A lot of people will tell you there is nothing unusual happening with cancer right now. I don’t agree. For those of us who have been observing carefully, there are patterns emerging that should concern us. And one of the clearest signals came not from a laboratory or a hospital ward, but from the financial markets — when investors were spooked by the results of a major cancer blood-test trial.
When I say investors were spooked, I mean they were genuinely alarmed by the outcomes. And when people start losing money, narratives begin to fracture. Investors don’t tolerate stories that cost them financially. That, more than any academic debate, is what forces honest reckoning with data.
The Promise of a Blood Test for Cancer
The Galleri blood test, developed by GRAIL Inc., was designed to be a breakthrough in cancer screening. The concept is elegant: detect fragments of tumour DNA circulating in the bloodstream, identify a cancer signal, and predict where in the body it might be originating — all from a simple annual blood draw.
Before the pandemic, early studies showed strong performance for certain cancers. The original validation study, published in 2021 in the Annals of Oncology, reported an overall specificity of 99.5% and demonstrated cancer signal detection across more than 50 cancer types, with an overall accuracy of cancer signal origin prediction of 88.7%. There was real optimism that this technology could transform how we approach population-level cancer screening.
The theory made intuitive sense: detect cancer earlier, intervene earlier, reduce late-stage disease. It was a compelling vision.
Klein, Eric A., et al. "Clinical validation of a targeted methylation-based multi-cancer early detection test using an independent validation set." Annals of Oncology 32.9 (2021): 1167-1177.
What the Test Actually Measures
To understand what went wrong, you first need to understand what the test is actually doing. The Galleri test measures patterns of DNA methylation in cell-free DNA circulating in the bloodstream. It is not detecting tumours directly. It is looking for chemical patterns on DNA fragments — almost like barcodes — that suggest cancer-like tissue is releasing DNA into the blood.
Every day, cells in your body die and release small pieces of DNA into your bloodstream. This is normal. The test tries to distinguish between fragments from healthy tissue turnover and fragments that carry methylation signatures consistent with malignancy. It then uses machine learning to answer two questions: is there a pattern that looks like cancer, and if so, where in the body might it be coming from?
The test is making an educated guess based on patterns — it is not seeing cancer directly. And that distinction matters enormously when you start interpreting population-level results.
The Outcome That Changed the Narrative
On the 19th of February 2026, GRAIL announced the latest results from the landmark NHS-Galleri trial. The headline was optimistic: a substantial reduction in Stage IV cancer diagnoses, increased Stage I and II detection for deadly cancers, and a fourfold higher cancer detection rate compared with standard care.
It sounded like good news. Then the market reacted.
GRAIL’s share price collapsed — dropping from approximately $100 to $50 virtually overnight, a fall of around 50%. That is not a minor correction. That is a fundamental reassessment of the business model.
Why? Because beneath the positive framing, the trial did not meet its primary endpoint: a statistically significant reduction in combined Stage III and IV cancer diagnoses. That single result changed everything. Detection had increased, but the expected population-level shift away from late-stage disease had not materialised convincingly.
The Timing Problem Nobody Wants to Discuss
This is where I believe the trial leadership made a critical error — and where, if they had been listening to people like me, they could have probably saved millions of dollars.
The NHS-Galleri trial enrolled approximately 142,000 participants aged 50 to 77 across eight Cancer Alliance regions in England. Recruitment ran from 31st August 2021 to 16th July 2022.
If you don’t remember what was happening during that period, let me remind you. This was right in the middle of the Delta variant wave. It was also the period during which the COVID vaccination programme was being rapidly scaled up across the UK, with boosters being mandated in many countries by late 2021. By mid-2022, Omicron had replaced Delta, and the population had been through multiple waves of infection and immune stimulation.
The assumption behind any large screening trial is that background population biology is relatively stable. During this period, it was anything but stable. There were widespread infection cycles, repeated immune activation from both natural infection and vaccination, significant healthcare disruption, and major changes in how people presented with disease.
Yet the trial proceeded as if biology had returned to normal.
The Variable That Wasn’t Considered
When you examine what the trial investigators did account for, the attention to demographics was thorough. They looked at age, sex, ethnicity, smoking status, alcohol consumption, BMI, and deprivation indices. They used sophisticated algorithms to ensure diverse recruitment across socioeconomic groups. They even deployed mobile clinics in deprived neighbourhoods and offered language interpretation services.
All appropriate and commendable.
But one factor was conspicuously absent from the core analytical framework: vaccination status.
Now, I want to be clear about what I am saying here. I am not claiming that vaccination caused specific cancer outcomes. What I am saying is methodological. When you are conducting a population-level screening trial during a period of unprecedented immune modulation, and you do not stratify for immune exposures, you leave an unanswered question hanging over interpretation.
The trial designers likely relied on the assumption that randomisation would balance vaccination status across both arms. And in classical trial methodology, that is a reasonable position. But randomisation protects internal comparison between groups. It does not restore a stable biological baseline. It does not account for whether population-level immune changes altered the very signals the test was designed to detect.
I suspect they could do this analysis retrospectively. And if they are serious about understanding how to use this valuable technology going forward, they should.
What I Think We Are Actually Seeing
Since around 2022, I have been concerned about changing cancer patterns. I have followed the idea of what some call “turbo cancer,” though I think that term has caused more confusion than clarity. It is not that cancers are growing faster. It is that they appear to be spreading earlier.
What seems to be happening is that cancers are behaving more like melanoma. If you are unfamiliar with melanoma, it is one of the most dangerous cancers precisely because it disseminates early. By the time you notice the primary lesion — sometimes just a mole — it has often already spread through the lymphatic system. That is why melanoma is so difficult to cure with surgery alone and frequently requires chemotherapy or radiotherapy.
This pattern, where cancers spread far earlier than anticipated, is what appears to be emerging in the post-pandemic period. The first step is to acknowledge it. The next question is to ask why. And that second question seems to be the one that many people are reluctant to pose.
CD147 and the Invasiveness Question
In my recent work, I have been exploring the role of a receptor called CD147 — also known as Basigin or EMMPRIN. This is a transmembrane glycoprotein that sits at the intersection of tumour invasion, metabolic adaptation, and microenvironment remodelling.
I use a simple analogy. Think of a cell as a castle. Outside, you have viruses representing the invading army. The main entrance is ACE2 — the receptor we heard so much about during the pandemic. But CD147 functions like hidden tunnels leading inside the castle. Once those tunnels open, access multiplies. And critically, when SARS-CoV-2 gets inside cells via these tunnels, it appears to upregulate CD147 further, opening even more entry points.
In the context of cancer, this matters enormously. CD147 is a known factor in how invasive cancers can be. It drives the production of matrix metalloproteinases that break down surrounding tissue, supports lactate export that allows tumours to thrive under metabolic stress, and creates acidic microenvironments that suppress local immune surveillance.
If CD147 is upregulated — whether through viral infection, immune activation, or other biological pressures — you would expect to see cancers that invade and spread more readily. This aligns with what we are observing: cancers that reach systemic competence earlier than traditional staging models predict.
The Stage III Problem
This brings me to what I consider one of the most revealing details in the trial data. Look carefully at where the blood test struggled, and a pattern emerges.
The original validation study showed that detection sensitivity increases dramatically with stage: approximately 17% for Stage I, 40% for Stage II, 77% for Stage III, and 90% for Stage IV. The test is fundamentally tracking tumour burden — how much biological disturbance the cancer has already created — rather than tumour potential.
In the NHS-Galleri trial outcomes, Stage IV diagnoses fell in certain groups, but Stage III remained stubbornly persistent. If cancers are reaching systemic competence earlier — spreading through lymphatics and acquiring metastatic capability before generating large cfDNA signals — then a blood-based screening test would consistently catch them slightly too late.
You would see exactly the pattern the trial produced: enough detection to reduce some Stage IV diagnoses, but not early enough to prevent the underlying biological transition that defines Stage III disease.
This is not a technological failure. It is a biological mismatch. And it aligns perfectly with the hypothesis that cancers are behaving more invasively than our screening models were designed to detect.
Why Investors Reacted So Strongly
Markets are brutally honest interpreters of evidence. The dramatic stock collapse reflected a single, clear realisation: detection alone does not guarantee population benefit.
The commercial viability of MCED testing depended on demonstrating that the test changes outcomes at scale — that it reduces late-stage disease, improves survival, and justifies the cost of annual population screening. Without that proof, the path to NHS or insurer reimbursement becomes uncertain, cost-effectiveness models deteriorate, and the timeline to profitability extends dramatically.
Investors were not saying the science was wrong. They were saying the path to mass adoption just became much longer and riskier. And in financial terms, a longer, more uncertain timeline means a lower present valuation.
Where the Conversation Needs to Go
I am not arguing that the Galleri technology is useless. The underlying principles remain strong, and the test may well prove valuable for specific populations and cancer types. But we need a different approach going forward.
Any company conducting major trials on drugs, therapeutics, or diagnostic technologies that does not take into consideration the biological changes of the pandemic period is inviting problems. The population biology has shifted. Ignoring that is not neutral — it is risky, both scientifically and economically.
If I were advising these companies, I would say: identify which biological cohorts benefit most from this technology. Stratify populations more precisely. Integrate immune status and metabolic context into trial design. And above all, stop assuming that the biological baseline of 2019 still applies.
The Science Will Tell the Answer
My view at this point is straightforward. I am no longer too concerned about people who are stuck in ideological positions on either side of these questions. The science is going to provide the answers regardless of who agrees with what. The outcomes will be visible for everybody to see. There is nowhere to hide.
What matters now is finding people who are willing to look carefully at the data — because the science is relevant to the long-term outcomes of population health, and increasingly, to the bottom line of every company building plans on assumptions about cancer behaviour that may no longer hold.
If biology has changed, and we continue to ignore it, the consequences will show up everywhere: in clinical trial results, in treatment outcomes, in share prices, and ultimately, in the health of populations. The question is not whether we will eventually face this reality. The question is how much time and money we are willing to lose before we do.
Vejon COVID-19 Review is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
