Developing and assessing a breast phantom for testing a fully automated density software: A
pilot study
Sammanfattning
Background: Mammographic density (MD) is an independent risk factor for breast
cancer and might be of particular role in the future of stratified screening for breast
cancer. MD can only be determined from a mammogram, by visual assessment, semiautomated
or automatic (volumetric) software. The correlation between compression
force and MD is not fully understood. This paper describes the pilot of a master thesis,
which was aimed at developing practical skills in designing phantoms for the main
experiment of the master thesis and to explore the relation between compression and
density at CT.
Material and methods: Poly (vinyl alcohol) (PVAL) phantoms with different designs
and complexity were created. The phantoms were CT scanned with and without compression.
The phantom’s density was measured by using Hounsfield Unit (HU). The
correlation was investigated by using the analysis of variance (ANOVA) and Tukey’s
Honestly significant different post-hoc comparison was conducted using SPSS.
Results: The pilot provided crucially needed experiences about the process of creating
PVAL phantoms. Reducing the thickness of the PVAL phantoms resulted in a decrease
in HU.
Conclusion: The experiences were used in the main experiment with multiple phantoms
with different designs. The main experiment and the outcome is described in
another paper.