Chapters Index (French) French Version Welcome page Contact

Delay effect and screening evaluation

Generally speaking, when patients have their cancer detected through a screening procedure, they survive longer than patients for whom diagnosis has been established after the appearance of symptoms. The diagram below shows that the time gained may be:

• either a genuine benefit for the patient (for a long evolving cancer for which treatment at a precocious stage will be more efficient and will bring less morbidity): for instance, for cervix uteri carcinoma, early screening (in situ carcinoma) will induce local treatment which may not compromise the reproductive possibilities of the patient
  
• or solely an advance in diagnosis without any modification in therapy. For instance, treatment of leukaemia at the pre-leukaemic stage is not very different from treatment at the leukaemic stage and will always involve relatively high morbidity.

What proves the efficiency of a screening policy is the diminution of mortality

The diagram below does not show any advantage in earlier screening since the mortality is identical between a population submitted to a screening policy and a population where no such policy has been instituted.

Among the reasons for such equivalence,

• screening reveals cancer at stages that are just as advanced as those of fortuitous or symptom-related discoveries,
• the cancer spontaneously has a slow evolution and screening does not bring a real benefit,
• the screening technique may have noticeable drawbacks (!)
• the treatment involves heavy morbidity, regardless of the stage of the treated cancer (i.e. leukaemia).

So, there is no real practical advantage in detecting this type of cancer (cf. the discussion about prostate cancer screening: some surgeons think that only slowly evolving cancers are detected by the PSA technique, the patient is not going to die from his cancer and surgeons do not dare to offer unnecessary treatment with evident morbidity).

Screening is of real interest if it brings longer survival for the whole population with a clear decrease in the mortality of the screened cancer. This is the classical scenario where, generally, small cancers are more often cured using simpler treatment modalities. As a matter of fact, the increasing number of small cancers in the screened population will bring a much higher survival curve than in the non screened population.

For the cancer screened on the above diagram, screening is very efficient with smaller tumours benefiting from treatment. Survival is clearly increased both on median survival and on the number of surviving patients 5 or 10 years after treatment.