Room: Davidson Ballroom B
Purpose: At present, there are no suitably complex phantoms commercially available for image quality optimization and ultrasound scanner selection for specific clinical applications such as breast ultrasound imaging. Such task-specific phantoms are important as they can enable an objective evaluation of image optimization of ultrasound systems using targets mimicking clinically-relevant pathology within tissue mimicking material (TMM). As well as replicating the different tissue typesâ€™, anthropomorphic and sonographic characteristics. This study aimed to design and fabricate task-specific anthropomorphic breast phantoms and make an initial assessment of their utility for breast ultrasound (US) image quality optimization.
Methods: Design specifications for anthropomorphic breast phantoms were developedÂ¹ through consultation with radiologists, breast US sonographers and medical physicists from a large radiology department. The specifications included thickness and acoustic characteristics of the skin, fat and fibroglandular tissues, as well as the number, location, size, echogenicity, margin characteristics, and associated artifacts of â€œmust seeâ€? lesions. These lesions included cancers with spiculations and angular margins; lipid, sebaceous and acorn cysts; fibroadenomas; lymph nodes; Mondorâ€™s disease; and papillary lesions.
Results: Based on these specifications, five phantoms were constructed from agar and polyvinyl-alcohol-based TMMs which included the different breast tissue types and between 6 and 13 lesions with the specified anthropomorphic and acoustic characteristics. An acoustic macroscope verified that the target acoustic properties were achieved for the different breast tissue types and lesions. The initial evaluation of these phantoms by a breast radiologist found very good clinical resemblance of between 60â€“92% of the â€˜must-seeâ€™ lesions.
Conclusion: The developed phantoms were successful in replicating the anatomical, morphological and sonographic appearance of â€œmust seeâ€? lesions and were found to be suitable for assessing image optimization through identification of individual lesion features. Other applications of these phantoms are training and demonstration of new imaging modes. Â¹Browne et al 2017 UMB
Funding Support, Disclosures, and Conflict of Interest: The authors which to acknowledge funding received from the Mayo Clinic Radiology Internal Grant, Grant Number RDCRNJH1 43487
Not Applicable / None Entered.