Review
Computer aids and human second reading as interventions in screening mammography: Two systematic reviews to compare effects on cancer detection and recall rate

https://doi.org/10.1016/j.ejca.2008.02.016Get rights and content

Abstract

Background

There are two competing methods for improving the accuracy of a radiologist interpreting screening mammograms: computer aids (CAD) or independent second reading.

Methods

Bibliographic databases were searched for clinical trials. Meta-analyses estimated impacts of CAD and double reading on odds ratios for cancer detection and recall rates. Sub-group analyses considered double reading with arbitration.

Results

Ten studies compared single reading with CAD to single reading. Seventeen compared double to single reading. Double reading increases cancer detection and recall rates. Double reading with arbitration increases detection rate (confidence interval (CI): 1.02, 1.15) and decreases recall rate (CI: 0.92, 0.96). CAD does not have a significant effect on cancer detection rate (CI: 0.96, 1.13) and increases recall rate (95% CI: 1.09, 1.12). However, there is considerable heterogeneity in the impact on recall rate in both sets of studies.

Conclusion

The evidence that double reading with arbitration enhances screening is stronger than that for single reading with CAD.

Introduction

In many countries, including the UK, it is standard practice for each screening mammogram to be viewed independently by two readers who either confer on discordant cases or refer them for arbitration. It is sometimes argued that this ‘double reading’ is too expensive or too demanding of radiologists’ time.1 An alternative is to use computer programs that process digitised mammograms and alert readers to possible abnormalities. A systematic review identified six studies comparing computer aids (CAD) to double reading but concluded that they were methodologically flawed and the evidence was limited.2 This paper takes a different approach: two sets of studies are reviewed:

  • studies comparing single reading with CAD to single reading without CAD;

  • studies comparing double reading to single reading.

We assess the impact of both interventions on cancer detection and recall rate since an improvement in cancer detection rate at the cost of an increased recall rate may not present an enhancement of the screening test.

Section snippets

Types of studies

Prospective and retrospective studies where the intervention was incorporated into routine screening work and all cases selected only on the basis of the usual screening criteria were included.

Types of participants

All the studies of women in a screening age range (aged 40 and above) were considered.

Types of interventions

Only the studies using commercially available CAD systems were included. The studies of double reading in which the second reader was a trained film reader but not a radiologist were included.

Types of outcome measures

Only the studies reporting

Description of studies

Ten prospective studies comparing single reading with CAD to single reading without CAD were identified,6, 7, 8, 9, 10, 11, 12, 13, 14, 15 and 17 comparing single reading with double reading.14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30

Impact of CAD and double reading on cancer detection and recall rate

Matched CAD studies measure its impact more directly, comparing assessments on individual images before and after looking at prompts. Although all of these studies show an improvement in cancer detection rate, none shows a statistically significant improvement and their combined effect is not statistically significant. Since it is impossible to detect fewer cancers after taking a second look than that were detected initially, these studies are biased in favour of CAD. They will not detect if

Conclusion

There is evidence that double reading increases cancer detection rate and that double reading with arbitration does so whilst lowering recall rate. There is insufficient evidence to claim that CAD improves cancer detection rates, but it does increase recall rate. Comparing CAD and double reading with arbitration, there is no difference in cancer detection rate, but double reading with arbitration shows a significantly better recall rate. Therefore, the best current evidence shows grounds for

Conflict of interest statement

None declared.

Acknowledgements

Dr. Given-Wilson provided advice. Dr. Liston supplied additional data. The work was partly supported by the NHS Breast Screening Programme which had no input into the research or presentation of results.

References (60)

  • E.R. Denton et al.

    Just how valuable is double reporting in screening mammography?

    Clin Radiol

    (1997)
  • B. Mucci et al.

    Double reading of screening mammograms: the use of a third reader to arbitrate on disagreements

    Breast

    (1999)
  • I. Doutriaux-Dumoulin et al.

    Cancers detectes par le deuxieme lecteur: analyse des donnees de la campagne de depistage du cancer du sein en Loire-Atlantique, 2003–2005 (nouveau cahier des charges)

    J Radiol

    (2007)
  • M.P. Becker et al.

    Marginal modeling of binary cross-over data

    Biometrics

    (1993)
  • M.J. Bradburn et al.

    metan – an alternative meta-analysis command

    STATA Tech Bull Reprints

    (1998)
  • M.J. Morton et al.

    Screening mammograms: interpretation with computer-aided detection – prospective evaluation

    Radiology

    (2006)
  • T.W. Freer et al.

    Screening mammography with computer-aided detection: prospective study of 12,860 patients in a community breast center

    Radiology

    (2001)
  • J.C. Dean et al.

    Improved cancer detection using computer-aided detection with diagnostic and screening mammography: prospective study of 104 cancers

    Am J Roentgenol

    (2006)
  • J.M. Ko et al.

    Prospective assessment of computer-aided detection in interpretation of screening mammography

    Am J Roentgenol

    (2006)
  • R.L. Birdwell et al.

    Computer-aided detection with screening mammography in a university hospital setting

    Radiology

    (2005)
  • J.J. Fenton et al.

    Influence of computer-aided detection on performance of screening mammography

    N Engl J Med

    (2007)
  • D. Gur et al.

    Changes in breast cancer detection and mammography recall rates after the introduction of a computer-aided detection system

    J Natl Cancer Inst

    (2004)
  • T.E. Cupples et al.

    Impact of computer-aided detection in a regional screening mammography program

    Am J Roentgenol

    (2005)
  • D. Georgian-Smith et al.

    Blinded comparison of computer-aided detection with human second reading in screening mammography

    Am J Roentgenol

    (2007)
  • M. Gromet

    Comparison of computer-aided detection to double reading of screening mammograms: review of 231,221 mammograms

    Am J Roentgenol

    (2008)
  • J. Brown et al.

    Mammography screening: an incremental cost effectiveness analysis of double versus single reading of mammograms

    BMJ

    (1996)
  • L.E. Duijm et al.

    Independent double reading of screening mammograms in The Netherlands: effect of arbitration following reader disagreements

    Radiology

    (2004)
  • R. Pauli et al.

    Comparison of radiographer/radiologist double film reading with single reading in breast cancer screening

    J Med Screen

    (1996)
  • J.M. Tonita et al.

    Medical radiologic technologist review: effects on a population-based breast cancer screening program

    Radiology

    (1999)
  • R. Renaud et al.

    Principes et premiers résultats de la campagne européenne de dépistage du cancer du sein dans le Bas-Rhin

    Bull Acad Natl Med

    (1991)
  • Cited by (122)

    View all citing articles on Scopus
    View full text