Tissue-type Imaging Research Articles

Tissue-Type Classification With Uncertainty Quantification of Microwave and Ultrasound Breast Imaging: A Deep Learning Approach

A deep learning approach is proposed for performing tissue-type classification of tomographic microwave and ultrasound property images of the breast. The approach is based on a convolutional neural network (CNN) utilizing the U-net architecture that also quantifies the uncertainty in the classification of each pixel. Quantitative tomographic reconstructions of dielectric properties (complex-valued permittivity), ultrasonic properties (compressibility and attenuation), as well as their combination, with the corresponding actual tissue-type classification constitute the training set. The CNN learns to map the quantitative property reconstructions to a single tissue-type image. The level of confidence in predicting a tissue-type at each pixel is determined. This uncertainty quantification is diagnostically critical for biomedical applications, especially when attempting to distinguish between cancerous and healthy tissues. The Gauss-Newton Inversion algorithm is used for the quantitative reconstruction of both dielectric and ultrasonic properties. Electromagnetic and ultrasound scattered-field data is obtained from MRI-derived numerical breast phantoms. Several numerical breast phantoms types, from fatty to dense, are considered. The proposed classification and uncertainty quantification approach is shown to outperform a previously studied tissue-type classification method based on a Bayesian approach.

Experimental Evaluation of Composite Tissue-Type Ultrasound and Microwave Imaging

The recently proposed concept of composite tissue-type image (cTTI), which provides an easy-to-interpret image constructed from quantitative ultrasonic and electromagnetic properties, is experimentally investigated. The experimental data set used for the ultrasound investigation is obtained from the multimodal ultrasound breast imaging system. The experimental data set utilized for microwave imaging is provided by an in-house system at the University of Manitoba. To this end, a tissue mimicking phantom and a human forearm are utilized for experimental ultrasound and microwave imaging. In addition, the cTTI algorithm is modified to take into account differences in the quantitative accuracy of reconstructing one property compared to other properties so as to increase the achievable accuracy in the resulting cTTI. In addition to experimental ultrasound and microwave data, the cTTI method is also applied against synthetic data obtained from an MRI-based numerical breast phantom to further demonstrate the performance of the cTTI not only with respect to microwave and ultrasound tomography data but also with respect to their combination. Finally, the improvements of the cTTI reconstructions based on changing the prior probabilities compared with equal prior probability distribution are shown for combined ultrasound and microwave tomography of a numerical breast phantom.

Automated characterisation of lipid core plaques in vivo by quantitative optical coherence tomography tissue type imaging.

Qualitative criteria for plaque tissue characterisation by OCT are well established, but quantitative methods lack systematic validation in vivo. High optical attenuation coefficient µt has been associated with unstable plaque features, such as lipid core. The purpose of this study was to validate optical coherence tomography (OCT) attenuation imaging for tissue characterisation in vivo, specifically to detect lipid core in coronary atherosclerotic plaques, and to evaluate quantitatively the ability of OCT attenuation imaging to differentiate thin-cap (TCFA) and thick-cap fibroatheroma (FA). We prospectively enrolled 85 patients undergoing imaging of a native coronary segment by both OCT and near-infrared spectroscopy and intravascular ultrasound (NIRS-IVUS). Ninety-eight NIRS-positive 4 mm plaque segments were selected and matched to the OCT data. Two experienced OCT readers classified the plaque type using OCT criteria. A cap thickness of 65 μm was used to differentiate TCFA and FA. We computed an index of plaque attenuation (IPA) in the 4 mm blocks, and assessed the association of this index with plaque type. IPA differentiated between TCFA and FA (AUC=0.82 in ROC analysis; p<0.0001). LCBI was numerically, but not significantly, higher in TCFA compared to FA (p=0.097). IPA is an unbiased reproducible measure of tissue optical properties. The fraction of pixels with attenuation coefficient ≥11 mm-1 can identify TCFA.

SU‐F‐I‐14: 3D Breast Digital Phantom for XACT Imaging

Purpose:The X‐ray induced acoustic computed tomography (XACT) is a new imaging modality which combines X‐ray contrast and high ultrasonic resolution in a single modality. Using XACT in breast imaging, a 3D breast volume can be imaged by only one pulsed X‐ray radiation, which could dramatically reduce the imaging dose for patients undergoing breast cancer screening and diagnosis. A 3D digital phantom that contains both X‐ray properties and acoustic properties of different tissue types is indeed needed for developing and optimizing the XACT system. The purpose of this study is to offer a realistic breast digital phantom as a valuable tool for improving breast XACT imaging techniques and potentially leading to better diagnostic outcomes.Methods:A series of breast CT images along the coronal plane from a patient who has breast calcifications are used as the source images. A HU value based segmentation algorithm is employed to identify breast tissues in five categories, namely the skin tissue, fat tissue, glandular tissue, chest bone and calcifications. For each pixel, the dose related parameters, such as material components and density, and acoustic related parameters, such as frequency‐dependent acoustic attenuation coefficient and bandwidth, are assigned based on tissue types. Meanwhile, other parameters which are used in sound propagation, including the sound speed, thermal expansion coefficient, and heat capacity are also assigned to each tissue.Results:A series of 2D tissue type image is acquired first and the 3D digital breast phantom is obtained by using commercial 3D reconstruction software. When giving specific settings including dose depositions and ultrasound center frequency, the X‐ray induced initial pressure rise can be calculated accordingly.Conclusion:The proposed 3D breast digital phantom represents a realistic breast anatomic structure and provides a valuable tool for developing and evaluating the system performance for XACT.

Composite Tissue-Type and Probability Image for Ultrasound and Microwave Tomography

The concept of creating a composite tissue-type-image (cTTI) along with an associated probability image is introduced for ultrasound and microwave tomography. The cTTI integrates information available within different quantitative property images, and the associated probability image provides an indication of the level of confidence regarding the reconstructed tissue types. It is shown that the cTTI concept can be applied to ultrasound tomography property images, microwave tomography property images, as well as to their combination. Thus, the concept is generalizable to the amalgamation of quantitative information derived from a wide variety of modalities with the goal of increasing the confidence in the reconstructed cTTI. Validation of the concept is performed on MRI-derived numerical breast phantoms containing up to five different tissue types.

Prospective phase I/II trial of dose-painting prostate brachytherapy guided by cancer-specific ultrasound-based tissue-type imaging (TTI).

246 Background: Standard treatments treat the entire prostate gland uniformly, while focal treatments target only the dominant tumor. We propose an alternative approach wherein tumor areas are treated with an escalated dose of radiation while the remainder of the prostate with dose de-escalation. Tissue-type images (TTIs), which identify intraprostatic tumors via ultrasound spectral analysis, were used to identify tumors. Methods: 14 low risk patients were enrolled. TTIs and standard B-mode ultrasound images were generated and fused intraoperatively. Based on TTIs, plans were developed to deliver 200% prescription dose to the tumors and 100% to the prostate. Doses above 100% to non-tumor regions were minimized and standard normal tissue constraints were respected. Results: 12 patients were successfully treated. Technical problems occurred in the first patient and another received a standard implant due to discordance of TTIs with MRI and biopsy findings. One patient passed away soon after the procedure of unrelated cause. A total of 28 tumors were identified in these 12 patients (median 2 tumors/patient, range 1-4). Mean tumor volume was 0.72cc (range 0.03cc -4.07cc). Tumor dose was significantly higher in TTI-based plans, with mean tumor V200%: TTI vs. standard 97% vs. 48%, p&lt;0.05. Modest dose reduction of the whole prostate was also achieved. Median follow-up was 25.7 months (range 21.3-48 months). A low incidence of grade 2 toxicities and no grade 3-4 toxicities were seen (Table). Two patients experienced a biochemical failure per the nadir+2 definition, both failures are attributable to PSA bounce (PSA down to 1.0 and 0.59 at 42 and 30 mos., respectively). Post-treatment biopsies have been performed on 5/11 patients (4 refused), with 2 patients with pending biopsies. 4/5 available biopsies are negative, with 1 patient having a small focus of residual cancer. Conclusions: Dose-painting prostate brachytherapy is technically feasible. This tumor-directed approach, more rational than uniform whole gland treatments, offers the promise of excellent cancer control, low toxicity and low risk of re-treatment. Clinical trial information: NCT01227642. [Table: see text]

Quantitative Optical Coherence Tomography Tissue-Type Imaging for Lipid-Core Plaque Detection

Tissue type in the coronary artery wall is an important parameter with significant clinical implications in interventional cardiology. Lipid-rich necrotic lesions may cause periprocedural complications, stent failure, and acute coronary events. Analysis of the optical properties of a lesion imaged

Advanced spectral analyses for real-time automatic echographic tissue-typing of simulated tumor masses at different compression stages

Prototypal software algorithms for advanced spectral analysis of echographic images were developed to perform automatic detection of simulated tumor masses at two different pathological stages. Previously published works documented the possibility of characterizing macroscopic variation of mechanical properties of tissues through elastographic techniques, using different imaging modalities, including ultrasound (US); however, the accuracy of US-based elastography remains affected by the variable manual modality of the applied compression and several attempts are under investigation to overcome this limitation. Quantitative US (QUS), such as Fourier- and wavelet-based analyses of the RF signal associated with the US images, has been developed to perform a microscopic-scale tissue-type imaging offering new solutions for operator-independent examinations. Because materials able to reproduce the harmonic behavior of human liver can be realized, in this study, tissue-mimicking structures were US imaged and the related RF signals were analyzed using wavelet transform through an in-house-developed algorithm for tissue characterization. The classification performance and reliability of the procedure were evaluated on two different tumor stiffnesses (40 and 130 kPa) and with two different applied compression levels (0 and 3.5 N). Our results demonstrated that spectral components associated with different levels of tissue stiffness within the medium exist and can be mapped onto the original US images independently of the applied compressive forces. This wavelet-based analysis was able to identify different tissue stiffness with satisfactory average sensitivity and specificity: respectively, 72.01% ± 1.70% and 81.28% ± 2.02%.

Sonoelastography: the method of choice for evaluation of tissue elasticity

Ultrasound devices and methodologies have been continuously evolving and becoming more important as tools in diagnostic medicine. Recently a new ultrasound diagnostic technique has been developed. Named sonoelastography, the technique enables evaluation of tissue elasticity and is based on differences in stiffness (hardness, compressibility, elasticity) of pathological changes and normal adjacent tissue. Sonoelastography (SE) is also known as Real-time tissue elastography (RTE), Real-time sonoelastography (RTSE), Tissue type imaging (TTI) and Ultrasound Strain Imaging Technology. It has been found useful in many medicine elds and adopted readily by clinicians of different specialties. It gives more information than conventional ultrasound in evaluation of tumors, liver disease, skeletal muscles, rheumatoid nodules and other pathological changes. This review covers the basics of elastography, its applications, instruments, techniques, the scoring system and the indications for elastography.

Ultrasonic tissue-type imaging of the prostate: Implications for biopsy and treatment guidance

Improved means of imaging prostate cancer would enable more-effective biopsy and treatment guidance and potentially would provide a reliable means of monitoring non-surgical therapy. Current, commonly used, conventional means of imaging the prostate do not reliably depict cancerous lesions, and as a result, biopsy needles are placed with respect to visible anatomic features of the gland, treatment tends to involve the entire gland, and monitoring of non-surgical therapy is based predominantly on blood PSA levels, and in many cases, periodic biopsies. Conventional transrectal ultrasound often is the imaging modality of choice for prostate biopsy and treatment procedures, but it offers no advantages in terms of reliably depicting cancerous regions of the gland. However, new methods of tissue-type imaging that are based on spectrum analysis of echo signals and that utilize artificial neural networks for classification offer promise of reliably distinguishing cancerous lesions from non-cancerous tissue in the prostate. The classifier produced an ROC-curve area of 0.84 for 617 biopsied locations compared to areas of 0.64 for conventional assessments of the same locations in biopsy-guidance B-mode images. The potential improvement in imaging sensitivity implied by the ROC curves is more than 50%. If current validation studies confirm these initial results, then an effective, inexpensive, noninvasive means of imaging prostate-cancer foci and therefore of guiding biopsies and treatments will be available to urologists and radiation oncologists.

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

This study aims to improve existing prostate tissue-type imaging methods by utilizing independent tissue properties sensed by MR and US. An artificial-neural-network classifier was trained using the blood PSA level and ultrasonic spectral parameters of echo signals derived from biopsied regions of 64 patients. Biopsy-core histology was used as the gold standard. Classifier performance was assessed using ROC analysis. We generated tissue-type images (TTIs) using a look-up table that returned cancer-likelihood scores for spectral parameter and PSA combinations. We assessed the feasibility of integrating MR and US parameters to improve classification. 3-D renderings of the prostate from transverse MR and US scans were generated from data acquired preoperatively. The 3-D rendering obtained from MR data showed that the prostate was considerably distorted (flattened) by the large MR endorectal probe compared to the distortion caused by the smaller US endorectal probe. The MR 3-D rendering was warped successfully to match the US volume. Warping successfully compensated for the deformations introduced during scanning. This suggests that accurate coregistration of 3-D data acquired using MR and US is possible. Such coregistration is essential for developing an improved classifier based on MR and US parameters.

283: Feasibility of Combining Ultrasonic and Magnetic-Resonance Methods for Improved Tissue-Type Imaging of Prostate Cancer

283: Feasibility of Combining Ultrasonic and Magnetic-Resonance Methods for Improved Tissue-Type Imaging of Prostate Cancer

166 Ultrasonic tissue-typing imaging in detecting and evaluating prostate cancer: In vivo and ex vivo studies vs. wholemount pathology

166 Ultrasonic tissue-typing imaging in detecting and evaluating prostate cancer: In vivo and ex vivo studies vs. wholemount pathology

Tissue-type imaging (TTI) based on ultrasonic spectral and clinical parameters for detecting, evaluating, and managing prostate cancer

This study seeks to develop more-sensitive and -specific ultrasonic methods of imaging cancerous prostate tissue and thereby to improve means of guiding biopsies and planning, targeting, and monitoring treatment. Ultrasonic radio-frequency, echo-signal data, and clinical variables, e.g., PSA, voiding function, etc., during biopsy examinations were acquired. Spectra of the radio-frequency signals were computed in each biopsied region, and used to train neural networks; biopsy results served as the gold standard. A lookup table gave scores for cancer likelihood on a pixel-by-pixel basis from locally computed spectral-parameter and global clinical-parameter values. ROC curves used leave-one-patient- and leave-one-biopsy-out approaches to minimize classification bias. Resulting ROC-curve areas were 0.80±0.03 for neural-networks versus 0.66±0.03 for conventional classification. TTIs generated from data acquired pre-surgically showed tumors that were unrecognized in conventional images and during surgery. 3-D renderings of prostatectomy histology and TTIs showed encouraging correlations, which shows promise for improving the detection and management of prostate cancer, e.g., for biopsy guidance, planning dose-escalation and tissue-sparing options for radiation or cryotherapy, and assessing the effects of treatment. Combining MRS parameters with US spectral parameters appears capable of further improving prostate-cancer imaging. [Work supported by NIH.]

642 poster Ultrasonic tissue-typing imaging for image-guided prostate cancer radiotherapy

642 poster Ultrasonic tissue-typing imaging for image-guided prostate cancer radiotherapy

In vivo and ex vivo prostate cancer imaging using an ultrasonic tissue-typing technique for imaging-guided dose escalation of prostate cancer radiotherapy

in vivo and ex vivo prostate cancer imaging using an ultrasonic tissue-typing technique for imaging-guided dose escalation of prostate cancer radiotherapy

In vivo and ex vivo prostate cancer imaging using an ultrasonic tissue-typing technique for imaging-guided dose escalation of prostate cancer radiotherapy

Materials/Methods: Ultrasonic tissue-typing imaging using spectrum analysis technique reveals the physical properties (e.g., size, shape and stiffness) of biological tissues noninvasively. Two studies were carried out, after IRB approval, to determine how well this ultrasound technique can differentiate cancerous tissues inside the prostate. Ten patients with localized prostate cancer were enrolled in the in vivo study. Each patient was scanned with a conventional transrectal ultrasound (B & K Medical System, Hawk, 7.5 MHz probe) before the brachytherapy procedure. 3-D ultrasound radio-frequency (rf) data were acquired in a sequence of parallel axial scans which were 1 mm apart. Prior prostate needle biopsies were used as the gold standard. For the ex vivo study, prostate specimens were immersed in saline and scanned with the same ultrasound system immediately after radical prostatectomy. Entire surgical specimens were then embedded for histologic examination. The ultrasound rf data were analyzed using a 2-D Fourier transform algorithm. The resulting spectral parameters of both cancerous regions and noncancerous regions were computed to quantitatively assess the prostatic tissue.

Recent Developments in Tissue-Type Imaging (TTI) for Planning and Monitoring Treatment of Prostate Cancer

Because current methods of imaging prostate cancer are inadequate, biopsies cannot be effectively guided and treatment cannot be effectively planned and targeted. Therefore, our research is aimed at ultrasonically characterizing cancerous prostate tissue so that we can image it more effectively and thereby provide improved means of detecting, treating and monitoring prostate cancer. We base our characterization methods on spectrum analysis of radiofrequency (rf) echo signals combined with clinical variables such as prostate-specific antigen (PSA). Tissue typing using these parameters is performed by artificial neural networks. We employed and evaluated different approaches to data partitioning into training, validation, and test sets and different neural network configuration options. In this manner, we sought to determine what neural network configuration is optimal for these data and also to assess possible bias that might exist due to correlations among different data entries among the data for a given patient. The classification efficacy of each neural network configuration and data-partitioning method was measured using relative-operating-characteristic (ROC) methods. Neural network classification based on spectral parameters combined with clinical data generally produced ROC-curve areas of 0.80 compared to curve areas of 0.64 for conventional transrectal ultrasound imaging combined with clinical data. We then used the optimal neural network configuration to generate lookup tables that translate local spectral parameter values and global clinical-variable values into pixel values in tissue-type images (TTIs). TTIs continue to show cancerous regions successfully, and may prove to be particularly useful clinically in combination with other ultrasonic and nonultrasonic methods, e.g., magnetic-resonance spectroscopy.

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

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

Recent developments in tissue-type imaging (TTI) for planning treatment of prostate cancer

Objective: The aim of this study was to develop ultrasonic methods for distinguishing cancerous from noncancerous prostate tissue or changes in those tissues so that therapy can be more effectively targeted and treatment more effectively monitored.