Elsevier

The Breast

Volume 15, Issue 1, February 2006, Pages 39-43
The Breast

ORIGINAL ARTICLE
Cancer detection and mammogram volume of radiologists in a population-based screening programme

https://doi.org/10.1016/j.breast.2005.04.005Get rights and content

Summary

This study investigates the relationship between the number of screening mammograms read by radiologists and the screening breast cancer detection rate. Cancer detection rates for incident screens (all women aged ⩾40 years) were compared by increasing categories of reader volume using Poisson regression. Data from New South Wales (NSW) for a 2 year period (2000–2001) were obtained from the BreastScreen NSW programme. Cancer detection rates increased with the number of mammograms read in the programme, reaching a plateau of approximately 40 per 10,000 after 1375 mammograms per year. No significant differences in cancer detection were evident above 875 mammograms (compared to below 875 mammograms) per year (RR=0.79, 95% CI 0.63–0.99).

Introduction

The efficacy of mammography screening in reducing breast cancer mortality has previously been shown in both randomised trials and service studies,1, 2, 3, 4 despite controversies regarding effect size and target age-groups.5, 6, 7 Randomised trials have shown that a 30% reduction in breast cancer mortality is achievable in women aged 50–69 years.3 BreastScreen New South Wales (BSNSW) is a population-based mammography screening programme that targets women aged 50–69 years. Women who attend BSNSW undergo bilateral two-view mammography, with all films read independently by two radiologists. Where there is disagreement between the two radiologists regarding a recall to assessment, the final recommendation is based on either the consensus opinion of both radiologists, or the recommendation of a third radiologist.

Radiologists in NSW contract their services to the State-funded screening programme with agreement to comply with established national accreditation standards and requirements.8, 9 Currently, the minimum number of reads per year required of radiologists in BSNSW is 2000,8 and was derived from considerations of standards in the United States (480 mammograms per year),10 the United Kingdom (5000 mammograms per year),11 and Canada (2000 mammograms per year).12

The aim of this study is to investigate the relationship between the number of mammograms read by radiologists and the breast cancer detection rate, and to determine if a cut-off point exists where the relationship changes significantly. This information is relevant to inform programme standards of minimum mammogram reads per radiologist for optimal cancer detection.

Section snippets

Data

The number of mammograms read for all incident screens and the number of cancers detected (invasive and ductal carcinoma in situ) were obtained for 134 radiologists contracted by BreastScreen NSW for the period January 2000 to December 2001 for all women (⩾40 years) attending 9 of the 10 Screening and Assessment Services (SAS) in New South Wales. One SAS was unable to participate in the study; this SAS conducted 11,055 incident screens, or 1.2% of all incident screens. Prevalent screens refer to

Results

Individual cancer detection rates ranged from 0 to 141 per 10,000 mammograms, however large variation in cancer detection was evident for low numbers of annual reads (<500) (Fig. 1). The mean cancer detection rate was 38.6 per 10,000 (95% confidence interval of 34.9–42.3) including all mammograms, and 42.0 per 10,000 (95% CI 39.4–44.6) excluding reader volumes of less than 500. Cancer detection rates increased with number of mammograms read, and peaked in quintile 2 (270–1459 mammograms per

Discussion

This study investigated the association between the number of mammograms read by radiologists in the screening programme and cancer detection rates for incident screens, using routinely collected mammography data from NSW (2000–2001). Cancer detection rates increased with the number of mammograms read, reaching a plateau at approximately 1300 mammograms per year. Poisson regression models of cancer detection rates for increasing cut-points of the number of mammograms read showed that cancer

References (16)

There are more references available in the full text version of this article.

Cited by (16)

  • Optimal screening mammography reading volumes; Evidence from real life in the East Midlands region of the NHS Breast Screening Programme

    2011, Clinical Radiology
    Citation Excerpt :

    An Australian study analysed the data from 133 radiologists working in a population-based screening programme for a 2 year period. No significant differences in cancer detection were evident above a film reading threshold of 875 mammograms/year; however, the study did not examine rates of recall to assessment.7 In a Canadian study, data from 304 radiologists working within organized screening programmes were analysed for a 3 year period.

  • Malignancy detection in digital mammograms. Important reader characteristics and required case numbers

    2010, Academic Radiology
    Citation Excerpt :

    No statistical differences for ROC scores between the first three categories were shown with very similar median values for each category of between 0.83 and 0.86; however, when the case number dropped below 1000 cases per year, ROC score dropped to 0.75, which was statistically below both categories ≥2000. This threshold at 1000 cases per year supports the current BreastScreen Australia national accreditation standard and is in broad agreement with earlier researchers in Australia, which showed that cancer detection dropped below 875 reads per year (21), and slightly less than that of Kan et al (20), who argued that cancer detection dropped below 2000. All three studies, however, have indicated higher levels of performance with annual case numbers well above the minimum value of 480 recommended in the United States.

View all citing articles on Scopus
View full text