CERVICAL CANCER: AGE AT REGISTRATION AND AGE AT DEATH

Summary.-The 5-year survival of women with localized (early-stage) cervical cancer is much higher than for women with non-localized (late-stage) cancer, but women with localized cancer tend also to be younger than those with advanced cancer. A new method of presenting the long-term survival is suggested, and the registrations of cervical cancers in South Wales are analysed in terms of average age at registration and average age at death. The observed average age at death was very close to 59 years regardless of stage (and age) at diagnosis, and calculations of expected ages at death of the whole populations suggest that more than half the advantage in survival shown by early stage cancers over late stage cancers is due to diagnosis of the former in younger women. CANCER of the cervix remains one of the most common cancers in women, although the crude mortality rate has been decreasing slowly during the last decade in England and Wales: from 10-4 per 100,000 in 1962 to 8-9 in 1973 (Registrar General, 1976). The registration rates are, how-ever, consistently some 80 % higher, at approximately 17-9 per 100,000 in 1962 and 16*3 in 1970 (OPCS, 1975). During that decade, screening by way of cervical cytology was introduced into many areas, with a somewhat ambiguous outcome (Br. med. J., 1973). It is not clear whether the mortality from cervical cancer is falling as a result of cervical cytology and intervention at the in situ stage or whether cervical screening has caused an increase in the proportion at diagnosis and registration of early stage cancers. What the cancer registry data do show is an increase in the registration rates of carcinoma in situ from negligible in 1962, through 1'6 per 100,000 in 1963, to a peak of 12*1 in 1967 and falling back to 8-9 in 1970 (OPCS, 1975). Cervical cancers were for the most part staged for cancer registration purposes according to internationally agreed criteria of tumour size, location and extent, until the Office of Population Censuses and Surveys (OPCS, 1970) relaxed the require-ment. The staging of cancer allows study of the comparative survival rates by stag