The long term goal of the project is to open an effective communication channel between the efforts of the clinical core and the engineering aspects of the overall proposal. Data will be obtained from Core A in a fashion identical to that acquired by the statistical groups. System parameters of the scanning procedure will be required for implementation in the feature set if available, in an attempt to study the possible role of physics in the speckle formation process, particularly with respect to frequency, frequency bandwidth, focusing and angle of incidence in B-mode generation. The feature set of primary importance will include parameters obtained from physician diagnosis, patient data, the statistical projects (#2 Texture Models for Soft Tissue Classification, #3 Application of Diversity Techniques and Scaling Concepts for Ultrasonic Tissue Characterization and Enhanced Tumor Detection, #4 Signal Processing for Enhanced Tissue Imaging and #9 Tissue Scatter Analysis with Spectral Redundancy), system parameters and other imaging modalities. Data fusion will be implemented with data acquired from the above mentioned resources. Appropriate features will be extracted from the entire feature set. The fractures are selected to give maximum differentiation between malignant and non-malignant lesions. The features from the above mentioned projects will be implemented in the pre-classification stage. The feature selection will be carried out in an iterative way by evaluating each feature set with a pattern classifier. The classification of an unknown pattern will be performed by a neural network (NN) structure (preferable error-backpropagation algorithm) which is a state-of-the-art technique. The technique of NN is as good as the traditional techniques of minimum distance classifiers, correlation- and multivariate-statistical classifiers, etc. This Tissue Diagnostic System (TDS) is a dynamic system with self-upgrading capabilities. Investigations will be initiated using sample problems originating from computed B-mode speckle patterns developed by our group during the previous NIH grant period. Future investigations will proceed based on Core A database and the developments of the above mentioned four projects. The verification phase of the project will be based on ROC analysis.
| Wheatley, Margaret A; Forsberg, Flemming; Oum, Kelleny et al. (2006) Comparison of in vitro and in vivo acoustic response of a novel 50:50 PLGA contrast agent. Ultrasonics 44:360-7 |
| Wheatley, Margaret A; Forsberg, Flemming; Dube, Neal et al. (2006) Surfactant-stabilized contrast agent on the nanoscale for diagnostic ultrasound imaging. Ultrasound Med Biol 32:83-93 |
| Mogatadakala, Kishore V; Donohue, Kevin D; Piccoli, Catherine W et al. (2006) Detection of breast lesion regions in ultrasound images using wavelets and order statistics. Med Phys 33:840-9 |
| Shankar, P M; Piccoli, C W; Reid, J M et al. (2005) Application of the compound probability density function for characterization of breast masses in ultrasound B scans. Phys Med Biol 50:2241-8 |
| Forsberg, Flemming; Lathia, Justin D; Merton, Daniel A et al. (2004) Effect of shell type on the in vivo backscatter from polymer-encapsulated microbubbles. Ultrasound Med Biol 30:1281-7 |
| El-Sherif, Dalia M; Lathia, Justin D; Le, Ngocyen T et al. (2004) Ultrasound degradation of novel polymer contrast agents. J Biomed Mater Res A 68:71-8 |
| Alacam, Burak; Yazici, Birsen; Bilgutay, Nihat et al. (2004) Breast tissue characterization using FARMA modeling of ultrasonic RF echo. Ultrasound Med Biol 30:1397-407 |
| Lathia, Justin D; Leodore, Lauren; Wheatley, Margaret A (2004) Polymeric contrast agent with targeting potential. Ultrasonics 42:763-8 |
| Oeffinger, Brian E; Wheatley, Margaret A (2004) Development and characterization of a nano-scale contrast agent. Ultrasonics 42:343-7 |
| Shankar, P Mohana; Dumane, Vishruta A; Piccoli, Catherine W et al. (2003) Computer-aided classification of breast masses in ultrasonic B-scans using a multiparameter approach. IEEE Trans Ultrason Ferroelectr Freq Control 50:1002-9 |
Showing the most recent 10 out of 44 publications
