Ultrasonic Tissue Microstructure Characterization Using Two-Dimensional Spectral Analysis and Its Potential Role in Conformal Radiation Therapy
T Liu,F.L Lizzi,R.H Silverman,G.J Kutcher
DOI: https://doi.org/10.1016/s0360-3016(02)03112-7
2002-01-01
Abstract:Purpose/Objective: The objective of this research is to develop an ultrasonic technique to characterize physical properties of different tissue microstructures. The long-term goal is to use this technique to localize cancerous regions (e.g. in prostate) for radiation treatment planning as well as to monitor tumor response and normal tissue damage of radiation therapy. Materials/Methods: One-dimensional (1-D) ultrasonic spectrum analysis has been found useful in clinical tissue characterization (Feleppa 1997). We have expanded the 1-D spectrum technique to 2-D to obtain 3-D distribution of microstructures using a sequence of 2-D spectrum analyses. This technique utilizes radio-frequency (RF) signals measured directly at the conventional medical ultrasound probe (transducer). 2-D power spectra are computed by performing 2-D Fourier transforms of the echo signals with respect to range (depth) and scan directions. Several spectral parameters are evaluated as quantitative assessments of tissue microstructures. These parameters are: 1. peak value of radially integrated spectral power (RISP), 2. peak shift of RISP; 3. 3-dB width of the RISP; 4. slope of angularly integrated spectral power (AISP), and 5. intercept of AISP. A theoretical model is developed to relate these measured spectral parameters to physical properties of tissue microstructures such as size, shape, orientation and acoustic impedance. Our 2-D ultrasound method has been applied to various experimental and clinical data including tissue-mimicking phantom, beef muscle fiber and human ocular tumor data. The phantom and ocular study employed a low frequency (10-MHz) transducer with 100 micron resolution. For in vivo beef muscle experiment, a high frequency (50-MHz) transducer was used to achieve a resolution of 20 micron. Results: The phantom study validated our theoretical model. In our beef muscle study, two independent muscle fiber groups with different fiber length, thickness and orientation were analyzes with our 2-D spectrum technique. Significant differences were shown in spectral parameters (p<0.0001). Muscle group one has a peak value of RISP of 20.7 dB, central shift of 8.80, 3-dB width of 7.50 and intercept of AISP of 41.4 dB. Respective values for group two were 28.7 dB, 3.30,10.30 and 62.9dB. We found that ultrasonic 2-D spectral parameters can provide a basis for estimating physical properties of muscle fiber cells such as size, shape, and orientation. In vivo results from 10 posterior intraocular tumors demonstrate that our 2-D technique can be adapted to meet specific clinical needs for specific organs. In these 10 patients with subsequent histology, we found our 2-D spectral parameters could significantly differentiate between benign and malignant melanoma by identifying vascular loops (50 65 micron, invisible to conventional Doppler ultrasound). Conclusions: Ultrasonic tissue characterization, using 2-D spectrum analysis, can contribute to the evaluation of the physical properties of tissue microstructures in a broad range of organs (e.g. eye, prostate and liver). This technique appears to be able to differentiate cancerous regions from normal tissue in prostate. We are currently working on incorporating this technique in intensity-modulated radiation therapy (IMRT) for prostate cancer. During the entire span of the treatment period, non-invasive ultrasonic monitoring of tissue properties will also enable the tracking of radiation induced changes of both diseased and normal tissue. Feleppa, E.J, et al., Ultrasonic Spectral-Parameter Imaging of the Prostate, Int. J. Imag. Sys. Tech., 8:11-25, 1997.