Debomita Basu, R. M. Arthur and J. W. Trobaugh, "Role of Collagen in the Temperature Dependence of Ultrasonic Backscatter", Memphis Biomedical Imaging Symposium 2007, Memphis, Tennessee, 1-2 November 2007. 

Purpose.  The combination of hyperthermia with conventional chemotherapy or radiation has proven to be beneficial in the treatment of cancer.  A major limitation for hyperthermia currently is the absence of adequate information about the temperature distribution in soft tissue to guide thermal dosimetry.  We predicted a monotonic change in ultrasound backscattered energy using lipid and aqueous sub-wavelength scatterers [1]. In-vitro experiments with porcine and turkey tissue verified this prediction.  Studies by others have indicated that the echo pattern in ultrasonic images is determined largely by the content, configuration, and distribution of collagen.  To better understand the phenomena that cause changes in backscattered energy (CBE), we have 1) refined our theoretical model to include collagen and 2) examined the relationship between CBE and tissue structure in both simulated and measured images.
Methods.  CBE images were formed from simulated B-mode images [2]. Simulated images were based on random distributions of populations of aqueous, lipid and collagen scatterers in an aqueous medium.  Measured CBE was found from ultrasonic images of turkey breast in a water bath during heating from 37^oC to 50^oC in 0.5^oC steps. Images were obtained with a Terason 2000 system (Teratech Corp., Burlington, MA) using a 7.5 MHz linear probe by placing the transducer both parallel and perpendicular to the turkey muscle fibers.  For both simulated and measured images, CBE images were calculated from the squared envelopes of the B-mode images.  CBE images were compared to their respective B-mode images.

Results.  Correlation coefficients from CBE based on simulated images showed significant correlation between the CBE and the B-mode images.  Simulated images used random placement of scatterers and therefore does not show any specific structure.  For the experimental results, there was a weak correlation (r = 0.2-0.34) for both alignments of the transducer with respect to the fiber direction of the turkey muscle.  Correlation was also found from regions of bright echo (r =0.39 and 0.36 for parallel and perpendicular orientations, respectively), but was strongest for regions with pure speckle texture (r = 0.8 and 0.65, respectively).
Conclusion.  Examination of the statistical correlation between CBE images and their respective B-mode images indicates strong correlation between speckle texture and CBE.  The effect, however, of tissue constituents and structure on CBE is not clear.  Further investigation of this phenomenon is being done by comparing histological studies with CBE images to correlate fiber structure with the change in backscattered energy.
 

Support:  R21-CA90531, R01-CA107558 and the Wilkinson Trust at Washington University, St. Louis.