Tissue classification with generalized spectrum parameters

Correlation length ratio as a parameter for determination of fiber-like structures in soft tissues

Quantitative ultrasound methods can provide valuable information about the microstructure of a material or tissue. This works well when the common assumptions of homogeneity, isotropy, and diffuse scattering conditions are valid. In biological tissues, however, these assumptions are often violated because the microstructure of biological tissues is often heterogeneous and anisotropic. The microstructure of biological tissues can change with disease, and therefore accurate identification and description of a tissue's microstructure can offer important clinical insight. To address the challenge of evaluating the microstructure of biological tissues, here we introduce a novel parameter called the correlation length ratio (CLR), a ratio of lateral to axial correlation lengths for backscattered echo signals. We developed it to determine the presence of fiber-like structures in soft tissues by comparing this value in tissue to a threshold determined from a reference material that is homogeneous, isotropic, and provides diffuse scattering. We tested this novel parameter in phantoms with spherical scattering sources, in an anisotropic phantom (containing elongated fibers), and in human biceps muscle. We found that the CLR accurately detected the presence of elongated structures in both the anisotropic phantom and muscle. These results encourage further exploration of this novel parameter in microstructurally complex tissues.

A Quantitative Ultrasound-Based Multi-Parameter Classifier for Breast Masses

This manuscript reports preliminary results obtained by combining estimates of two or three (among seven) quantitative ultrasound (QUS) parameters in a model-free, multi-parameter classifier to differentiate breast carcinomas from fibroadenomas (the most common benign solid tumor). Forty-three patients scheduled for core biopsy of a suspicious breast mass were recruited. Radiofrequency echo signal data were acquired using clinical breast ultrasound systems equipped with linear array transducers. The reference phantom method was used to obtain system-independent estimates of the specific attenuation (ATT), the average backscatter coefficients, the effective scatterer diameter (ESD) and an effective scatterer diameter heterogeneity index (ESDHI) over regions of interest within each mass. In addition, the envelope amplitude signal-to-noise ratio (SNR), the Nakagami shape parameter, m, and the maximum collapsed average (maxCA) of the generalized spectrum were also computed. Classification was performed using the minimum Mahalanobis distance to the centroids of the training classes and tested against biopsy results. Classification performance was evaluated with the area under the receiver operating characteristic (ROC) curve. The best performance with a two-parameter classifier used the ESD and ESDHI and resulted in an area under the ROC curve of 0.98 (95% confidence interval [CI]: 0.95-1.00). Classification performance improved with three parameters (ATT, ESD and ESDHI) yielding an area under the ROC curve of 0.999 (0.995-1.000). These results suggest that system-independent QUS parameters, when combined in a model-free classifier, are a promising tool to characterize breast tumors. A larger study is needed to further test this idea.

EEMD Domain AR Spectral Method for Mean Scatterer Spacing Estimation of Breast Tumors From Ultrasound Backscattered RF Data

We present a novel method for estimating the mean scatterer spacing (MSS) of breast tumors using ensemble empirical mode decomposition (EEMD) domain analysis of deconvolved backscattered radio frequency (RF) data. The autoregressive (AR) spectrum from which the MSS is estimated is obtained from the intrinsic mode functions (IMFs) due to regular scatterers embedded in RF data corrupted by the diffuse scatterers. The IMFs are chosen by giving priority to the presence of an enhanced fundamental harmonic and the presence of a greater number of higher harmonics in the AR spectrum estimated from the IMFs. The AR model order is chosen by minimizing the mean absolute percentage error (MAPE) criterion. In order to ensure that the backscattered data is indeed from a source of coherent scattering, we begin by performing a non-parametric Kolmogorov-Smirnov (K-S) classification test on the backscattered RF data. Deconvolution of the backscattered RF data, which have been classified by the K-S test as sources of significant coherent scattering, is done to reduce the system effect. EEMD domain analysis is then performed on the deconvolved data. The proposed method is able to recover the harmonics associated with the regular scatterers and overcomes many problems encountered while estimating the MSS from the AR spectrum of raw RF data. Using our technique, a mean absolute percentage error (MAPE) of 5.78% is obtained while estimating the MSS from simulated data, which is lower than that of the existing techniques. Our proposed method is shown to outperform the state of the art techniques, namely, singular spectrum analysis, generalized spectrum (GS), spectral autocorrelation (SAC), and modified SAC for different simulation conditions. The MSS for in vivo normal breast tissue is found to be 0.69 ± 0.04 mm; for benign and malignant tumors it is found to be 0.73 ± 0.03 and 0.79 ± 0.04 mm, respectively. The separation between the MSS values of normal and benign tissues for our proposed method is similar to the separations obtained for the conventional methods, but the separation between the MSS values for benign and malignant tissues for our proposed method is slightly higher than that for the conventional methods. When the MSS is used to classify breast tumors into benign and malignant, for a threshold-based classifier, the increase in specificity, accuracy, and area under curve are 18%, 19%, and 22%, respectively, and that for statistical classifiers are 9%, 13%, and 19%, respectively, from that of the next best existing technique.

A Review of Ultrasound Tissue Characterization with Mean Scatterer Spacing

Tissues exhibiting quasi-periodic structures can be modeled as a collection of diffuse scatterers and coherent scatterers. The mean scatterer spacing (MSS) of coherent and quasi-periodic components is directly related to tissue microstructure and has become an important quantitative ultrasound (QUS) parameter in the characterization of quasi-periodic tissues. In this paper, a review of the literature on the development of MSS as a QUS parameter was conducted. First, a unified theoretical background of MSS estimates was provided. Then, the application of MSS estimates was summarized with respect to liver, spleen, breast, bone, muscle, and other tissues. MSS estimation techniques were applied to (a) the diagnosis of hepatitis, liver fibrosis and cirrhosis, and lesions in tissues such as liver, breast, and spleen; (b) the differentiation between benign and malignant breast tumors, and the grading of breast cancer; (c) the detection of cancellous bone; and (d) the monitoring of the efficacy of treatments such as thermal ablation, with various levels of success. Future developments were also discussed in terms of real-time implementation of MSS estimates, local MSS estimation, relationship of MSS to other QUS parameters, combination of MSS with other QUS parameters, in vivo validation of MSS estimates, MSS parametric imaging, and three-dimensional ultrasound tissue characterization.

Analysis of Coherent and Diffuse Scattering Using a Reference Phantom

The estimation of many spectral-based quantitative ultrasound parameters assumes that backscattered echo signals are from a stationary, incoherent scattering process. The accuracy of these assumptions in real tissue can limit the diagnostic value of these parameters and the physical insight about tissue microstructure they can convey. This work presents an empirical decision test to determine the presence of significant coherent contributions to echo signals and whether they are caused by low scatterer number densities or the presence of specular reflectors or scatterers with periodic spacing. This is achieved by computing parameters from echo signals that quantify stationary or nonstationary features related to coherent scattering, and then comparing their values to thresholds determined from a reference material providing diffuse scattering. The paper first presents a number of parameters with demonstrated sensitivity to coherent scattering and describes criteria to select those with the highest sensitivity using simulated and phantom-based echo data. Results showed that the echo amplitude signal-to-noise ratio and the multitaper-generalized spectrum were the parameters with the highest sensitivity to coherent scattering with stationary and nonstationary features, respectively. These parameters were incorporated into the reference-based decision test, which successfully identified regions in simulated and tissue-mimicking phantoms with different incoherent and coherent scattering conditions. When scatterers with periodic organization were detected, the combination of stationary and nonstationary analysis permitted the estimation of the mean spacing below and above the resolution limit imposed by the pulse size. Preliminary applications of this algorithm to human cervical tissue ex vivo showed correspondence between regions of B-mode images showing bright reflectors, tissue interfaces, and hypoechoic regions with regions classified as specular reflectors and low scatterer number density. These results encourage further application of the algorithm to more structurally complex phantoms and tissue.

Multidimensional spectral analysis of the ultrasonic radiofrequency signal for characterization of media

The importance of the analysis of the radiofrequency signal is by now recognized in the field of tissue characterization via ultrasound. The RF signal contains a wealth of information and structural details that are usually lost in the B-Mode representation. The HyperSPACE (Hyper SPectral Analysis for Characterization in Echography) algorithm presented by the authors in previous papers for clinical applications is based on the radiofrequency ultrasonic signal. The present work describes the method in detail and evaluates its performance in a repeatable and standardized manner, by using two test objects: a commercial test object that simulates the human parenchyma, and a laboratory-made test object consisting of human blood at different dilution values. In particular, the sensitivity and specificity in discriminating different density levels were estimated. In addition, the robustness of the algorithm with respect to the signal-to-noise ratio was also evaluated.

Automatic Cataract Hardness Classification Ex Vivo by Ultrasound Techniques

To demonstrate the feasibility of a new methodology for cataract hardness characterization and automatic classification using ultrasound techniques, different cataract degrees were induced in 210 porcine lenses. A 25-MHz ultrasound transducer was used to obtain acoustical parameters (velocity and attenuation) and backscattering signals. B-Scan and parametric Nakagami images were constructed. Ninety-seven parameters were extracted and subjected to a Principal Component Analysis. Bayes, K-Nearest-Neighbours, Fisher Linear Discriminant and Support Vector Machine (SVM) classifiers were used to automatically classify the different cataract severities. Statistically significant increases with cataract formation were found for velocity, attenuation, mean brightness intensity of the B-Scan images and mean Nakagami m parameter (p < 0.01). The four classifiers showed a good performance for healthy versus cataractous lenses (F-measure ≥ 92.68%), while for initial versus severe cataracts the SVM classifier showed the higher performance (90.62%). The results showed that ultrasound techniques can be used for non-invasive cataract hardness characterization and automatic classification.

Computer Aided Theragnosis Using Quantitative Ultrasound Spectroscopy and Maximum Mean Discrepancy in Locally Advanced Breast Cancer

A noninvasive computer-aided-theragnosis (CAT) system was developed for the early therapeutic cancer response assessment in patients with locally advanced breast cancer (LABC) treated with neoadjuvant chemotherapy. The proposed CAT system was based on multi-parametric quantitative ultrasound (QUS) spectroscopic methods in conjunction with advanced machine learning techniques. Specifically, a kernel-based metric named maximum mean discrepancy (MMD), a technique for learning from imbalanced data based on random undersampling, and supervised learning were investigated with response-monitoring data from LABC patients. The CAT system was tested on 56 patients using statistical significance tests and leave-one-subject-out classification techniques. Textural features using state-of-the-art local binary patterns (LBP), and gray-scale intensity features were extracted from the spectral parametric maps in the proposed CAT system. The system indicated significant differences in changes between the responding and non-responding patient populations as well as high accuracy, sensitivity, and specificity in discriminating between the two patient groups early after the start of treatment, i.e., on weeks 1 and 4 of several months of treatment. The proposed CAT system achieved an accuracy of 85%, 87%, and 90% on weeks 1, 4 and 8, respectively. The sensitivity and specificity of developed CAT system for the same times was 85%, 95%, 90% and 85%, 85%, 91%, respectively. The proposed CAT system thus establishes a noninvasive framework for monitoring cancer treatment response in tumors using clinical ultrasound imaging in conjunction with machine learning techniques. Such a framework can potentially facilitate the detection of refractory responses in patients to treatment early on during a course of therapy to enable possibly switching to more efficacious treatments.

Computer-Aided Prostate Cancer Detection Using Ultrasound RF Time Series: In Vivo Feasibility Study

This paper presents the results of a computer-aided intervention solution to demonstrate the application of RF time series for characterization of prostate cancer, in vivo.

METHODS

We pre-process RF time series features extracted from 14 patients using hierarchical clustering to remove possible outliers. Then, we demonstrate that the mean central frequency and wavelet features extracted from a group of patients can be used to build a nonlinear classifier which can be applied successfully to differentiate between cancerous and normal tissue regions of an unseen patient.

RESULTS

In a cross-validation strategy, we show an average area under receiver operating characteristic curve (AUC) of 0.93 and classification accuracy of 80%. To validate our results, we present a detailed ultrasound to histology registration framework.

CONCLUSION

Ultrasound RF time series results in differentiation of cancerous and normal tissue with high AUC.

Backscatter coefficient estimation using tapers with gaps

When using the backscatter coefficient (BSC) to estimate quantitative ultrasound parameters such as the effective scatterer diameter (ESD) and the effective acoustic concentration (EAC), it is necessary to assume that the interrogated medium contains diffuse scatterers. Structures that invalidate this assumption can affect the estimated BSC parameters in terms of increased bias and variance and decrease performance when classifying disease. In this work, a method was developed to mitigate the effects of echoes from structures that invalidate the assumption of diffuse scattering, while preserving as much signal as possible for obtaining diffuse scatterer property estimates. Backscattered signal sections that contained nondiffuse signals were identified and a windowing technique was used to provide BSC estimates for diffuse echoes only. Experiments from physical phantoms were used to evaluate the effectiveness of the proposed BSC estimation methods. Tradeoffs associated with effective mitigation of specular scatterers and bias and variance introduced into the estimates were quantified. Analysis of the results suggested that discrete prolate spheroidal (PR) tapers with gaps provided the best performance for minimizing BSC error. Specifically, the mean square error for BSC between measured and theoretical had an average value of approximately 1.0 and 0.2 when using a Hanning taper and PR taper respectively, with six gaps. The BSC error due to amplitude bias was smallest for PR (Nω = 1) tapers. The BSC error due to shape bias was smallest for PR (Nω = 4) tapers. These results suggest using different taper types for estimating ESD versus EAC.

Ultrasound-Based Characterization of Prostate Cancer Using Joint Independent Component Analysis

OBJECTIVE

This paper presents the results of a new approach for selection of RF time series features based on joint independent component analysis for in vivo characterization of prostate cancer.

METHODS

We project three sets of RF time series features extracted from the spectrum, fractal dimension, and the wavelet transform of the ultrasound RF data on a space spanned by five joint independent components. Then, we demonstrate that the obtained mixing coefficients from a group of patients can be used to train a classifier, which can be applied to characterize cancerous regions of a test patient.

RESULTS

In a leave-one-patient-out cross validation, an area under receiver operating characteristic curve of 0.93 and classification accuracy of 84% are achieved.

CONCLUSION

Ultrasound RF time series can be used to accurately characterize prostate cancer, in vivo without the need for exhaustive search in the feature space.

SIGNIFICANCE

We use joint independent component analysis for systematic fusion of multiple sets of RF time series features, within a machine learning framework, to characterize PCa in an in vivo study.

Ultrasound RF time series for classification of breast lesions

This work reports the use of ultrasound radio frequency (RF) time series analysis as a method for ultrasound-based classification of malignant breast lesions. The RF time series method is versatile and requires only a few seconds of raw ultrasound data with no need for additional instrumentation. Using the RF time series features, and a machine learning framework, we have generated malignancy maps, from the estimated cancer likelihood, for decision support in biopsy recommendation. These maps depict the likelihood of malignancy for regions of size 1 mm(2) within the suspicious lesions. We report an area under receiver operating characteristics curve of 0.86 (95% confidence interval [CI]: 0.84%-0.90%) using support vector machines and 0.81 (95% CI: 0.78-0.85) using Random Forests classification algorithms, on 22 subjects with leave-one-subject-out cross-validation. Changing the classification method yielded consistent results which indicates the robustness of this tissue typing method. The findings of this report suggest that ultrasound RF time series, along with the developed machine learning framework, can help in differentiating malignant from benign breast lesions, subsequently reducing the number of unnecessary biopsies after mammography screening.

Quantification of heterogeneity as a biomarker in tumor imaging: a systematic review

BACKGROUND

Many techniques are proposed for the quantification of tumor heterogeneity as an imaging biomarker for differentiation between tumor types, tumor grading, response monitoring and outcome prediction. However, in clinical practice these methods are barely used. This study evaluates the reported performance of the described methods and identifies barriers to their implementation in clinical practice.

METHODOLOGY

The Ovid, Embase, and Cochrane Central databases were searched up to 20 September 2013. Heterogeneity analysis methods were classified into four categories, i.e., non-spatial methods (NSM), spatial grey level methods (SGLM), fractal analysis (FA) methods, and filters and transforms (F&T). The performance of the different methods was compared.

PRINCIPAL FINDINGS

Of the 7351 potentially relevant publications, 209 were included. Of these studies, 58% reported the use of NSM, 49% SGLM, 10% FA, and 28% F&T. Differentiation between tumor types, tumor grading and/or outcome prediction was the goal in 87% of the studies. Overall, the reported area under the curve (AUC) ranged from 0.5 to 1 (median 0.87). No relation was found between the performance and the quantification methods used, or between the performance and the imaging modality. A negative correlation was found between the tumor-feature ratio and the AUC, which is presumably caused by overfitting in small datasets. Cross-validation was reported in 63% of the classification studies. Retrospective analyses were conducted in 57% of the studies without a clear description.

CONCLUSIONS

In a research setting, heterogeneity quantification methods can differentiate between tumor types, grade tumors, and predict outcome and monitor treatment effects. To translate these methods to clinical practice, more prospective studies are required that use external datasets for validation: these datasets should be made available to the community to facilitate the development of new and improved methods.

Clinical study of a noninvasive multimodal sono-contrast induced spectroscopy system for breast cancer diagnosis

PURPOSE

To present a noninvasive multimodal sono-contrast induced spectroscopy (SCIS) system for breast cancer detection.

METHODS

An IRB approved clinical study was carried out to evaluate its diagnostic power. A total of 66 subjects were enrolled with informed consent. The study data were grouped into healthy breast tissue (26), histologically proven cancer (14), and benign mass (26). The diffuse reflectance optical intensity and low intensity focused ultrasound (LIFU) signals, as well as ultrasound images, were collected during each study. The ratio of optical intensities at wavelengths 685 and 830 nm was analyzed using wavelet technique to compare the LIFU effects in cancer and noncancerous tissues. The ultrasound images were also processed to obtain tissue texture parameters, such as correlation, energy, contrast, homogeneity, etc. Backward stepwise regression method was performed to identify the statistically significant factors correlating to tissue types (cancer vs benign mass).

RESULTS

Comparison of the optical signals showed that LIFU induced transitory fluctuation in noncancerous tissue, but not in malignant tissue, as quantified by the ratio of mean absolute deviation (RMAD) of the high frequency component. Statistical analysis revealed that the RMAD ratios were significantly different in tumor vs noncancerous masses (p ≪ 0.01). For tissue texture parameters, energy and correlation were found to statistically correlate with the tissue types. A cancer characterization model was developed using the weighted factors to differentiate the tumor from the benign mass. Trade-off between sensitivity and specificity was obtained by varying the threshold value that estimated the upper-bound of the cancer output factor, from which the receiver-operating characteristic (ROC) curve was generated. The characterization model was optimized using ten modeling datasets and verified using another ten validation datasets randomly generated from the database. The optimization results show that an AUC of 0.93 can be achieved. With threshold 0.3, sensitivity of 96.0%, specificity of 84.1%, and negative predictive value (NPV) of 97.3% can be achieved.

CONCLUSIONS

The feasibility of the multimodal system in characterizing breast cancer vs benign mass is established.

Mean scatterer spacing estimation using multi-taper coherence

It has been hypothesized that estimates of mean scatterer spacing are useful indicators for pathological changes to the liver. A commonly employed estimator of the mean scatterer spacing is the location of the maximum of the collapsed average of coherence of the ultrasound radio-frequency signal. To date, in ultrasound, estimators for this quantity have been calculated with a single taper. Using frequency-domain Monte Carlo simulations, we demonstrate that multi-taper estimates of coherence are superior to single-taper estimates for predicting mean scatterer spacing. Scattering distributions were modeled with Gamma-distributed scatterers for fractional standard deviations in scatterer spacings of 5, 10, and 15% at a mean scatterer spacing of 1 mm. Additionally, we demonstrate that we can distinguish between ablated liver tissue and unablated liver tissue based on signal coherence. We find that, on the average, signal coherence is elevated in the liver relative to signal coherence of received echoes from thermally ablated tissue. Additionally, our analysis indicates that a tissue classifier utilizing the multi-taper estimate of coherence has the potential to distinguish between ablated and unablated tissue types better than a single-taper estimate of coherence. For a gate length of 5 mm, we achieved an error rate of only 8.7% when sorting 23 ablated and 23 unablated regions of interest (ROIs) into classes based on multi-taper calculations of coherence.

Quantitative assessment of in vivo breast masses using ultrasound attenuation and backscatter

Clinical analysis of breast ultrasound imaging is done qualitatively, facilitated with the ultrasound breast imaging-reporting and data system (US BI-RADS) lexicon, which helps to standardize imaging assessments. Two descriptors in that lexicon, "posterior acoustic features" and the "echo pattern" within a mass, are directly related to quantitative ultrasound (QUS) parameters, namely, ultrasound attenuation and the average backscatter coefficient (BSC). The purpose of this study was to quantify ultrasound attenuation and backscatter in breast masses and to investigate these QUS properties as potential differential diagnostic markers. Radio frequency (RF) echo signals were from patients with breast masses during a special ultrasound imaging session prior to core biopsy. Data were also obtained from a well characterized phantom using identical system settings. Masses include 14 fibroadenomas and 10 carcinomas. Attenuation for the acoustic path lying proximal to the tumor was estimated offline using a least squares method with constraints. BSCs were estimated using a reference phantom method (RPM). The attenuation coefficient within each mass was assessed using both the RPM and a hybrid method, and effective scatterer diameters (ESDs) were estimated using a Gaussian form factor model. Attenuation estimates obtained with the RPM were consistent with estimates done using the hybrid method in all cases except for two masses. The mean slope of the attenuation coefficient versus frequency for carcinomas was 20% greater than the mean slope value for the fibroadenomas. The product of the attenuation coefficient and anteroposterior dimension of the mass was computed to estimate the total attenuation for each mass. That value correlated well with the BI-RADS assessment of "posterior acoustic features" judged qualitatively from gray scale images. Nearly all masses were described as "hypoechoic," so no strong statements could be made about the correlation of echo pattern findings in BI-RADS with the averaged BSC values. However, most carcinomas exhibited lower values for the frequency-average BSC than fibroadenomas. The mean ESD alone did not differentiate the mass type, but fibroadenomas had greater variability in ESDs within the ROI than that found for invasive ductal carcinomas. This study demonstrates the potential to use attenuation and QUS parameters associated with the BSC as quantitative descriptors.

Assessing the combined performance of texture and morphological parameters in distinguishing breast tumors in ultrasound images

PURPOSE

This work aims to investigate the combination of morphological and texture parameters in distinguishing between malignant and benign breast tumors in ultrasound images.

METHODS

Linear discriminant analysis was applied to sets of up to five parameters, and then the performances were assessed using the area A(z) (± standard error) under the receiver operator characteristic curve, accuracy (Ac), sensitivity (Se), specificity (Sp), positive predictive value, and negative predictive value.

RESULTS

The most relevant individual parameter was the normalized residual value (nrv), calculated from the convex polygon technique. The best performance among all studied combinations was achieved by two morphological and three texture parameters (nrv, con, std, R, and asm(i)), which correctly distinguished nearly 85% of the breast tumors.

CONCLUSIONS

This result indicates that the combination of morphological and texture parameters may be useful to assist physicians in the diagnostic process, especially if it is associated with an automatic classification tool.

Nonlinear optical microscopy and ultrasound imaging of human cervical structure

The cervix softens and shortens as its collagen microstructure rearranges in preparation for birth, but premature change may lead to premature birth. The global preterm birth rate has not decreased despite decades of research, likely because cervical microstructure is poorly understood. Our group has developed a multilevel approach to evaluating the human cervix. We are developing quantitative ultrasound (QUS) techniques for noninvasive interrogation of cervical microstructure and corroborating those results with high-resolution images of microstructure from second harmonic generation imaging (SHG) microscopy. We obtain ultrasound measurements from hysterectomy specimens, prepare the tissue for SHG, and stitch together several hundred images to create a comprehensive view of large areas of cervix. The images are analyzed for collagen orientation and alignment with curvelet transform, and registered with QUS data, facilitating multiscale analysis in which the micron-scale SHG images and millimeter-scale ultrasound data interpretation inform each other. This novel combination of modalities allows comprehensive characterization of cervical microstructure in high resolution. Through a detailed comparative study, we demonstrate that SHG imaging both corroborates the quantitative ultrasound measurements and provides further insight. Ultimately, a comprehensive understanding of specific microstructural cervical change in pregnancy should lead to novel approaches to the prevention of preterm birth.

Effect of instrument settings on liquid-containing lesion images characterized by radiofrequency ultrasound local estimators

The purpose of this study was to investigate the effects of ultrasonic instrument gain, transducer frequency, and depth on the color variety and color filling of radiofrequency ultrasonic local estimators (RULES) images which indicated specific physical representation of liquid-containing lesions in order to find the optimal settings for the clinical application of RULES in liquid-containing lesions. Changing the ultrasonic instrument gain, transducer frequency, and depth affected the color filling and color variety of 21 pathologically-confirmed liquid-containing lesion images analyzed by RULES. Blue colored fill dominated the RULES images to represent the liquid-containing lesions. A frequency of 12.5MHz led to red and green colors along the inner edges of the liquid-containing lesions. Changing the gain resulted in significantly different blue colored filling that was highest when the gain was 90 to 100. Changing the frequency also significantly changed the blue color filling, with the highest filling occurring at 12.5MHz. Changing the depth did not affect the blue color filling. The liquid components of the lesions may be identified by their characteristic manifestations in RULES, where color variety is affected by transducer frequency and blue color filling which represent liquid-containing lesions in RULES images is affected by frequency and gain.

Quantitative ultrasonic characterization of diffuse scatterers in the presence of structures that produce coherent echoes

Quantitative ultrasound (QUS) techniques that parameterize the backscattered power spectrum have demonstrated significant promise for ultrasonic tissue characterization. Some QUS parameters, such as the effective scatterer diameter (ESD), require the assumption that the examined medium contains uniform diffuse scatterers. Structures that invalidate this assumption can significantly affect the estimated QUS parameters and decrease performance when classifying disease. In this work, a method was developed to reduce the effects of echoes that invalidate the assumption of diffuse scattering. To accomplish this task, backscattered signal sections containing non-diffuse echoes were identified and removed from the QUS analysis. Parameters estimated from the generalized spectrum (GS) and the Rayleigh SNR parameter were compared for detecting data blocks with non-diffuse echoes. Simulations and experiments were used to evaluate the effectiveness of the method. Experiments consisted of estimating QUS parameters from spontaneous fibroadenomas in rats and from beef liver samples. Results indicated that the method was able to significantly reduce or eliminate the effects of nondiffuse echoes that might exist in the backscattered signal. For example, the average reduction in the relative standard deviation of ESD estimates from simulation, rat fibroadenomas, and beef liver samples were 13%, 30%, and 51%, respectively. The Rayleigh SNR parameter performed best at detecting nondiffuse echoes for the purpose of removing and reducing ESD bias and variance. The method provides a means to improve the diagnostic capabilities of QUS techniques by allowing separate analysis of diffuse and non-diffuse scatterers.

Ultrasonic multi-feature analysis procedure for computer-aided diagnosis of solid breast lesions

We have developed quantitative descriptors to provide an objective means of noninvasive identification of cancerous breast lesions. These descriptors include quantitative acoustic features assessed using spectrum analysis of ultrasonic radiofrequency (rf) echo signals and morphometric properties related to lesion shape. Acoustic features include measures of echogenicity, heterogeneity and shadowing, computed by generating spectral-parameter images of the lesion and surrounding tissue. Spectral-parameter values are derived from rf echo signals at each pixel using a sliding-window Fourier analysis. We derive quantitative acoustic features from spectral-parameter maps of the lesion and adjacent areas. We quantify morphometric features by geometric and fractal analysis of traced lesion boundaries. Initial results on biopsy-proven cases show that although a single parameter cannot reliably discriminate cancerous from noncancerous breast lesions, multi-feature analysis provides excellent discrimination for this data set. We have processed data for 130 biopsy-proven patients, acquired during routine ultrasonic examinations at three clinical sites and produced an area under the receiver-operating-characteristics (ROC) curve of 0.947 +/- 0.045. Among the quantitative descriptors, lesion-margin definition, spiculation and border irregularity are the most useful; some additional morphometric features (such as border irregularity) also are particularly effective in lesion classification. Our findings are consistent with many of the BI-RADS (Breast Imaging Reporting and Data System) breast-lesion-classification criteria in use today.

Application of a real-time, calculable limiting form of the Renyi entropy for molecular imaging of tumors

Previously, we reported new methods for ultrasound signal characterization using entropy, H(f); a generalized entropy, the Renyi entropy, I(f)(r); and a limiting form of Renyi entropy suitable for real-time calculation, I(f),(infinity). All of these quantities demonstrated significantly more sensitivity to subtle changes in scattering architecture than energy-based methods in certain settings. In this study, the real-time calculable limit of the Renyi entropy, I(f),(infinity), is applied for the imaging of angiogenic murine neovasculature in a breast cancer xenograft using a targeted contrast agent. It is shown that this approach may be used to reliably detect the accumulation of targeted nanoparticles at five minutes post-injection in this in vivo model.

A feasibility study of novel ultrasonic tissue characterization for prostate-cancer diagnosis: 2D spectrum analysis of in vivo data with histology as gold standard

This study demonstrates the feasibility of using a novel 2D spectrum ultrasonic tissue characterization (UTC) technique for prostate-cancer diagnosis. Normalized 2D spectra are computed by performing Fourier transforms along the range (beam) and the cross-range directions of the digital radio-frequency echo data, then dividing by a reference spectrum. This 2D spectrum method provides axial and lateral information of tissue microstructures, an improvement over the current 1D spectrum analysis which only provides axial information. A pilot study was conducted on four prostate-cancer patients who underwent radical prostatectomies. Cancerous and noncancerous regions of interest, identified through histology, were compared using four 2D spectral parameters: peak value and 3 dB width of the radially integrated spectral power (RISP), slope and intercept of the angularly integrated spectral power (AISP). For noncancerous and cancerous prostatic tissues, respectively, our investigation yielded 23 +/- 1 and 26 +/- 1 dB for peak value of RISP, 7.8 +/- 0.5 degrees and 7.6 +/- 0.6 degrees for 3 dB of RISP, -2.1 +/- 0.2 and -2.7 +/- 0.4 dB/MHz for slope of AISP, and 92 +/- 5 and 112 +/- 6 dB for intercept of AISP. Preliminary results indicated that 2D spectral UTC has the potential for identifying tumor-bearing regions within the prostate gland.

Tissue characterization using multiscale products of wavelet transform of ultrasound radio frequency echoes

This paper presents a novel method for tissue characterization using wavelet transform of ultrasound radio frequency (RF) echo signals. We propose the use of multiscale products of wavelet transform sequences of RF echoes to estimate the scatterer distribution in the tissue. The proposed method is based on the fact that when emitted ultrasound beams interact with scatterers in the tissue, backscattered beams contain singularities corresponding to the location of the scatterers. The singularities will exist in multiple scales of wavelet sequences of the echo signals. Therefore, peaks of wavelet transform multiscale products correspond to the location of scatterers. Estimation of scatterer spacing can be used for tissue characterization. The efficacy of the proposed method was validated in RF echo signals of in-vitro human prostate to characterize normal and cancerous tissue. The results confirm that wavelet transform multiscale products of RF echo signals contain tissue typing information that can be used as an effective tool to differentiate normal and cancerous prostate tissue.

Simplified inverse filter tracking algorithm for estimating the mean trabecular bone spacing

Ultrasonic backscatter signals provide useful information relevant to bone tissue characterization. Trabecular bone microstructures have been considered as quasi-periodic tissues with a collection of regular and diffuse scatterers. This paper investigates the potential of a novel technique using a simplified inverse filter tracking (SIFT) algorithm to estimate mean trabecular bone spacing (MTBS) from ultrasonic backscatter signals. In contrast to other frequency-based methods, the SIFT algorithm is a time-based method and utilizes the amplitude and phase information of backscatter echoes, thus retaining the advantages of both the autocorrelation and the cepstral analysis techniques. The SIFT algorithm was applied to backscatter signals from simulations, phantoms, and bovine trabeculae in vitro. The estimated MTBS results were compared with those of the autoregressive (AR) cepstrum and quadratic transformation (QT) . The SIFT estimates are better than the AR cepstrum estimates and are comparable with the QT values. The study demonstrates that the SIFT algorithm has the potential to be a reliable and robust method for the estimation of MTBS in the presence of a small signal-to-noise ratio, a large spacing variation between regular scatterers, and a large scattering strength ratio of diffuse scatterers to regular ones.

Computer-aided diagnosis of prostate cancer with emphasis on ultrasound-based approaches: a review

This paper reviews the state of the art in computer-aided diagnosis of prostate cancer and focuses, in particular, on ultrasound-based techniques for detection of cancer in prostate tissue. The current standard procedure for diagnosis of prostate cancer, i.e., ultrasound-guided biopsy followed by histopathological analysis of tissue samples, is invasive and produces a high rate of false negatives resulting in the need for repeated trials. It is against these backdrops that the search for new methods to diagnose prostate cancer continues. Image-based approaches (such as MRI, ultrasound and elastography) represent a major research trend for diagnosis of prostate cancer. Due to the integration of ultrasound imaging in the current clinical procedure for detection of prostate cancer, we specifically provide a more detailed review of methodologies that use ultrasound RF-spectrum parameters, B-scan texture features and Doppler measures for prostate tissue characterization. We present current and future directions of research aimed at computer-aided detection of prostate cancer and conclude that ultrasound is likely to play an important role in the field.

Ultrasonic tissue characterization via 2-D spectrum analysis: theory and in vitro measurements

A theoretical model is described for application in ultrasonic tissue characterization using a calibrated 2-D spectrum analysis method. This model relates 2-D spectra computed from ultrasonic backscatter signals to intrinsic physical properties of tissue microstructures, e.g., size, shape, and acoustic impedance. The model is applicable to most clinical diagnostic ultrasound systems. Two experiments employing two types of tissue architectures, spherical and cylindrical scatterers, are conducted using ultrasound with center frequencies of 10 and 40 MHz, respectively. Measurements of a tissue-mimicking phantom with an internal suspension of microscopic glass beads are used to validate the theoretical model. Results from in vitro muscle fibers are presented to further elucidate the utility of 2-D spectrum analysis in ultrasonic tissue characterization.

Benign versus malignant classification of breast tumors based on the the PLSN model for the ultrasound RF echo and homomorphic filtering

The Power-law Shot Noise (PLSN) model has been recently proposed for modeling the ultrasound radio-frequency echo. According to it, the spectrum of the in-phase/quadrature/envelope components are power-law functions. The corresponding power-law exponents were shown to possess good tissue characterization ability. A crucial step in the computation of in-phase/quadrature/envelope components is the estimation of the echo center frequency at different depths. We here propose a robust way of estimating the center frequency. We employ a well known convolutive model for the rf echo that views the echo as convolution of the tissue response and a component that represents the combined effect of the ultrasound impulse response and frequency dependent attenuation. Via low-pass filtering in the cepstrum domain, the combined ultrasonic contribution and attenuation term is extracted and used to estimate the center frequency. Furthermore, the tissue contribution is used to construct two new tissue characterization features. ROC analysis of 65 clinical ultrasound images of the breast indicates that the proposed features combined yield an area of 0.963.

Complexity curve and grey level co-occurrence matrix in the texture evaluation of breast tumor on ultrasound images

This work aims at investigating texture parameters in distinguishing malign and benign breast tumors on ultrasound images. A rectangular region of interest (ROI) containing the tumor and its neighboring was defined for each image. Five parameters were extracted from the complexity curve (CC) of the ROI. Another five parameters were calculated from the grey-level co-occurrence matrix (GLCM) also for the ROI. The same was carried out for internal tumor region, hence, totaling 20 parameters. The linear discriminant analysis was applied to sets of up to five parameters and then the performances were assessed. The most relevant individual parameters were the contrast (con) (from the GLCM over the ROI) and the maximum value (mvi) from the CC just for the tumor internal region). When they were taken together, a correct classification slightly over 80% of the breast tumors was achieved. The highest performance (accuracy=84.2%, sensitivity=87.0%, and specificity=78.8%) was obtained with mvi, con, the standard deviation of the pixel pairs and the entropy, both for GLCM, and the internal region contrast also from GLCM. Parameters extracted from the internal region generally performed better and were more significant than those from the ROI. Moreover, parameters calculated only from CC or GLCM resulted in no statistically significant performance difference. These findings suggest that the texture parameters can be useful to help radiologist in distinguishing between benign or malign breast tumors on ultrasound images.

Improved visualization of high-intensity focused ultrasound lesions

Spectral parameter imaging in both the fundamental and harmonic of backscattered radio-frequency (RF) data were used for immediate visualization of high-intensity focused ultrasound (HIFU) lesion sites. A focused 5-MHz HIFU transducer with a coaxial 9-MHz focused single-element diagnostic transducer was used to create and scan lesions in chicken breast and freshly excised rabbit liver. B-mode images derived from the backscattered RF signal envelope were compared with midband fit (MBF) spectral parameter images in the fundamental (9-MHz) and harmonic (18-MHz) bands of the diagnostic probe. Images of HIFU-induced lesions derived from the MBF to the calibrated spectrum showed improved contrast (approximately 3 dB) of tumor margins versus surround compared with images produced from the conventional signal envelope. MBF parameter images produced from the harmonic band showed higher contrast in attenuated structures (core, shadow) compared with either the conventional envelope (3.3 dB core; 11.6 dB shadow) or MBF images of the fundamental band (4.4 dB core; 7.4 dB shadow). The gradient between the lesion and surround was 3.4 dB/mm, 6.9 dB/mm and 17.2 dB/mm for B-mode, MBF-fundamental mode and MBF-harmonic mode, respectively. Images of threshold and "popcorn" lesions produced in freshly excised rabbit liver were most easily visualized and boundaries best-defined using MBF-harmonic mode.

Detection of breast lesion regions in ultrasound images using wavelets and order statistics

Accurate detection and segmentation of suspicious regions within the complex and irregular tissues of the breast, as depicted with ultrasonic B scans, typically require human analysis and decision making. Tissue characterization methods for classifying suspicious regions often depend on identifying and then accurately segmenting these regions. Motivated by an ultimate goal to automate this critical identification and segmentation step for tissue characterization problems, this work examines ultrasonic signal characteristics between various regions of breast tissue broadly classified as normal tissue and breast lesions. This paper introduces a nonparametric model based on order statistics (OS) estimated from multiresolution (MR) decompositions of energy-normalized subregions. Experimental results demonstrate the classification performance of the OS-based features extracted from the tumor and normal tissue regions in multiple scans from 84 patients, which resulted in a total of 204 tumor regions (from 43 malignant and 161 benign) and 816 normal tissue regions. Performance results indicate that OS-based features achieved an area under the receiver-operator characteristic curve of 91% in the discrimination between breast lesions and surrounding normal tissues.

Characterization of in vitro healthy and pathological human liver tissue periodicity using backscattered ultrasound signals

This work studied the periodicity of in vitro healthy and pathologic liver tissue, using backscattered ultrasound (US) signals. It utilized the mean scatterer spacing (MSS) as a parameter of tissue characterization, estimated by three methods: the spectral autocorrelation (SAC), the singular spectrum analysis (SSA) and the quadratic transformation method (SIMON). The liver samples were classified in terms of tissue status using the METAVIR scoring system. Twenty tissue samples were classified in four groups: F0, F1, F3 and F4 (five samples for each). The Kolmogorov-Smirnov test (applied on group pairs) resulted as nonsignificant (p > 0.05) for two pairs only: F1/F3 (for SSA) and F3/F4 (for SAC). A discriminant analysis was applied using as parameters the MSS mean (MSS) and standard deviation (sigmaMSS), the estimates histogram mode (mMSS), and the speed of US (mc(foie)) in the medium, to evaluate the degree of discrimination among healthy and pathologic tissues. The better accuracy (Ac) with SAC (80%) was with parameter group (MSS, sigmaMSS, mc(foie)), achieving a sensitivity (Ss) of 92.3% and a specificity (Sp) of 57.1%. For SSA, the group with all four parameters showed an Ac of 75%, an Ss of 78.6% and an Sp of 66.70%. SIMON obtained the best Ac of all (85%) with group (MSS, mMSS, mc(foie)), an Ss of 100%, but with an Sp of 50%.

Noninvasive estimation of tissue temperature via high-resolution spectral analysis techniques

We address the noninvasive temperature estimation from pulse-echo radio frequency signals from standard diagnostic ultrasound imaging equipment. In particular, we investigate the use of a high-resolution spectral estimation method for tracking frequency shifts at two or more harmonic frequencies associated with temperature change. The new approach, employing generalized second-order statistics, is shown to produce superior frequency shift estimates when compared to conventional high-resolution spectral estimation methods Seip and Ebbini (1995). Furthermore, temperature estimates from the new algorithm are compared with results from the more commonly used echo shift method described in Simon et al. (1998).

Ultrasonic multifeature tissue characterization for prostate diagnostics

A new system for prostate diagnostics based on multifeature tissue characterization is proposed. Radiofrequency (RF) ultrasonic echo data are acquired during the standard transrectal ultrasound (US) imaging examination. Nine spectral, texture, first order and morphologic parameters are calculated and fed into two adaptive neuro-fuzzy inference systems (FIS) working in parallel. The outputs of the FISs are fed into a postprocessing procedure evaluating contextual information before being combined to form a malignancy map in which areas of high cancer probability are marked in red. The malignancy map is presented to the physician during the examination to improve the early detection of prostate cancer. The system has been evaluated on 100 patients undergoing radical prostatectomy. The ROC curve area using leave-one-out cross-validation over patients is A(Z) = 0.86 when distinguishing between hyperechoic and hypoechoic tumors and normal tissue and A(Z) = 0.84 when distinguishing between isoechoic tumors and healthy tissue, respectively. Tumors that are not visible in the conventional B-mode image can be located. Diagnosis of the prostate carcinoma using multifeature tissue characterization in combination with US imaging allows the detection of tumors at an early stage. Also, biopsy guidance and therapy planning can be improved.

Spectral parameter imaging for detection of prognostically significant histologic features in uveal melanoma

Specific extracellular matrix patterns in uveal melanoma are associated with metastatic risk. The laminin-rich composition and dimensions (on the order of a wavelength or less) of these structures suggest that acoustic backscatter might be affected by their presence. In this study, 10-MHz radiofrequency (RF) ultrasound (US) data were acquired before surgical removal of 117 eyes with uveal malignant melanoma. Histologic sections were evaluated for the presence of matrix patterns and acoustic backscatter was characterized using calibrated spectrum analysis. Statistical correlations between acoustic and histologic patterns were determined and linear discriminant analysis (LDA) and radial basis networks (RBN) were used to develop classification models for histologically based risk groups. Statistically significant correlations were found between acoustic parameters and the presence of histologic matrix-rich patterns. Retrospective classification accuracies of 74.4% and 78.6% were obtained with LDA and RBN, respectively. Leave-one-out analyses indicated estimated predictive accuracies of 71.8% and 75.0% for LDA and RBN, respectively.

Malignant and benign breast tissue classification performance using a scatterer structure preclassifier

Benign and malignant breast tissue classification is examined for generalized-spectrum parameters computed from RF ultrasound data when a preclassification of subregions based on general scattering properties is performed. Results using a clinical database of 84 patients show statistically significant improvements (over 10% in receiver operation characteristic (ROC) areas) when only coherent scatterer subregions are used as compared to using all subregions within the region of interest.

ROC analysis of ultrasound tissue characterization classifiers for breast cancer diagnosis

Breast cancer diagnosis through ultrasound tissue characterization was studied using receiver operating characteristic (ROC) analysis of combinations of acoustic features, patient age, and radiological findings. A feature fusion method was devised that operates even if only partial diagnostic data are available. The ROC methodology uses ordinal dominance theory and bootstrap resampling to evaluate A(z) and confidence intervals in simple as well as paired data analyses. The combined diagnostic feature had an A(z) of 0.96 with a confidence interval of at a significance level of 0.05. The combined features show statistically significant improvement over prebiopsy radiological findings. These results indicate that ultrasound tissue characterization, in combination with patient record and clinical findings, may greatly reduce the need to perform biopsies of benign breast lesions.

Computer aided classification of masses in ultrasonic mammography

Frequency compounding was recently investigated for computer aided classification of masses in ultrasonic B-mode images as benign or malignant. The classification was performed using the normalized parameters of the Nakagami distribution at a single region of interest at the site of the mass. A combination of normalized Nakagami parameters from two different images of a mass was undertaken to improve the performance of classification. Receiver operating characteristic (ROC) analysis showed that such an approach resulted in an area of 0.83 under the ROC curve. The aim of the work described in this paper is to see whether a feature describing the characteristic of the boundary can be extracted and combined with the Nakagami parameter to further improve the performance of classification. The combination of the features has been performed using a weighted summation. Results indicate a 10% improvement in specificity at a sensitivity of 96% after combining the information at the site and at the boundary. Moreover, the technique requires minimal clinical intervention and has a performance that reaches that of the trained radiologist. It is hence suggested that this technique may be utilized in practice to characterize breast masses.