Early indicators of efficacy of breast cancer screening programmes. Results of the florence district programme

Abstract

A mammographic breast cancer screening programme has been ongoing in the Florence District (Italy) since 1970 and a favourable impact of screening on breast cancer mortality of women aged 50-70 has been shown by means of a case-control study. Two hundred and eleven screen- and 116 interval-detected cancers in the period 1975-1986 have been identified, and detection rates calculated, for first and repeated screening test (2nd to 7th). Overall, 22,980 subjects were screened and 44,988 repeated tests performed. The observed number of interval cancers has been compared with the expected incident cancers and their ratio (O/E) studied at different time intervals since last test. The O/E ratio at the third year since the last test was 0.98 for the age-group 40-49 0.50 (95% CI: 0.23-0.95) and 0.39 (95% CI: 0.26-0.94) for the 50-59 and 60-69 groups, respectively. The prevalence/incidence ratio (P/I) was then calculated as an early indicator of efficacy. For the 40-49 age-group the P/I ratio at first test was 1.09, suggesting poor anticipation of diagnosis. In contrast, for women 50-59 and 60-69 results suggest quite a good diagnosis anticipation (P/I: 3.14; 4.82), confirming the result of the previous case-control study on mortality reduction. The proportion of advanced carcinomas (stage II or worse) and 5-year survival have been analysed and discussed. The study confirms the opportunity of using early indicators of screening efficacy for monitoring of screening services.

Age-specific interval breast cancers in New South Wales and meta-analysis of studies of women aged 40–49 years

“…Results from service studies that include women aged <50 years indicate significant breast cancer mortality reduction. 28,29,31,38 The biology of breast cancer and the results of screening trials in women aged <50 years suggest that screening intervals of less than two years would be required for mortality reduction (such as the 18 month interval in the Gothenburg study). 9,10,[22][23][24][25][26] One factor for consideration when considering systematic screening of women aged 40-49 years is the magnitude of interval cancer rates achievable locally, because trial results may not be readily replicated in the service context.…”

Comparison of standard and double reading and computer-aided detection (CAD) of interval cancers at prior negative screening mammograms: blind review

“…In the Florence District programme (Paci et al, 1999), the sensitivity of biennial screening (with single reading) of women aged 50 -69 years has been calculated on the basis of proportional interval cancer incidence and resulted to be 75%. Based on the expected distribution of interval cancer by error type (FN ¼ 11.9%, MS ¼ 26.1% and occult ¼ 61.9%) (Ciatto et al, 1995a, b) and on the results observed in the present study (sensitivity increase at CAD-assisted reading; FN ¼ þ 12.4%, MS ¼ þ 8.4%), the absolute increase in screening sensitivity of CAD-assisted reading as compared to single reading might be estimated around þ 0.9%.…”

“…Sensitivity estimates based on proficiency tests of screening mammography (Ciatto et al, 1996;Ciatto et al, 1999), or according to interval cancer proportional incidence, suggest that approximately one in four cancers may be missed by biennial mammographic screening in women aged 50 -69 years (Paci et al, 1999). Adequate training is currently recommended to improve diagnostic accuracy, as well as double reading (DOUBLE) (Bird, 1990;Kirkpatrick et al, 1992;Thurfjell et al, 1994;Ciatto et al, 1995a, b;Kopans, 1998).…”

Breast density as a determinant of interval cancer at mammographic screening

“…Breast radiological density may have a masking effect on small cancer lesions; it is likely to be a determinant of false-negative mammography (Peeters et al, 1989), and might be associated with IC, particularly in cases reviewed as 'occult'. Other variables, such as age (Wald et al, 1993), use of single or double view (Bryan et al, 1995) and use of single or double reading (Ciatto et al, 1995a, b) have been shown to be associated with mammography sensitivity and might also be associated to IC.In Florence District, mammographic screening has been ongoing at the Centro per lo Studio e la Prevenzione Oncologica (CSPO) since 1970, being extended to Florence City in 1990.Screening efficacy has been demonstrated by means of a casecontrol study (Palli et al, 1986), and the proportional IC rate in women aged 50 -69 years has been reported to be 18% (95% CI 10 -24) and 42% (95% CI 29 -59) in the first or second year of the interval, respectively: the sensitivity of biennal screening, estimated on the basis of proportional IC incidence, was 75% (Paci et al, 1990).In the present study, a consecutive series of IC was compared to screen-detected cancers (SDC), and the association of breast density and other variables to IC was determined by univariate and multivariate analysis. Parameters significantly and independently associated to IC were identified, and the related implications on screening modalities were discussed.…”