Visually assessed breast density, breast cancer risk and the importance of the craniocaudal view

Stephen W Duffy¹ , Iris D Nagtegaal² , Susan M Astley³ , Maureen GC Gillan⁴ , Magnus A McGee⁵ , Caroline RM Boggis⁶ , Mary Wilson⁶ , Ursula M Beetles⁶ , Miriam A Griffiths⁶ , Anil K Jain⁶ , Jill Johnson⁶ , Rita Roberts⁶ , Heather Deans⁷ , Karen A Duncan⁷ , Geeta Iyengar⁷ , Pam M Griffiths³ , Jane Warwick¹ , Jack Cuzick¹ and Fiona J Gilbert⁴

¹Department of Epidemiology, Mathematics and Statistics, Wolfson Institute of Preventive Medicine, Charterhouse Square, London EC1M 6BQ, UK

²Department of Pathology, University Medical Center St Radboud, Geert Grooteplein 9, 6500 HB Nijmegen, the Netherlands

³Department of Imaging Science and Biomedical Engineering, University of Manchester, Oxford Road, Manchester M13 9PT, UK

⁴Department of Radiology, University of Aberdeen, Foresterhill Aberdeen AB25 2ZD, UK

⁵Department of Public Health and General Practice, Christchurch School of Medicine & Health Sciences, Riccarton Avenue, Christchurch 8140, New Zealand

⁶Nightingale Centre, Withington Hospital, Southmoor Road, Manchester M23 9LT, UK

⁷Northeast Scotland Screening Centre, Forester Hill Road, Aberdeen AB25 2XF, UK

author email corresponding author email

Breast Cancer Research 2008, 10:R64doi:10.1186/bcr2123


Published:	23 July 2008

Abstract

Introduction

Mammographic density is known to be a strong risk factor for breast cancer. A particularly strong association with risk has been observed when density is measured using interactive threshold software. This, however, is a labour-intensive process for large-scale studies.

Methods

Our aim was to determine the performance of visually assessed percent breast density as an indicator of breast cancer risk. We compared the effect on risk of density as measured with the mediolateral oblique view only versus that estimated as the average density from the mediolateral oblique view and the craniocaudal view. Density was assessed using a visual analogue scale in 10,048 screening mammograms, including 311 breast cancer cases diagnosed at that screening episode or within the following 6 years.

Results

Where only the mediolateral oblique view was available, there was a modest effect of breast density on risk with an odds ratio for the 76% to 100% density relative to 0% to 25% of 1.51 (95% confidence interval 0.71 to 3.18). When two views were available, there was a considerably stronger association, with the corresponding odds ratio being 6.77 (95% confidence interval 2.75 to 16.67).

Conclusion

This indicates that a substantial amount of information on risk from percentage breast density is contained in the second view. It also suggests that visually assessed breast density has predictive potential for breast cancer risk comparable to that of density measured using the interactive threshold software when two views are available. This observation needs to be confirmed by studies applying the different measurement methods to the same individuals.