What does the "survival" term mean, and first, what does it not mean?
- Being alive 5 years after diagnosis does not mean being cured.
- Survival has no impact on longevity or life expectancy, which are unaffected by improved survival.
Consider a woman with a life expectancy of 65 years.
- A 90% survival rate at 5 years means that 9 out of 10 women will be alive after 5 years. But this is only the case for cancers that can be detected at an early stage, for those less aggressive.
Cancers with aggressive and progressive characteristics cannot be detected at an early stage, which would significantly improve their management, so survival rates are much lower…
A well-known bias that improves survival artificially.
This is known as "lead-time bias."
Survival is defined as the length of time a patient lives while being knowledgeable of his cancer. We anticipate the "start point" of cancer that would have manifested itself later in the absence of screening. The life expectancy of the patient with his cancer thus appears longer.
Here is a diagram to understand :
To use an analogy:
A train traveling to Paris derails in Orleans at 3 pm and causes the death of many passengers. If you get on that train in Bordeaux, you will live another three and a half hours. If you get on that same train in Tours, you will live another 30 minutes. However, your train will always derail at 3 pm.
Or better, let us use an analogy from Pr Michael Baum in a letter published in The Times, March 24, 2019
"Lead time reflects a frame shift in observing the natural history of the disease that can be understood with this analogy. If you get on a train bound for Edimburgh at Durham that crashes at Newcastle you live for 20 minutes yet if you board the same train at King Cross you live for two and a half hours. Over diagnosis results from the detection of sub-clinical foci of disease that microscopically look like cancer yet are not programmed to progress. These account for about 30% of screen detected "cancers" that are then over-treated by surgery, radiotherapy and chemotherapy, all of which have toxic consequences."
Survival depends mainly on two parameters that amplify it
1°- Therapeutic efficacy
The accumulation of epidemiological data provides, especially since 2015, strong evidence that improvements in patient management have played a significant role in the reductions in breast cancer mortality observed in Europe, Oceania, and North America, while the role of screening mammography is marginal.
While survival is an inappropriate criterion for measuring the effectiveness of screening, it remains the most widely used and most likely relevant marker for assessing the efficacy of a therapy.
2° - Overdiagnosis
The more lesions are overdiagnosed, which by definition do not cause death, the more lesions that would never have led to death are detected and counted, and the more survival will tend to be 100 %.
Let's use an analogy here as well.
If we treat all of the minor winter colds with triple antibiotic therapy, we can claim we have saved all patients from severe pneumonia. However, we know that the vast majority of seasonal colds heal on their own without the use of antibiotics. This is a misleading embellishment of the effect of human action, presenting it as beneficial when the counterpart will be the induction of numerous drug resistances.
Where does the 99% survival figure come from?
The figures come from American data from the American SEER program , but this survival is not the same depending on the stage of cancer:
- localized breast cancer: 99% survival at 5 years
- breast cancer with regional spread: 86% survival at 5 years
- breast cancer with distant metastases: 29% survival at 5 years.
When it is stated that the 5-year survival rate for breast cancer is 99 percent, it leads one to believe that 99 percent of cancers will be cured with screening. However, as we can see above, survival is better in the early stages than in the advanced stages; these advanced stages would benefit from early detection, but the real question is: is screening able to avoid the advanced stages?
Severe forms of cancers are often dangerous from the start, and because they are inherently aggressive, they progress quickly, so screening misses them . Screening is more likely to detect too many small, non-progressive, and non-threatening cancers, many of which are not necessary to detect. This ability to detect essentially non-progressive cancers increases overdiagnosis by creating the illusion of cures for lesions that would not have killed anyway and thus artificially improves survival rates.
Increased survival with no benefit to women
The two diagrams below (reproduced with the author's kind permission ) show how survival is “improved” with increasing overdiagnosis, without any benefit in terms of mortality for women.
Caution, this is a hypothetical scenario used to demonstrate how a survival rate can vary significantly depending on the extent of overdiagnosis. However, it shows how closely linked the two parameters are; the more significant the overdiagnosis, the greater the "improved" survival rate.
If survival is not a good criterion for screening effectiveness, then what are the good criteria?
The three main criteria for screening effectiveness are:
- a significant decrease in specific mortality
- the decrease of advanced forms of the disease
- the decrease of the heaviest treatments.
1. Localized cancer survival is excessively optimistic due to increased overdiagnosis and the resulting illusion of cure.
2. Survival would indeed be better in advanced forms, but screening is unable to detect them in time.
3. Survival is not a good marker of the effectiveness of screening but of the therapy efficacy.
 The Surveillance, Epidemiology, and End Results (SEER) program of the National Cancer Institute (NCI) is a source of epidemiologic information on cancer incidence and survival rates in the United States.