Although shown to decrease breast cancer mortality, mammographic screening suffers from less than optimal sensitivity and specificity. This has led to the development of new breast screening technologies such as digital breast tomosynthesis (DBT) and dedicated breast CT. These new technologies have shown promise in early studies, but their innate complexity challenges optimization. Breast imaging represents a complex chain of image acquisition, display and interpretation that is best tested in a clinical trial. Conducting clinical trials comparing all potential configurations of DBT systems is simply not practical. In response, we propose an innovative system to perform virtual clinical trials (VCT) of breast imaging technologies. We envision VCTs being complementary to clinical trials, and having a critical role in preclinical testing of imaging devices, so that human clinical trials may be targeted to the most promising devices and appropriate clinical roles. This approach leverages unique expertise of the collaborators. The University of Pennsylvania has developed a 3D virtual breast phantom that is well suited for use in VCTs;it provides infinite anatomic variation and can simulate a multitude of breast lesions. The Barco MeVIC (Medical Virtual Imaging Chain) is among the most sophisticated observer modeling tools available, incorporating image acquisition, image processing, image reconstruction, image display, and the human visual and perceptual systems. We propose to integrate these technologies into a VCT system for digital mammography (DM) and DBT. The VCT system will be used to simulate an actual trial of DBT and DM being conducted by the ACRIN clinical trials network in order to compare relative performance in terms of sensitivity and specificity with an actual clinical trial. The richness of the VCT components will allow exploration of the broadest possible sampling of breast parenchyma, lesions, and observer performance. We propose the following. The existing voxel phantom and lesion models will be refined. The image simulation environment will be adapted to model the Hologic DM/DBT system. The existing MeVIC 2D channelized Hotelling observer (CHO) will be adapted to 3D (2D + time) by studying and then implementing the human spatio-temporal contrast sensitivity function. Various channeling mechanisms will be tested and perceptual factors such as degradation of memory with time will be included. 2D and 3D 2-AFC experiments will be conducted with human observers to estimate observer performance for each observer, modality and lesion type. The performance of the MeVIC observer models will be tested with real and synthetic DM and DBT images. The resultant validated 2D and 3D observer models will be used to conduct a VCT of DBT and DM screening. The results of the VCT will be compared to the results of the ongoing ACRIN PA clinical trial of DBT and DM screening. If successful, the VCT system can be utilized to test technology variations in breast cancer screening technology, accelerating development and clinical implementation of improved imaging systems in a more cost-efficient fashion.
Medical imaging is undergoing a rapid expansion both in terms of new devices and new methodologies;as the pace of medical device development increases, one is faced with the quandary of increasing the pace of clinical trials or finding effective and safe alternatives to some clinical trials. In response, we propose the development of an innovative system to perform virtual clinical trials (VCT) of breast cancer screening technology that builds on our prior work in developing a virtual breast phantom and complex observer model for 3D breast imaging. If successful, the VCT system can be utilized to test technology variations in breast cancer screening technology, accelerating development and clinical implementation of improved imaging systems in a more cost-efficient fashion.
|Karunamuni, Roshan; Naha, Pratap C; Lau, Kristen C et al. (2016) Development of silica-encapsulated silver nanoparticles as contrast agents intended for dual-energy mammography. Eur Radiol 26:3301-9|
|Borges, Lucas R; Oliveira, Helder C R de; Nunes, Polyana F et al. (2016) Method for simulating dose reduction in digital mammography using the Anscombe transformation. Med Phys 43:2704-2714|
|Lago, M A; Rúperez, M J; Martínez-Martínez, F et al. (2015) Methodology based on genetic heuristics for in-vivo characterizing the patient-specific biomechanical behavior of the breast tissues. Expert Syst Appl 42:7942-7950|
|Chen, Feiyu; Bakic, Predrag R; Maidment, Andrew D A et al. (2015) Description and characterization of a novel method for partial volume simulation in software breast phantoms. IEEE Trans Med Imaging 34:2146-61|
|Skoura, Angeliki; Bakic, Predrag R; Megalooikonomou, Vasilis (2013) Analyzing tree-shape anatomical structures using topological descriptors of branching and ensemble of classifiers. J Theor Appl Comput Sci 7:3-19|
|Young, Stefano; Bakic, Predrag R; Myers, Kyle J et al. (2013) A virtual trial framework for quantifying the detectability of masses in breast tomosynthesis projection data. Med Phys 40:051914|
|Lau, Beverly A; Reiser, Ingrid; Nishikawa, Robert M et al. (2012) A statistically defined anthropomorphic software breast phantom. Med Phys 39:3375-85|
|Pokrajac, David D; Maidment, Andrew D A; Bakic, Predrag R (2012) Optimized generation of high resolution breast anthropomorphic software phantoms. Med Phys 39:2290-302|