Why Can’t Scientists Agree on What Counts as Cancer?

The Label That Won’t Stick

In 2015 a huge team of scientists published a massive catalog of mutations found in head and neck cancers. They had hoped to tidy the disease up — to find a clean genetic fingerprint for each type. Instead they discovered that cancers linked to smoking and cancers linked to a virus were wildly different, even though they grew in the same part of the body. And within each group, the tumors varied further. The catalog didn’t settle any sorting problem. It made the sorting problem feel much bigger.

That is the puzzle at the heart of cancer research: you can look at two clumps of cells, both called “cancer,” and they may have almost nothing in common except that they grow out of control. Some spread through the body like wildfire; others sit quietly for decades and never cause harm. Some have one dominant mutation; others have thousands. So how many diseases is cancer really? And why is answering that question so much harder than it looks?

What Makes a Disease a “Kind”?

Philosophers have a name for categories that carve nature at its joints: natural kinds. Think of chemical elements. Every atom of gold has the same number of protons. That one fact gives you a crisp, shared feature that makes gold gold. No matter where you find it, it’s the same stuff.

But diseases like cancer seem to be a different kind of category. The contemporary philosopher Marc Lange argues that cancer isn’t a natural kind at all. The “typical patterns of disruption” that make cells grow too fast are multiply realized — they can happen through many different routes. There’s no single broken part, no one master switch. When a medical textbook says “breast cancer,” it doesn’t point to a single destructive machinery; it points to a crowd of somewhat similar disruptions that happen to show up in the same organ. Lange even says modern medicine is leading toward “the end of diseases” as natural kinds, because we keep splitting them into hundreds of molecular subtypes that don’t share an essence.

Not everyone agrees. Some philosophers think cancer can still be a natural kind if we stop looking for one simple essence. Muhammad Ali Khalidi, a contemporary philosopher, borrows an idea called a homeostatic property cluster. If a bunch of features tend to travel together — like how various mutations all tend to disable the cell’s built-in safety switches — then the category can count as a natural kind, even without a single golden property. Think of a flock of birds: no one bird is the flock, but they cluster and move together. Khalidi points to “caretaker” genes that normally tell damaged cells to self-destruct. When those caretakers break, a whole set of cancer hallmarks — such as ignoring signals to stop growing or building their own blood supply — tends to appear. The cluster is kept together by the breakdown of a few key control mechanisms. That, he says, is enough for a kind to be real.

The debate matters because it shapes what scientists expect to find. If you believe cancer must have one hidden essence, you might pour all your energy into finding it — and miss the fact that the disease may simply be many loosely connected things.

When a Cancer Isn’t Really a Cancer

Here’s a surprising fact: more than half of men over sixty have small clumps of abnormal cells in their prostate, but most of them will never develop deadly cancer. Similar “lazy” growths appear in the thyroid and breast. These are called indolent lesions — they sit still, or grow so slowly that a person will die of something else long before the lump causes trouble.

This raises a thorny question: should we even call these things cancer? The philosopher Robert Schwartz argues that early-stage, non-aggressive growths are not disease. They are not yet disrupting normal function in a way that’s unusual for a person’s age. On his view, they should be reclassified as risk factors — like high cholesterol, which raises your chance of a future heart problem but isn’t itself the heart attack. Calling them cancer, Schwartz says, leads to overdiagnosis and overtreatment: people get surgery, radiation, or anxiety for a condition that would never have harmed them.

Lynette Reid, another philosopher, pushes back. She points out that diagnosis is never a pure, value‑free measurement. It’s a judgment call that balances risks and benefits, and those calculations depend partly on what we value — how precautionary we want to be, how much we fear missing a deadly case versus harming a healthy person. The line between “risk factor” and “disease” is not just a scientific fact waiting to be discovered; it’s also a choice with consequences for real patients.

A Thousand Clues, No Single Answer

The dream of precision medicine was that if we read every mutation in a person’s tumor, we could sort cancers into perfect categories — each one matching a targeted drug. Yet the closer researchers look, the messier it gets.

One complication is intratumor heterogeneity: a single tumor isn’t made of one kind of cancer cell. It’s a miniature forest of slightly different cells, each with its own set of accumulated changes. A sample taken from one edge might tell a totally different genetic story from a sample taken from the center. And tumors that arise in the same organ can be more genetically different from each other than tumors that arise in completely different organs. That means genomic information doesn’t draw a clean family tree; it gives us what philosophers call cross-cutting classifications — groups that overlap and slice through each other, like sorting toys by color, size, and material all at once.

Then there’s a deeper twist: context changes everything. Some mutations that look terrifying in a breast cell are completely harmless in a skin cell. Certain mutations only become dangerous after chemotherapy is applied, because the drug kills off all the cells except the ones that happen to be resistant. Even the order in which mutations appear — gene A first, then gene B, or B first, then A — can change whether a patient develops a slow-growing or an aggressive disease. So focusing only on a static list of broken genes can blind us to the dynamic, unfolding story of a disease.

Why It Matters for You

You might think these are abstract, nerdy debates that happen only in philosophy journals. But they land in real clinics every day. When a screening test finds an indolent lump, the words a doctor uses — “this is cancer” versus “this is a slow-growing change we should watch” — can decide whether a person spends long months treating something that would never have bothered them. The very definition of cancer determines who gets counted as a patient, which drugs get developed, and how terrified a family feels.

Cancer teaches us that not all important categories have sharp edges. That doesn’t make the science weak; it makes the world interesting. The job of sorting cancers isn’t like sorting marbles, where each one is clearly red or blue. It’s more like sorting clouds — shapes that shift, blend, and change depending on the weather. A good label is one that helps you predict what will happen and choose what to do, even when the clouds refuse to stay still.

Think about it

1. If a tiny lump in your body will probably never grow or cause harm, do you want to be told you have cancer, or would you rather a doctor call it a “risk shadow” and keep watching?
2. A disease category that is too broad might freak people out unnecessarily; one that is too narrow might miss something dangerous. How would you decide where to draw the line?
3. If two cancers look completely different under a microscope but act the same in the body, should they get the same name? What matters more in naming a disease — its cause or its behavior?