Interval cancers, positive predictive value for malignance and other early quality indicators in mammographic screening for breast cancer

• The study was designed to evaluate early quality indicators in screening for breast cancer. lt comprised, (i) investigations of compliance, recall, referral, cancer detection, and interval cancer rates in the Östergötland screening programme of women aged 40-74 years; (ii) evaluation of the three-stage diagnostic procedure for pre-operative assessment of women with suspicious mammographic findings; (iii) comprehensive analysis of all invasive interval cancers detected; and (iv) investigation into how to reduce the number of interval cancers without a concomitant increase in false positives.The attendance rates were 85.5% at the initial screen and 81.3% at subsequent screens. The referral rates for further examination at subsequent screens were roughly half of th'lt at the initial screen. The cancer detection rates were 6.4/1000 at the initial and 2.6/1000 at subsequent screens. The positive predictive value for malignancy at surgery was 96% for ages 50-74 years. Apart from the tumour grade (not analysed) the diagnostic outcome tallied with the quality targets stipulated by Tabár et al.The study confirmed the less favourable prognosis for interval cancer patients and cancers in non-attenders compared to screen-detected patients, but the mode of detection was not an independent predictor of metastatic capacity (hazard rate ratio [RR] of distant recurrence /adjusted for other variables/ RR=1.39, 95%confidence interval [95%CI] 0.78-2.46 for interval cancers; and RR=1.6, 95%CI 0.76-3.36 for non-attenders). The higher metastatic potential in these tumours could be explained by the differences in tumour characteristics at the time of diagnosis.The incidence risk of interval cancer was 0.46/1000 for tumours detected within 1 year of the latest screen, and 1.2/1000 for tumours detected within 2 years of the latest screen. Despite the lower age-specific breast cancer incidence in women aged 40-49 years, these women ran roughly the same risk of interval cancer after a negative screen as did the other age groups in the screening programme.Patients with potential iatrogenic delay in diagnosis (overlooked or misinterpreted cancers) constituted 25% of all patients with invasive interval cancer, and patients with true interval cancer 49%. The radiological category of interval cancer had no significant influence on the survival (overall comparison, p=0.1202; comparison of true interval with missed interval cancers, p=0.3175). In the present study there was no clear evidence of difference in prognosis between true interval and overlooked or misinterpreted interval cancers.The interval between the latest screen and diagnosis was not an independent prognostic factor in patients with true interval cancers (RR=0.47, 95%CI 0.16-1.35 for tumours detected 21 year after the latest screen), and there was no significant difference in survival according to this interval (comparison of tumours detected <1 year with tumours detected 21 year of the latest screen, p=0.3844).The study confirmed the association of criteria for referral for further examination with number of false positives. Efforts to reduce the number of interval cancers by lowering the mammographic threshold for recall are likely to be counterproductive. The early quality indicators constitute. an excellent means of monitoring of the quality of screening.

Nyckelord

Mammography

Breast neoplasm

Breast cancer screening

Interval breast cancer

Metastasis

Mortality

MEDICINE

MEDICIN

Publikations- och innehållstyp

vet (ämneskategori)

dok (ämneskategori)