Abstract

Abstract Introduction: A decrease in mammographic density is increasingly used as an intermediate marker to evaluate the effect of interventions for reducing breast cancer risk. Technical aspects, such as a different amount of compression, could influence density measurements. We investigated the inter-scan reproducibility of mammographic density measurements to see whether this may influence measurements of true changes in density. Methods: We included women with two mammograms made closely after one another (median 12 days), with no true changes in breast density expected to happen: 106 had two Full-Field Digital Mammograms (FFDM), 12 had two Film-Screen Mammograms (FSM) and 52 had one FFDM and one FSM. We estimated intraclass correlation coefficients, and mean absolute density differences between both mammograms. Results: The intraclass correlation between percent density measurements on two FFDMs was 0.93 (95%CI 0.91-0.95), between two FSMs 0.94 (95%CI 0.84-0.98), and between FFDM and FSM 0.72 (95%CI 0.61-0.80). The mean absolute difference in percent density between two FFDMs was 4.3% (95%CI 3.8-4.9) and between two FSMs 6.5% (95%CI 3.4-9.6). On average, percent density in the same woman was higher on FFDM than on FSM (difference 12.9% (95%CI 11.0-14.8)). Conclusion: When studying changes in mammographic density, it is important to compare two mammograms of the same type (FSM or FFDM). Even then, although intraclass correlation coefficients are high, the absolute differences in percent density are still of a magnitude that may influence detectability of true changes in mammographic density. Citation Format: Mariëtte Lokate, Marleen Emaus, Michiel Kallenberg, Marije Bakker, Nico Karssemeijer, Wouter Veldhuis, Petra Peeters, Carla van Gils. Inter-scan reproducibility of mammographic density measurements. [abstract]. In: Proceedings of the Eleventh Annual AACR International Conference on Frontiers in Cancer Prevention Research; 2012 Oct 16-19; Anaheim, CA. Philadelphia (PA): AACR; Cancer Prev Res 2012;5(11 Suppl):Abstract nr A116.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call