA) announced that consultants in England will go on strike in November. Consultants are the most…
76% of the dose, and a recall that goes all the way to the patient
On 26 October 2026, the Medicines and Healthcare products Regulatory Agency (MHRA) issued a recall for a liquid version of clopidogrel, a medicine that helps st…
1% against 3.3%, and the bleeding figure nobody should skip
Researchers at Monash University in Australia have reported that low-dose aspirin was linked to a much lower risk of dementia — but only in one particular genet…