Objective: The two objectives of this study were to create an ex vivo phantom model that closely mimics human breast cancer for detection tasks and to compare the performance of full-field digital mammography with screen-film mammography in detecting and characterizing small breast masses in a phantom with a spectrum of complex tissue backgrounds.

Materials And Methods: Sixteen phantom breast masses of varying sizes (0.3-1.2 cm), shapes (round and irregular), and densities (high and low) were created from shaved tumor specimens and imaged using both full-field digital and screen-film mammography techniques. We created 408 detection tasks that were captured on 68 films. On each radiograph, six detection tasks were partially obscured by areas of varying breast-pattern complexity, including low (predominantly fatty), mixed (scattered fibroglandular densities and heterogeneously dense), and high (extremely dense) density patterns. Each detection task was scored using a five-point confidence scale by three mammographers. Receiver operating characteristic (ROC) curve analysis was performed to analyze differences in detection of masses between the two imaging systems, and sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were computed.

Results: Full-field digital mammography showed higher area under the ROC curve than screen-film mammography for detecting masses in each breast background and performed significantly better than screen-film mammography in mixed (p = 0.010), dense (p = 0.029), and all breast backgrounds combined (p = 0.004). Full-field digital mammography was superior to screen-film mammography for characterizing round and irregular masses and low- and high-density masses.

Conclusion: Full-field digital mammography was significantly superior to screen-film technique for detecting and characterizing small masses in mixed and dense breast backgrounds in a phantom model.

Download full-text PDF

Source
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1892902PMC
http://dx.doi.org/10.2214/AJR.05.0126DOI Listing

Publication Analysis

Top Keywords

full-field digital
24
screen-film mammography
24
digital mammography
20
detection tasks
12
mammography
11
mammography screen-film
8
small masses
8
phantom model
8
mammography detecting
8
detecting characterizing
8

Similar Publications

Purpose: Digital breast tomosynthesis (DBT) has been introduced more than a decade ago. Studies have shown higher breast cancer detection rates and lower recall rates, and it has become an established imaging method in diagnostic settings. However, full-field digital mammography (FFDM) remains the most common imaging modality for screening in many countries, as it delivers high-resolution planar images of the breast.

View Article and Find Full Text PDF

 Synthesized mammography (SM) refers to two-dimensional (2D) images derived from the digital breast tomosynthesis (DBT) data. It can reduce the radiation dose and scan duration when compared with conventional full-field digital mammography (FFDM) plus tomosynthesis.  To compare the diagnostic performance of 2D FFDM with synthetic mammograms obtained from DBT in a diagnostic population.

View Article and Find Full Text PDF

Introduction: Research concerning artificial intelligence in breast cancer detection has primarily focused on population screening. However, Hong Kong lacks a population-based screening programme. This study aimed to evaluate the potential of artificial intelligence-based computer-assisted diagnosis (AI-CAD) program in symptomatic clinics in Hong Kong and analyse the impact of radio-pathological breast cancer phenotype on AI-CAD performance.

View Article and Find Full Text PDF

This study examines the specimen size-dependent deformation behavior of commercially pure titanium grade 4 (cp-Ti grade 4) sheets under tension, with strain paths between uniaxial tension (UT) and plane-strain tension and compares the results with cyclic bending under tension (CBT) data. Specimens of varying widths (11.7, 20, 60, 100, and 140 mm) were tested in both rolling (RD) and transverse (TD) directions.

View Article and Find Full Text PDF

Eddy-resolving turbulence simulations are essential for understanding and controlling complex unsteady fluid dynamics, with significant implications for engineering and scientific applications. Traditional numerical methods, such as direct numerical simulations (DNS) and large eddy simulations (LES), provide high accuracy but face severe computational limitations, restricting their use in high-Reynolds number or real-time scenarios. Recent advances in deep learning-based surrogate models offer a promising alternative by providing efficient, data-driven approximations.

View Article and Find Full Text PDF

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!