A novel and automatic mammographic texture resemblance marker is an independent risk factor for breast cancer M. Nielsen a,b, *, G. Karemore a,b , M. Loog a,b,c , J. Raundahl a , N. Karssemeijer d , J.D.M. Otten d , M.A. Karsdal b , C.M. Vachon e , C. Christiansen b a University of Copenhagen, Copenhagen, Denmark b Nordic Bioscience A/S, Herlev, Denmark c Delft University of Technology, Delft, The Netherlands d Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands e Mayo Clinic, Rochester, MN, USA 1. Introduction The second most common form of cancer in Europe is breast cancer with an estimated 429,900 cases in 2006 [1], and therefore screening and risk profiling are essential tools for early detection in order to improve survival [2]. The Gail model [3] is an established tool to establish the risk of breast cancer in clinical practice. Recent reports indicate that the addition of breast density and other risk factors to the Gail model may increase the ability to predict cases of breast cancer [4]. Among the numerous methodologies for density scoring, the categorical density scoring of the Breast Imaging Report and Data System 1 (BI-RADS 1 ) was originally proposed by the American College of Radiology for quantifying masking effects [5]. Before this, Wolfe patterns [6] were introduced to assess the risk of breast cancer. Both methods rely on an expert’s categorisation of the mammographic appearance. Advantages in terms of a continuous score, but still requiring the interaction of a radiologist, are provided by an interactive threshold measurement method that expresses the area of dense tissue as a percentage of the total breast area [7]. Although it is well established that density measurements are associated with breast cancer risk, still several aspects are questioned. Hormone therapy is known to increase breast density, certain regimes are known to relate to breast cancer risk [8], but the degree to which density and risk are causally related during hormone therapy is not known [9,10]. Recent studies have indicated Cancer Epidemiology 35 (2011) 381–387 A R T I C L E I N F O Article history: Accepted 29 October 2010 Available online 10 December 2010 Keywords: Breast cancer risk Mammographic density Texture CAD HRT A B S T R A C T Objective: We investigated whether breast cancer is predicted by a breast cancer risk mammographic texture resemblance (MTR) marker. Methods: A previously published case–control study included 495 women of which 245 were diagnosed with breast cancer. In baseline mammograms, 2–4 years prior to diagnosis, the following mammographic parameters were analysed for relation to breast cancer risk: (C) categorical parenchymal pattern scores; (R) radiologist’s percentage density, (P) computer-based percentage density; (H) computer-based breast cancer risk MTR marker; (E) computer-based hormone replacement treatment MTR marker; and (A) an aggregate of P and H. Results: Density scores, C, R, and P correlated (tau = 0.3–0.6); no other pair of scores showed large (tau > 0.2) correlation. For the parameters, the odds ratios of future incidence of breast cancer comparing highest to lowest categories (146 and 106 subject respectively) were C: 2.4(1.4–4.2), R: 2.4(1.4–4.1), P: 2.5(1.5–4.2), E: non- significant, H: 4.2(2.4–7.2), and A: 5.6(3.2–9.8). The AUC analysis showed a similarly increasing pattern (C: 0.58  0.02, R: 0.57  0.03, P: 0.60  0.03, H: 0.63  0.02, A: 0.66  0.02). The AUC of the aggregate marker (A) surpasses others significantly except H. HRT–MTR (E) did not significantly identify future cancers or correlate with any other marker. Conclusions: Breast cancer risk MTR marker was independent of density scores and more predictive of risk. The hormone replacement treatment MTR marker did not identify patients at risk. ß 2010 Elsevier Ltd. All rights reserved. Abbreviations: MTR, mammographic texture resemblance; C, categorical paren- chymal pattern scores; P, computer-based percentage density scores; H, computer- based breast cancer risk mammographic texture resemblance score; E, computer- based hormone replacement treatment mammographic texture resemblance score; A, aggregate of P and H; ROC, receiver-operator characteristics; AUC, area under the ROC curve; BI-RADS, Breast Imaging Report and Data System 1 ; HRT, hormone replacement treatment; TIMP, tissue inhibitors of matrix metalloproteinasis; ECN, extra cellular matrix. * Corresponding author at: Nordic Bioscience Imaging, Herlev Hovedgade 207, 2730 Herlev, Denmark. Tel.: +45 4454 7777; fax: +45 4454 8888. E-mail addresses: madsn@nordicbioscience.com, madsn@diku.dk (M. Nielsen). Contents lists available at ScienceDirect Cancer Epidemiology The International Journal of Cancer Epidemiology, Detection, and Prevention jou r nal h o mep age: w ww.c an cer ep idem io log y.n et 1877-7821/$ – see front matter ß 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.canep.2010.10.011