What a Five-Year Cancer Survival Rate Actually Means

A five-year survival rate is a backward-looking count. It answers one question. Of the people diagnosed with this cancer during a past period, what share were still alive five years later? It is not a prediction. It is not a countdown. It is not a claim that anything happens at year five. Five years is simply a reporting convention. It lets registries compare different cancers and different eras on the same footing.

Most published figures are relative survival, not overall survival. SEER defines relative survival as "an estimate of the percentage of patients who would be expected to survive the effects of their cancer," and adds that "it excludes the risk of dying from other causes."

In practice, it compares people who have the cancer with people of the same age and sex in the general population, then divides out the background death rate. This is why a relative survival figure can reach 100 percent. SEER's current figure for localized female breast cancer is 100.0 percent. That means those women died at roughly the same rate over five years as women of the same age without breast cancer. It does not mean none of them died.

Across all cancer types combined, SEER reports 70.5 percent five-year relative survival for people diagnosed in 2016 to 2022.

Survival statistics need follow-up time before they can exist. To know who was alive five years after diagnosis, five years have to pass. Then the registry data are compiled, cleaned and published. So any figure you read describes people diagnosed roughly five to ten years earlier. They were treated with the drugs, radiation techniques and surgery of that period.

Some cancers have seen fast change in treatment: melanoma, lung cancer, kidney cancer, several lymphomas and myeloma. For those, the published number can badly understate what is possible now. Where little has changed, it stays close to current reality. Ask your oncologist which situation applies to your cancer. It is one of the more useful questions available.

A headline figure pools every stage, age, tumor subtype, molecular profile and level of background health. It includes someone diagnosed at 85 with heart failure and someone diagnosed at 40 with no other illness. It includes someone whose tumor was found on a screening test and someone who arrived in an emergency department with widespread disease.

SEER states the limitation plainly: "because survival statistics are based on large groups of people, they cannot be used to predict exactly what will happen to an individual patient." An average cannot tell you which side of it you are on. It describes a crowd, and you are not a crowd.

Registries group the extent of disease as localized, regional or distant. The gaps are enormous, and the headline figure hides them.

Female breast cancer (2016-2022): 91.9 percent overall — but 100.0 percent localized, 87.5 percent regional, 33.8 percent distant.

Pancreatic cancer (2016-2022): 13.7 percent overall — but 43.6 percent localized, 17.0 percent regional, 3.4 percent distant.

Someone with localized pancreatic cancer who reads "13.7 percent" is reading a number weighted by the large majority of cases found after spread. Someone with distant breast cancer who reads "91.9 percent" is reading a number dominated by early-stage disease. In both directions, the headline is the wrong figure for the individual.

Survival statistics are usually quoted from the day of diagnosis. But risk is rarely spread evenly across the five years. For most aggressive cancers the highest risk sits early. So surviving the first year or two removes a large part of the risk that the day-one figure was averaging in.

This is called conditional survival, and SEER*Explorer publishes it: five-year relative survival rates "conditioned on the patient having already survived 0, 1, 3, or 5 years since the cancer diagnosis." For many cancers, the five-year outlook for someone already two years out is much better than the figure they were given at diagnosis. It is a fair question to ask at follow-up appointments. The answer is often more encouraging than anything read on the day of diagnosis.

Statistics are useful for choosing between treatments and for planning. They are poor at telling you what will happen to you. Ask for figures specific to your stage and your subtype. Ask how old the underlying data are. Ask about validated prediction tools that take your own features into account. And ask how the picture changes with time survived.

For stage-by-stage SEER figures broken out by cancer type — localized, regional, and distant, for 22 cancers — see Understanding Cancer Survival Statistics, which links out to each cancer-specific survival and outlook page.

• SEER*Explorer — survival and conditional survival statistics
• SEER Cancer Stat Facts — Female Breast Cancer
• SEER Cancer Stat Facts — Pancreatic Cancer
• NCI — Understanding Cancer Statistics
• SEER — Types of cancer statistics