Cancer Treatment: How Each Weapon Works

TL;DR. There are six ways to attack cancer, and each exploits a different weakness. Surgery removes it. Radiotherapy shatters its DNA with precisely aimed energy. Chemotherapy poisons dividing cells, hitting the tumour hardest because it divides fastest, which is also exactly why it causes hair loss, mouth ulcers, and low blood counts. Hormone therapy starves cancers that depend on oestrogen or testosterone. Targeted therapy blocks the specific broken protein driving one tumour, which is why it requires knowing the tumour's genetics. Immunotherapy does not attack the cancer at all: it removes the brake the cancer put on your own immune system. Understanding which mechanism is in play tells you what to expect, including which side effects are unavoidable consequences of the mechanism working.

Key takeaways

• Chemotherapy's side effects are not incidental toxicity; they are the same mechanism acting on your other fast-dividing tissues: bone marrow, hair follicles, gut lining.
• Targeted therapy needs a target. Sequencing the tumour is now part of diagnosis, because a drug matched to a driver mutation can work spectacularly and the same drug in an unselected patient does nothing.
• Checkpoint inhibitors release the immune system rather than attacking the tumour, which is why their side effects are autoimmune: the immune system attacks the thyroid, bowel, skin, liver, or lungs.
• Early palliative care extends life. In a randomised trial in metastatic lung cancer, patients getting early palliative care alongside standard treatment had better quality of life, less depression, and lived about 2.7 months longer.
• Most cures come from combinations, because a single agent selects for the cells that resist it.
• Financial toxicity is a measurable clinical outcome. Cancer treatment cost is a leading cause of personal bankruptcy in countries without universal coverage, and cost drives people to stop treatment.

Surgery

In short: Still the treatment that cures the most people, and the trend for forty years has been doing less of it, safely.

The oldest treatment and still the one that cures the most people. If the entire tumour and a margin of normal tissue can be removed before it has spread, the patient is cured. Surgery also stages the disease (sampling lymph nodes tells you how far it has gone) and relieves symptoms in advanced disease.

The trend over the last forty years has been toward doing less, safely. Radical mastectomy, which removed breast, chest wall muscle, and axillary nodes, was standard for most of the twentieth century until randomised trials showed that lumpectomy plus radiotherapy gave the same survival. Sentinel node biopsy, in which a dye or tracer identifies the first node draining the tumour so only that one is removed unless it is involved, spared enormous numbers of patients the lifelong arm swelling (lymphoedema) of full node clearance. Laparoscopic and robotic approaches reduce recovery time.

Limitations: it cannot treat disease that has already spread microscopically, which is why adjuvant (after surgery) drug therapy exists, and it is bounded by what can be removed without destroying function.

Radiotherapy

In short: Radiation breaks DNA, and splitting the dose across weeks lets normal tissue repair between fractions while cancer cells accumulate damage.

How it works. High-energy X-rays, gamma rays, or particles deposit energy in tissue, ionising water molecules and generating free radicals that break DNA strands. Cells with badly broken DNA die when they next try to divide. Cancer cells are more vulnerable because they divide more often and because their DNA repair machinery is frequently defective already.

Fractionation is the central trick. The total dose is split into many small daily doses over weeks, which allows normal tissue, with intact repair machinery, to recover between fractions while cancer cells accumulate damage. This is why radiotherapy is a six-week appointment rather than a single treatment.

Precision is the other. Modern techniques shape the beam to the tumour and spare surrounding tissue: intensity-modulated radiotherapy, image guidance, stereotactic radiotherapy delivering very high doses to small targets in a few sessions, and proton therapy, which deposits its energy at a defined depth and stops, sparing tissue behind the tumour. Protons matter most in children and in tumours next to critical structures.

Used for: curative treatment (prostate, head and neck, cervix, early lung), after surgery to sterilise the tumour bed (breast), before surgery to shrink a tumour (rectal), and for symptom relief, where a single fraction to a painful bone metastasis often produces rapid relief.

Side effects are local and follow the tissue in the beam: skin reaction, fatigue, mucositis for head and neck, bowel irritation for pelvic treatment, and late effects months to years later including fibrosis, and a small risk of a second cancer.

Chemotherapy

In short: It poisons dividing cells, which is exactly why hair, gut lining, and bone marrow suffer, and why combinations beat single drugs.

How it works. Cytotoxic drugs interfere with DNA replication or cell division. Because cancer cells divide more often, they are hit disproportionately, but the drug does not know what a cancer is. Everything in the body that divides quickly is affected too.

Why the side effects are what they are:

Nausea, once the defining misery of chemotherapy, is now largely controlled by modern antiemetics (5-HT3 antagonists such as ondansetron, NK1 antagonists such as aprepitant, and steroids). Growth factor injections (G-CSF) shorten the period of low white cells. Chemotherapy given in combination is the norm, because different mechanisms hit cells at different points, and because resistance to several drugs at once is far less likely than resistance to one. That principle, established in childhood leukaemia in the 1960s, is why some cancers became curable at all.

Hormone therapy

In short: Some cancers depend on oestrogen or testosterone, so removing the hormone stalls them, at the cost of a medical menopause or androgen deprivation.

Some cancers depend on a hormone for growth. Remove the hormone and the cancer stalls.

Breast cancer that is oestrogen receptor positive (about 70 percent of cases) is treated with tamoxifen, which blocks the receptor, or aromatase inhibitors (anastrozole, letrozole), which stop the conversion of androgens into oestrogen in postmenopausal women. Five to ten years of endocrine therapy after surgery substantially reduces recurrence and death. The trade-off is menopausal symptoms, bone loss with aromatase inhibitors, and a small risk of uterine cancer and clots with tamoxifen.

Prostate cancer depends on testosterone. Androgen deprivation therapy uses drugs that shut down the pituitary signal to the testes, plus newer agents (abiraterone, which blocks androgen synthesis everywhere including inside the tumour, and enzalutamide, which blocks the receptor). Side effects mirror the mechanism: hot flushes, loss of libido and erectile function, muscle loss, bone loss, fatigue, and metabolic changes.

Hormone therapy is not chemotherapy, does not cause hair loss, and is often given for years.

Targeted therapy

In short: Match the drug to the tumour's specific broken protein, which is why sequencing the tumour is now part of diagnosis rather than a luxury.

How it works. Identify the specific broken protein driving this tumour and design a drug that binds it. The result can be dramatic, and it only works in patients whose tumour carries that alteration, which is why tumour sequencing has become routine.

The founding case: imatinib. Chronic myeloid leukaemia is caused by a chromosomal swap that fuses two genes, BCR and ABL, producing a permanently active enzyme that drives white cell proliferation. Imatinib, approved in 2001, sits in that enzyme's ATP pocket and switches it off. Ten-year survival in CML rose from roughly 20 percent to over 80 percent, and most patients take a daily tablet and live a normal lifespan. It turned a fatal leukaemia into a chronic condition and set the template for everything that followed.

PARP inhibitors deserve a note because they illustrate an elegant idea called synthetic lethality. Cells have two main DNA repair systems. BRCA-mutated cancer cells have already lost one. Blocking PARP disables the backup, so the cancer cell dies while normal cells, which still have both, survive. The drug is selectively lethal only in the context of the mutation.

Antibody-drug conjugates are a hybrid: an antibody that binds a tumour surface protein carries a potent chemotherapy payload directly to the cell, delivering a dose that would be intolerable if given systemically.

Resistance is the rule. Targeted drugs select for cells with a second mutation that restores the pathway or bypasses it. Second and third generation inhibitors, and combinations blocking two points in one pathway, are the response.

Immunotherapy

In short: These drugs do not attack the cancer; they release the brake the cancer applied to your immune system, which is why their side effects are autoimmune.

How it works. Cancer cells display abnormal proteins and should be visible to the immune system. Many tumours survive by exploiting the immune system's own off-switches, which exist to prevent autoimmunity. Checkpoint inhibitors block those switches.

• CTLA-4 is an early brake on T cell activation. Ipilimumab blocks it.
• PD-1 on T cells binds PD-L1, which many tumours display in quantity, effectively showing a "friendly" badge. Pembrolizumab and nivolumab block PD-1; atezolizumab blocks PD-L1.

The consequences are unlike anything before. In metastatic melanoma, a disease with a median survival under a year in 2010, a substantial minority of patients treated with checkpoint inhibitors are alive with no evidence of disease a decade later. Similar long-tail survival appears in some lung, kidney, bladder, and head and neck cancers.

But not for everyone. Response rates in unselected patients are often 15 to 30 percent. Predictors of response include high tumour mutation burden (more abnormal proteins to recognise), PD-L1 expression, and mismatch repair deficiency, which produces so many mutations that response rates can exceed 50 percent regardless of where the cancer started. That last point produced the first approvals based on a tumour's genetics rather than its organ of origin.

CAR-T cell therapy goes further: T cells are collected from the patient, genetically engineered to display a receptor recognising a protein on the cancer cell, expanded in a lab, and infused back. In refractory B-cell leukaemias and lymphomas it produces remission in patients who had exhausted every other option. It has so far worked much less well in solid tumours, which hide behind dense stroma and a hostile local environment.

Side effects are autoimmune, by mechanism. Releasing the brakes on the immune system means it may attack normal tissue: colitis, hepatitis, thyroid failure (often permanent), pituitary inflammation, pneumonitis, rash, and type 1 diabetes. Most are manageable with steroids if caught early, and some are life-threatening. CAR-T carries two specific dangers: cytokine release syndrome, a massive inflammatory response, and neurotoxicity, both requiring intensive care management.

Bone marrow and stem cell transplantation

For leukaemias, lymphomas, and myeloma. Autologous transplant harvests the patient's own stem cells, gives very high-dose chemotherapy that would otherwise permanently destroy the marrow, then returns the cells to rescue it. Allogeneic transplant uses a donor's stem cells, and its benefit comes partly from high-dose therapy and partly from the graft-versus-tumour effect, in which donor immune cells attack residual cancer. The price is graft-versus-host disease, in which those same donor cells attack the patient's skin, gut, and liver, which can be chronic and disabling. Allogeneic transplant remains one of the most toxic treatments in medicine and one of the few that cures otherwise incurable disease.

Supportive and palliative care

Palliative care is not end-of-life care, and the confusion costs lives. It is specialist management of symptoms, side effects, and the practical and emotional burden of serious illness, delivered alongside active treatment.

The evidence is unusually clear. Temel and colleagues randomised patients with newly diagnosed metastatic lung cancer to early palliative care alongside standard oncology care or standard care alone. The early palliative care group had better quality of life, less depression, received less aggressive treatment at the end of life, and lived about 2.7 months longer on average. A supportive intervention outperformed the survival benefit of several drugs approved in the same period.

Also in this category: pain control (including opioids, whose availability is severely restricted in many countries, leaving millions to die in pain), nutrition support, management of nausea and fatigue, psychological support, and honest advance care planning.

What treatment costs, beyond side effects

Financial toxicity is a recognised clinical problem with measurable outcomes. Newer cancer drugs frequently cost six figures per year. Patients who face high out-of-pocket costs are more likely to skip doses, delay care, deplete savings, and go bankrupt, and financial distress independently predicts worse survival. In countries with universal coverage the burden shifts to the system, appearing as access delays and restricted formularies.

Time toxicity is the newer concept: how many days of a limited remaining life are spent in clinics, scanners, and infusion chairs. For a treatment that extends life by six weeks, that number matters and is rarely presented.

What the patient can do

In short: Get the tumour properly tested, ask whether the goal is cure or extension, and ask for absolute numbers rather than whether it works.

• Get the diagnosis right before treatment starts. Ask whether the tumour has been tested for the markers relevant to it (hormone receptors and HER2 in breast cancer, EGFR/ALK/PD-L1 in lung, mismatch repair in colorectal), because they change treatment entirely.
• Ask for the goal in plain terms. Is this treatment intended to cure, to extend life, or to relieve symptoms? Patients and oncologists frequently believe different things here, and studies show a large share of patients with incurable cancer believe their treatment may cure them.
• Ask about absolute benefit. "How many months, on average, and what proportion of people benefit at all?" is a better question than "does it work?"
• Ask about a clinical trial. Trial participation is how treatment improves, and outcomes for participants are generally at least as good as standard care.
• Deal with fertility before starting, if relevant. It is time-critical and easy to miss.
• Keep moving. Exercise during and after treatment reduces fatigue, preserves muscle, and is associated with better outcomes in breast and colorectal cancer cohorts.
• Stop smoking, even after diagnosis. It improves treatment response and survival, and it is often assumed to be pointless at that stage, which is wrong.
• Be sceptical of anything sold as an alternative to treatment. A cohort study found that patients who chose alternative medicine instead of conventional treatment for curable cancers had substantially higher death rates. Complementary approaches for symptoms and wellbeing are a different matter and can be genuinely useful alongside treatment.

What's next

• Multi-cancer early detection blood tests, screening for circulating tumour DNA from many cancers at once. Promising and unproven: the crucial question is whether they reduce mortality or mainly generate overdiagnosis and anxiety.
• Circulating tumour DNA to guide treatment, detecting microscopic residual disease after surgery and identifying resistance mutations from a blood draw rather than a biopsy.
• Therapeutic cancer vaccines, including individualised mRNA vaccines encoding a patient's own tumour mutations, in trials combined with checkpoint inhibitors.
• Next-generation cell therapies engineered to work in solid tumours.
• Better KRAS drugs, opening the largest previously undruggable target.
• De-escalation trials, deliberately testing whether less surgery, less radiation, or shorter chemotherapy gives the same result with less harm. This is where a lot of the real quality-of-life gain now comes from.

Sources and notes

Mechanisms and drug classes follow standard oncology references (DeVita, Hellman, and Rosenberg, Cancer: Principles and Practice of Oncology). Imatinib and CML survival: IRIS trial long-term follow-up, NEJM, 2017. Early palliative care: Temel et al., NEJM, 2010 (2.7-month survival difference). Alternative medicine outcomes: Johnson et al., JNCI, 2018, and JAMA Oncology, 2018. Checkpoint inhibitor long-term melanoma survival: pooled nivolumab and ipilimumab follow-up analyses. Mismatch repair deficiency and tissue-agnostic approval: Le et al., Science, 2017. Breast surgery de-escalation: NSABP B-06 and successors. Financial toxicity: Ramsey et al., Health Affairs, 2013, and subsequent work. Patient understanding of curative intent: Weeks et al., NEJM, 2012.

Open questions. Why checkpoint inhibitors work spectacularly in a minority and not at all in most patients is only partly explained. Whether multi-cancer early detection tests save lives is unknown. Optimal duration of many targeted and immune therapies has never been formally tested, and patients are often treated longer than may be necessary.

Next: the other half of the world's disease burden, the one caused by things that are alive. 👉