A tumour biology primer that starts with embryology and ends with why your DNA is trying to kill you, several thousand times a day, and mostly failing
Before we get into pathways and mutations — the kind of thing you can look up on Google or ask ChatGPT in thirty seconds — let's spend some time on something more useful: how to actually think about tumour biology, in a way that shapes how you make decisions as a clinician.
Where every cell came from
Start with embryology. Every complex multicellular organism, humans included, begins life as a single cell — the zygote. That one fact has a huge consequence: every cell in your body, no matter how specialized or how different it looks under a microscope, carries the exact same DNA. A skin cell and a neuron are running the same source code.
The zygote first divides into a featureless ball of identical cells — no head, no tail, no front, no back. Then something remarkable happens: the ball breaks symmetry. A head–tail axis appears. Anterior–posterior orientation appears. Left–right distinction, inside–outside distinction — all of it emerges out of what started as undifferentiated sameness. The cells sort themselves into three primitive germ layers — endoderm, mesoderm, ectoderm — and from the ectoderm, the neural crest arises. From there, cells specialize, tissues organize, organs form.
All of this is a fight against entropy. Left alone, complex systems drift toward chaos and decay. To resist that, a multicellular organism needs a stable internal environment, coordination across millions of simultaneous functions, and constant, controlled exchange of matter and energy with the outside world. None of that is possible without one thing underneath it all: cooperative cellular behavior. Every cell has to know when to specialize, when to divide and replace a damaged neighbor, and — critically — when to die.
What a tumour actually is
Strip away the complexity and a tumour is simple: an abnormal mass of tissue caused by unregulated cell division. If it stays put — doesn't invade nearby tissue — it's benign. If it invades surrounding structures, and potentially spreads to distant sites, it's malignant.
The question everyone asks — and the question we should be asking
People want a single cause for cancer. There's a comforting logic to it: find the one cause, find the magic-bullet cure. Every few months a headline announces some lab has "found the cure for cancer," feeding exactly that hope.
But that's the wrong question. The right question is the opposite one: why doesn't cancer happen constantly? Because once you look at the numbers, the odds actually favor cancer forming, not the other way around.
The staggering scale of cell division
You have trillions of cells, and billions of them divide every single day, each division running through a tightly choreographed sequence called the cell cycle.
The history here is worth knowing. In 1824, French biologists Prévost and Dumas were the first to actually observe cell division — before that, the working assumption was that development happened through the growth of a single undifferentiated blob of tissue. They showed it was division and multiplication, not growth of a blob. Later in the 1800s, German anatomist Walther Flemming described mitosis in detail, breaking it into four visible stages: prophase (chromatin condenses into chromosomes), metaphase (chromosomes line up at the cell's equator, attached to spindle fibers from either pole), anaphase (sister chromatids are pulled apart toward opposite poles), and telophase (the chromatids arrive, nuclear envelopes reform, and the cell is ready to physically split — cytokinesis).
But here's the catch: those four stages are only what's visible under a microscope. They're the tip of the process. The full picture — the molecular preparation happening before any of this becomes visible — wasn't worked out until 1953, when Howard and Pelc used radioactive DNA labeling to map it. What Flemming saw was only the M phase. The rest of the cycle includes G0, G1, S, and G2.
G0 (quiescence) is the resting phase — the cell isn't dividing, isn't preparing to replicate DNA, but is still metabolically active. Some cells park here permanently (neurons). Others can be called back into the cycle when needed (epithelial cells, lymphocytes, hepatocytes).
G1 (the first growth phase) is where the cell grows, builds proteins, adds organelles, and assembles the machinery it'll need for DNA replication. Buried in late G1 is one of the most important decision points in the entire cycle: the restriction point, discovered by Arthur Pardee in 1974 and now known as Pardee's point. Before this point, the cell still needs growth factors, adequate size, and favorable conditions to proceed. The moment it crosses Pardee's point, though, the decision is irreversible — the cell commits to dividing and will carry through S, G2, and M even if growth factors disappear afterward.
S phase (synthesis) is where the actual DNA replication happens — the entire genome duplicated, producing sister chromatids, with centrosomes duplicating alongside.
G2 (the second growth phase, or pre-mitotic phase) is quality control: the cell checks the freshly copied DNA for errors, repairs what it can, builds spindle proteins, and grows a bit more before hitting the G2/M checkpoint.
Then comes M phase — mitosis, the four visible stages Flemming originally described, ending in a fully divided cell.
The math that should terrify you
The single most important requirement of this whole process is that DNA replicates perfectly, so that both daughter cells inherit identical copies. It doesn't. Every single division introduces not one or two, but millions of DNA errors — and somewhere between 1,000 and 10,000 of those are the specific kind of errors typically seen in malignant cells.
Do that math across billions of divisions happening every day, and the conclusion is unavoidable: your body is generating somewhere in the range of 1,000 to 10,000 potentially malignant cells every single day. By the raw numbers, cancer should be a constant, ongoing event. It isn't. So something extraordinary is holding the line — and it isn't one thing. It's six layers, stacked on top of each other.
Layer One: Ultra-efficient DNA repair
To understand why this layer matters so much, remember what DNA is: a double helix, two strands running in complementary directions, held together by base pairing — adenine with thymine, guanine with cytosine — via hydrogen bonds. The precise sequence of these bases is what eventually becomes RNA, and then protein. Change the sequence, even slightly, and you can change the protein. The cell's entire identity and future behavior is written into that sequence.
And that sequence is fragile. Damage comes from everywhere. Replication itself introduces errors — the replication fork can stall, leaving unstable single-stranded DNA exposed and vulnerable, bases can be mis-incorporated or skipped or duplicated. The cell's own metabolism generates reactive oxygen species as an unavoidable side effect of using oxygen for energy, and those species chemically damage bases. And then there's everything from outside: ionizing radiation, UV light, environmental toxins, chemotherapy, mutagenic chemicals — all capable of distorting, oxidizing, or outright breaking the DNA strand.
Different kinds of damage need different repair crews. A single altered base gets pulled and replaced by base excision repair. Bulkier distortions — the kind UV light tends to cause — need nucleotide excision repair. Mismatches introduced during replication are caught by mismatch repair, a system you'll recognize clinically from its failure state: Lynch syndrome and microsatellite instability. Single-strand breaks are handled routinely by several overlapping pathways. The most dangerous lesion of all is the double-strand break, where both strands sever and the chromosome is physically split — repaired either by non-homologous end joining (fast, but sloppy) or homologous recombination (slow, but accurate). And some especially nasty lesions — DNA-protein crosslinks, interstrand crosslinks — require multiple systems working in coordination, including the Fanconi anemia pathway.
Layer Two: Cell cycle checkpoints
If damage does occur, the cell cycle itself has brakes built in — molecular checkpoints run by p53, RB, ATM, ATR, and CHK1/2. Their job is to halt the cycle the moment DNA damage, replication errors, or spindle defects are detected, buying time for repair before the cell is allowed to proceed. Lose these checkpoints, and cells start dividing with damaged DNA intact — one of the earliest steps on the road to a tumor.
Layer Three: Apoptosis
When damage can't be repaired, the cell is supposed to destroy itself — programmed cell death, via one of two routes. The intrinsic pathway runs through mitochondrial BAX/BAK activation, triggering a caspase cascade. The extrinsic pathway runs through death ligands like Fas and TRAIL binding to surface receptors. If repair fails, self-destruction is the fallback — and it's one of the single most important cancer-prevention systems the body has.
Layer Four: Immune surveillance
Say a cell slips past all of that — repair fails, checkpoints miss it, apoptosis doesn't trigger. The immune system is the next gatekeeper, scanning for abnormal antigens and neoantigens using cytotoxic T-cells, NK cells, and macrophages.
This plays out as a three-act process called immune editing: elimination, where immune cells wipe out the emerging clone; equilibrium, a long standoff where the tumor and the immune system essentially fight to a draw; and escape, where the tumor finally evolves a way around immune detection. That equilibrium phase can stretch on for 7 to 10 years before a tumor ever becomes clinically detectable.
This has real clinical weight. If a teenage patient receives radiation — for lymphoma, say — and a radiation-induced mutant cell forms that same day, it may take 7 to 10 years before it becomes apparent. That timeline is exactly why long-term screening and survivorship programs matter as much as they do.
Layer Five: Cellular senescence
Even a cell that survives all of the above doesn't get to divide forever. Senescence caps unlimited proliferation, through two routes: replicative senescence, driven by telomere shortening, and stress-induced senescence, triggered by oncogene activation or accumulated damage. A normal cell gets roughly 50 to 70 divisions before telomere loss shuts the process down. Cancer cells get around this limit by switching on telomerase, or by using the ALT pathway instead.
Layer Six: The microenvironment
The last line of defense isn't inside the cell at all — it's the surrounding tissue architecture. Contact inhibition, cell polarity relative to the basement membrane and lumen, limited physical niche space (think intestinal crypts), and tightly controlled growth-factor signaling all work together to box in abnormal cells before they can expand. Disturb that architecture, and an abnormal clone suddenly has room to grow.
Putting it together
Once you stack all six layers, a different picture of cancer comes into focus. It isn't a single failure, a single mutation, a single bad break. It's a multi-step escape from an extraordinarily robust, redundant defense system — one that's already stopping thousands of potential cancers a day without you ever knowing it happened.
Understanding tumour biology, in the end, isn't about memorizing every pathway in this list. It's about understanding how the layers interact, and exactly where and how cancer manages to slip through the gaps. That framework is what actually shapes how we think about radiosensitivity, fractionation, long-term cancer risk, screening strategy, and treatment design — the parts of the job that matter once you're standing in front of a patient.