The objective of this research is to improve detection sensitivity and specificity in diagnostic imaging of the breast. This will be accomplished by exploiting a novel hybrid imaging method, which combines near infrared (NIR) optical techniques based on diffusive waves with ultrasound. NIR imaging affords high sensitivity/specificity in the detection of breast tumors, However, it is intrinsically limited in spatial resolution because of the diffuse nature of NIR light in human tissue. In addition, NIR imaging reconstruction is in general underdetermined and ill posed. On the other hand, ultrasound has good imaging capability and can resolve a few mm lesions in breast tissues. However, ultrasound sensitivity and specificity for breast cancer detection are not high due to the overlap in acoustic features of normal tissue, benign and malignant lesions. Our hybrid imaging system is designed to combine the high contrasts of NIR imaging in distinguishing among normal, benign and malignant tissues with the high spatial resolution inherent in ultrasound imaging. Specifically, optical image reconstruction algorithms will be optimized by using ancillary knowledge about lesion geometry provided by co-registered ultrasound images. A cohort of patients with benign and malignant lesions, as determined by physical exam and radiologic study, will be examined, and the data thus obtained will be used to assess the hybrid NIR/ultrasound method and to compare it with conventional ultrasound diagnosis. The detection sensitivity and specificity of the combined method compared to ultrasound alone will be estimated using biopsy results as """"""""gold' standard. Completion of this research plan will serve to validate a novel method for detection and diagnosis of breast lesions, resulting in new opportunities for early detection and appropriate treatment of breast carcinoma.
Showing the most recent 10 out of 59 publications
