Abstract:Purpose/Objective: The objective of this research is to demonstrate the feasibility of using an ultrasonic tissue-typing technique to differentiate and potentially grade cancerous tissue inside the prostate. These studies were undertaken with the ultimate goal of localizing cancerous regions inside the prostate to guide tumor dose escalation for both IMRT and brachytherapy. Materials/Methods: Ultrasonic tissue-typing imaging reveals the physical properties of biological tissues noninvasively. Ultrasonic tissue typing using 1-D spectrum analysis has demonstrated the ability of identifying cancer-containing regions in the prostate (Feleppa 1997). Comparing the results from this ultrasonic technique to the results of prostate biopsies showed a high degree of accuracy with the area under the receiver-operating characteristic (ROC) curve of 0.85, a result superior to other known imaging methods. We have expanded the 1-D spectrum technique to 2-D to obtain 3-D distributions of microstructures using a sequence of 2-D spectrum analyses. Our 2-D spectrum technique analyzes radio-frequency ultrasonic echo signals and extracts features of the ultrasonic signals that are associated with properties of tissue microstructures. The 2-D method is capable of accurate assessment of asymmetric tissue features, which is a key factor in grading cancerous tissue, and hence overcomes the primary limitation of the original 1-D analysis. 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 (stiffness). These parameters are: peak value of radially integrated spectral power (RISP); peak shift of RISP; 3-dB width of RISP; slope of angularly integrated spectral power (AISP); and intercept of AISP. Four patients with localized prostate cancer were enrolled in a feasibility study. In vivo 3-D prostate ultrasound radio-frequency (RF) data were acquired in a sequence of parallel scans which were 1 mm apart using a B&K Medical System Hawk with 7.5 MHz probe. The RF data were analyzed using a 2-D Fourier transform algorithm. The resulting spectral parameters were computed to quantitatively assess the prostatic tissue. Prostate needle biopsies were used as the gold standard. Results: From the theoretical model, we have derived closed form solutions that explicitly relate spectral parameters to the physical properties of internal biological tissue. Among these parameters, slope of AISP is sensitive to the size of the tissue microstructures; peak shift of RISP is sensitive to the orientation; peak value of RISP is sensitive to the concentration; 3-dB width of RISP is sensitive to the shape. In the initial study with four prostate cancer patients, we found that the 2-D spectral analysis results were consistent with the theoretical predictions in differentiating between cancerous and non-cancerous tissues, and low-grade and high-grade tumors. For example, decrease in the 3-dB width of RISP for the patient with high Gleason score could be due to the lack of symmetry or the loss of homogeneity of the prostatic tissue, which is indicative of high-grade cancer. Conclusions: Ultrasonic tissue-typing imaging using 2-D spectrum analysis method can contribute to the evaluation of physical properties of prostate tissue microstructures that can become the basis for segmenting cancer-containing regions and grading cancers inside prostate. Initial studies demonstrated the feasibility of clinically employing this ultrasonic imaging method for prostate cancer. As the radiation treatment technique advances, this localized tumor information will become indispensable for intraprostatic tumor dose escalation. We are now incorporating this ultrasonic tissue typing technique in our clinical trials of dose escalation (with IMRT and brachytherapy) for patients with localized prostate cancer.

2-D images for biopsy guidance and 3-D images for treatment planning and monitoring of prostate cancer based upon spectrum analysis and neural-network classification

Spectrum-Analysis and Neural Networks for Imaging to Detect and Treat Prostate Cancer

Three-dimensional Ultrasound Analyses of the Prostate.

3-D Multi-parametric Contrast-Enhanced Ultrasound for the Prediction of Prostate Cancer.

Targeting and monitoring radiation therapy of prostate cancer using ultrasonic spectrum-analysis and neural-network classification for tissue-type imaging

Machine Learning for the Prediction of Prostate Cancer Biopsy Based on 3D Dynamic Contrast-Enhanced Ultrasound Quantification

Can 3D Multiparametric Ultrasound Imaging Predict Prostate Biopsy Outcome?

Retrospective T2 Quantification from Conventional Weighted MRI of the Prostate Based on Deep Learning

3-D tissue typing of prostate tissue based on spectral parameters

PREDICTION OF PROSTATE BIOPSY OUTCOME BY MULTIPARAMETRIC ULTRASOUND IMAGING

Classification of in Vivo Autofluorescence Spectra Using Support Vector Machines

Multiparametric Ultrasound and Machine Learning for Prostate Cancer Localization

Improving tissue‐type imaging of prostate cancer by combining magnetic‐resonance spectroscopy and ultrasound spectral parameters

Ultrasonic tissue-typing imaging for guiding dose escalation of prostate cancer radiotherapy

Multiparametric Ultrasound Imaging of Prostate Cancer Using Deep Neural Networks

ProsDectNet: Bridging the Gap in Prostate Cancer Detection via Transrectal B-mode Ultrasound Imaging

Prostate Cancer Identification: Quantitative Analysis of T2-weighted MR Images Based on a Back Propagation Artificial Neural Network Model.

Detection and grading of prostate cancer using temporal enhanced ultrasound: combining deep neural networks and tissue mimicking simulations

Ultrasonic Tissue Microstructure Characterization Using Two-Dimensional Spectral Analysis and Its Potential Role in Conformal Radiation Therapy

Using Spatio-Temporal Dual-Stream Network with Self-Supervised Learning for Lung Tumor Classification on Radial Probe Endobronchial Ultrasound Video

Application of Spectral Algorithm Applied to Spatially Registered Bi-Parametric MRI to Predict Prostate Tumor Aggressiveness: A Pilot Study