Research Article

Comparison of breast carcinomas diagnosed in the 1980s with those diagnosed in the 1940s to 1960s.

BMJ 1991; 303 doi: http://dx.doi.org/10.1136/bmj.303.6795.155 (Published 20 July 1991) Cite this as: BMJ 1991;303:155
  1. H Joensuu,
  2. S Toikkanen
  1. Department of Radiotherapy and Oncology, University Central Hospital of Turku, Finland.

    Abstract

    OBJECTIVE--To find out if breast carcinomas diagnosed in the 1980s differ from those diagnosed a few decades ago. DESIGN--Retrospective comparative cohort study. SETTING--City of Turku, south western Finland. PATIENTS--439 patients with breast carcinoma diagnosed in 1945-65 with a median follow up of 27 years (95% of all those histologically diagnosed in Turku); and 370 patients with breast carcinoma diagnosed in 1980-4 with a median follow up of 6 years (94% of all those histologically diagnosed in Turku). MAIN OUTCOME MEASURES--Breast cancer incidence, mortality from breast cancer, age at diagnosis, stage of cancer, seven histological factors, DNA ploidy. RESULTS--Age adjusted incidence of breast cancer increased from 30.8/100,000 person years in 1953-7 to 62.2/100,000 in 1983-7; mortality in breast cancer increased from 16.7 to 17.2/100,000 person years. Survival of patients with stages II to IV (but not with stage I) improved. The mean age at diagnosis increased from 55.5 years in 1945-55 to 62.5 years in 1980-4 (p less than 0.0001); the proportion of patients with T0-1 carcinomas increased from 13% to 41% (p less than 0.0001) and with pN0 carcinomas from 43% to 55% (p = 0.003) from 1945-65 to 1980-4. There was no change in histological type or DNA ploidy of breast cancer, but in the 1980-4 cohort carcinomas were more often well differentiated, had lower mitotic counts and less nuclear pleomorphism, more often had a well defined tumour margin, and had less tumour necrosis. There was, however, no difference between the two cohorts in these histological characteristics except for tumour necrosis when they were compared in a multivariate log linear model at a given size of primary tumour. CONCLUSION--Improved survival with breast cancer can at least partially be explained by detection of increased numbers of small carcinomas with favourable histological characteristics.