Search

Thyroid nodules suspicious for malignancy during clinical-ultrasound (US) evaluation undergo fine-needle-aspiration biopsy (FNAB) followed by cytological evaluation and, in indeterminate cases, an additional surgical intervention or molecular testing is recommended. Quantitative ultrasound (QUS) was investigated as a non-invasive approach to detect thyroid cancer and potentially reduce the currently large number of unnecessary FNABs. RF data were acquired from 225 nodules of 208 patients using a GE Logiq E9 US equipped with a 10-MHz linear-array probe. Five QUS estimates were computed throughout each nodule. Specifically, effective scatter diameter, effective scatter concentration, midband fit, and intercept were obtained using a reference-phantom method. The Nakagami shape parameter was computed using a maximum-likelihood estimator. Mean and standard deviations (SD) of QUS estimates within nodules were used for classification. QUS-based classification performance was quantified using area-under-the-curve (AUC) values obtained from receiver-operating- characteristic analyses. An AUC value of 0.77 was obtained using the SD of intercept alone; the value increased to 0.86 when five QUS estimates were linearly combined. Our in vivo results demonstrate the potential of QUS methods to detect cancerous nodules. Ultimately, the clinical protocol could be altered to prevent FNABs of nodules highly unlikely to be malignant based on QUS analysis.

Ultrasound-guided brachytherapy using titanium-shelled radioactive seeds is a popular, effective means of treating prostate cancer. Unfortunately, implantation using needles inserted transperitoneally causes gland movement and distortion, which often results in seed misplacement and dosimetry errors. If actual seed locations could be determined in the operating room, then corrections to dosimetry errors could be made immediately. However, seed specularity, shadowing, and tissue clutter make imaging seeds difficult using conventional ultrasound. Singular spectrum analysis (SSA) shows promise for reliably imaging radioactive seeds implanted in the prostate and enabling additional corrective implantations to be made in the operating room. SSA utilizes eigenvalues derived from the diagonalized correlation matrix of envelope-detected radio-frequency echo signals to yield a P value indicative of the likelihood of a seed-specific repetitive signal. We demonstrated the potential of SSA for seed detection and imaging and illustrated the trade-off considerations for optimization of SSA in clinical applications using simulations assessing performance as a function of different levels of noise and the presence of repetitive signals with various repetition periods; experiments in an ideal scattering environment; and experiments using seeds implanted in beef.

Currently, transrectal ultrasound (TRUS) guided biopsy is the only method for definitive diagnosis of prostate cancer. Our current study used a high-frequency (i.e., 29-MHz), transrectal, micro-ultrasound system (ExactVu™ micro-ultrasound, Exact Imaging) to acquire RF data from 163 patients immediately before needle firing during 12-core biopsy examinations. Quantitative ultrasound (QUS) estimates of effective scatter diameter (ESD), effective acoustic concentration (EAC), midband (M), intercept (I), and slope (S) were calculated. Additional QUS estimates were derived including envelope statistics employing a Nakagami distribution and the envelope signal-to-noise ratio (ESNR). Estimate values were used to train linear-discriminant classifiers and performance was assessed using area-under-the-curve (AUC) values obtained from receiver operating characteristic (ROC) analyses based on 10-fold cross validation. A combination of ESS- and EAC-related parameters produced an AUC value of 0.75. When ESNR or PSA value was added as a feature, the AUC increased significantly to 0.77 or 0.78, respectively. The best classifier performance was obtained by combining envelope statistics, PSA, ESD, and EAC, which produced an AUC of 0.80. Our initial results with AUC values of 0.80 are very encouraging for developing a new tool for prostate-cancer biopsy guidance and treatment.

Brachytherapy to treat prostate cancer uses transrectal ultrasound to guide implantation of titanium-shelled radioactive seeds. Transperitoneal implantation allows errors in placement that cause suboptimal dosimetry. Conventional ultrasound cannot reliably image implanted seeds; therefore, seed misplacements cannot be corrected in the operating room. Previously, an algorithm based on singular spectrum analysis was shown to image palladium seeds better than B-mode ultrasound could. The algorithm is now applied to imaging an iodine seed in gel and in beef tissue as a function of seed angle relative to the incident ultrasound. Results indicate that both seed types are imaged reliably by the algorithm.

Histological assessment of lymph nodes excised from cancer patients suffers from an unsatisfactory rate of false-negative determinations. We are evaluating high-frequency quantitative ultrasound (QUS) to detect metastatic regions in lymph nodes freshly excised from cancer patients. Three-dimensional (3D) RF data were acquired from 289 lymph nodes of 82 colorectal-, 15 gastric-, and 70 breast-cancer patients with a custom scanner using a 26-MHz, single-element transducer. Following data acquisition, individual nodes underwent step-sectioning at 50-µm to assure that no clinically significant cancer foci were missed. RF datasets were analyzed using 3D regions-of-interest that were processed to yield 13 QUS estimates including spectral-based and envelope-statistics-based parameters. QUS estimates are associated with tissue microstructure and are hypothesized to provide contrast between non-cancerous and cancerous regions. Leave-one-out classifications, ROC curves, and areas under the ROC (AUC) were used to compare the performance of support vector machines (SVMs) and step-wise linear discriminant analyses (LDA). Results showed that SVM performance (AUC = 0.87) was superior to LDA performance (AUC = 0.78). These results suggest that QUS methods may provide an effective tool to guide pathologists towards suspicious regions and also indicate that classification accuracy can be improved using sophisticated and robust classification tools. [Supported in part by NIH grant CA100183.]

We employed quantitative acoustic microscopy (QAM) to measure acoustic properties of tissue microstructure. 32 QAM datasets were acquired from 2, fresh and 11, deparaffinized, 12-µm-thick lymph-node samples obtained from cancer patients. Our custom-built acoustic microscope was equipped with an F-1.16, 250-MHz transducer having a 160-MHz bandwidth to acquire reflected signals from the tissue and a substrate that intimately contacted the tissue. QAM images with a spatial resolution of 7 µm were generated of attenuation (A), speed of sound (SOS), and acoustic impedance (Z). Samples then were stained using hematoxylin and eosin, imaged by light microscopy, and co-registered to QAM images. The spatial resolution and contrast of QAM images were sufficient to distinguish tissue regions consisting of lymphocytes, fat cells and fibrous tissue. Average properties for lymphocyte-dominated tissue were 1552.6 ± 30 m/s for SOS, 9.53 ± 3.6 dB/MHz/cm for A, and 1.58 ± 0.08 Mrayl for Z. Values for Z obtained from fresh samples agreed well with those obtained from 12-µm sections from the same node. Such 2D images provide a basis for developing improved ultrasound-scattering models underlying quantitative ultrasound methods currently used to detect cancerous regions within lymph nodes. [NIH Grant R21EB016117.]

Quantitative-ultrasound (QUS) estimates obtained from spectrum analysis of pulse-echo data are sensitive to tissue microstructure. We investigated the feasibility of obtaining quantitative photoacoustic (QPA) estimates for simultaneously providing sensitivity to microstructure and optical specificity, which could more robustly differentiate among tissue constituents. Experiments were conducted using four, gel-based phantoms (1 × 1 × 2 cm) containing black polyethylene spheres (1E5 particles/ml) that had nominal mean diameters of 23.5, 29.5, 42, or 58 μm. A pulsed, 532-nm laser excited the photoacoustic (PA) response. A 33-MHz transducer was raster scanned over the phantoms to acquire 3D PA data. PA signals were processed using rectangular-cuboidal regions-of-interests to yield three quantitative QPA estimates associated with tissue microstructure: spectral slope (SS), spectral intercept (SI), and effective-absorber size (EAS). SS and SI were computed using a linear-regression approximation to the normalized spectrum. EAS was computed by fitting the normalized spectrum to the multi-sphere analytical solution. The SS decreased and the SI increased with an increase in particle size. While EAS also was correlated with particle size, particle aggregation resulted in EAS estimates that were greater than the nominal particle size. Results indicated that QPA estimates potentially can be used for tissue classification. [Work supported by NIH grant EB015856.]

Histology performed to assess lymph nodes excised during node-dissection surgeries from cancer patients suffers an unsatisfactory rate of false-negative determinations due to labor and time constraints. In this study, more than 300 lymph nodes were scanned in 3D using a 26-MHz high-frequency ultrasound transducer. Following scanning, individual nodes underwent a special histology procedure that involved step-sectioning each node at 50-µm intervals to guarantee that no significant cancer foci were missed. The 3D radio-frequency ultrasound dataset was analyzed using overlapping 3D regions-of-interests that were individually processed to yield thirteen quantitative ultrasound (QUS) estimates associated with tissue microstructure and were hypothesized to show contrast between normal and cancerous regions in lymph nodes. Step-wise linear discriminant analyses were performed to yield an optimal QUS-based classifier. ROC curves and areas under the ROC curves (AUCs) were obtained to assess cancer-detection performance. The AUC for the linear combination of four QUS estimates was 0.83 for a dataset of 110 axillary nodes of breast-cancer patients. Similarly, using five QUS estimates, an AUC of 0.97 was obtained for a dataset of 180 nodes of gastrointestinal-cancer patients. These studies demonstrate that QUS methods may provide an effective tool to guide pathologist towards suspicious regions in lymph nodes.

Human lymph nodes excised from cancer patients during lymphadenectomy can contain small clinically-important metastatic regions that can be missed because conventional histopathology methods do not allow nodes to be examined over their entire volume. In this study, more than 250 lymph nodes were scanned in 3D using a 26-MHz ultrasound transducer before histology processing. Acquired radio-frequency data were processed using 3D regions-of-interest to yield thirteen quantitative ultrasound (QUS) estimates. The QUS estimates are related to tissue microstructure and are hypothesized to be different in normal nodal tissue and metastatic tissue. Four QUS estimates were obtained from backscattered spectra and the remaining nine were derived from envelope statistics. Following ultrasound scanning, serial-section histology was performed at 50-μm intervals to depict cancer foci in 3D. Classification based on QUS estimates was performed using linear-discriminant analyses in a step-wise approach, and areas under ROC ...

High‐frequency quantitative ultrasound (QUS) may offer a reliable means of identifying tumor foci in dissected lymph nodes. Detection of metastases is essential for staging and treatment planning. Conventional histopathology methods do not allow nodes to be examined over their entire volume. Therefore, our objective is to develop QUS methods to improve detection of clinically significant lymph‐node metastases. A single‐element 26‐MHz ultrasound transducer was used to scan and digitally acquire rf echo‐signal data in three‐dimensional (3D) from more than 200 lymph nodes. Thirteen QUS estimates based on backscatter spectra and envelope statistics were computed in 3D. Serial‐sectioning histology was performed at 50‐μm intervals to depict cancer foci in 3D. Classification based on QUS estimates was performed using linear‐discriminant analyzes; areas under ROC curves (AUCs) were computed. The most‐significant QUS estimates for metastases detection were identified. Comparison of the 3D QUS results and 3D histol...

Brachytherapy for prostate‐cancer treatment involves permanent implantation of radioactive seeds into the gland. Reliable imaging of implanted seeds is needed to correct dosimetry errors in the operating room. A singular‐spectrum analysis algorithm previously was able to detect palladium seeds over various angles between the seed and ultrasound‐beam axes. We subsequently evaluated algorithm performance with iodine seeds. The algorithm extracts pairs of eigenvalues from the autocorrelation matrix of seed echo signals. Selected eigenvalues are used to compute a P‐value for seed presence. The algorithm was applied to echo signals obtained using a 5‐MHz transducer to scan seeds implanted into an acoustically transparent gel pad and a piece of ex vivo beef. The angle of the seed axis with respect to the beam axis was varied from normal to the beam axis to 20 deg from normal. Algorithm performance was denoted by a score computed from P‐values. Scores for both seed types in beef varied from 70 to 40 dB over the ...

High‐frequency (>15 MHz, i.e., wavelength

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 successfull...

Backscattered echo signals are related to scatterer properties in theoretical frameworks published independently by Lizzi and Insana. These frameworks are based on assumptions that may not consistently apply to tissue scatterers. Our study aims to confirm and expand the theory to include scattering under a variety of scatterer conditions. A 35‐MHz transducer was used to scan suspended 12‐μm‐diam uveal melanoma cells. Using Fourier methods, spectral parameters were computed from ultrasonic echo data acquired from the suspension. Normalization was implemented in two ways. First, a glass‐plate calibration spectrum was subtracted from the experimental data. Second, the spectrum acquired from suspended 20‐μm beads was compared to the spectrum predicted by Faran’s theory. The difference between expected and experimental spectra was attributed to system effects. Cell acoustic impedance was estimated from cell‐pellet density and propagation velocity. Theoretical spectral values were computed from cell size, conce...

A singular‐spectrum‐analysis (SSA) algorithm previously was shown to be successful in detecting brachytherapy seeds in B‐mode images when seeds were orthogonal to the ultrasound beam. In this study, the algorithm was extended to detect seeds at angles to the beam. The SSA algorithm derives P‐values from selected eigenvalue pairs of the autocorrelation matrix of seed echo signals. The P‐value indicates the probability of presence of a seed. A seed inserted in a gel pad and another inserted in beef tissue were scanned with a 5‐MHz transducer. The seed angle was varied from orthogonal to the beam to 23 deg from orthogonal, using 1‐degree increments. Simulations of seed echoes based on empirical data were generated and used to test the algorithm. Success in seed detection was expressed by a score computed from P‐values. Scores for simulations decreased from 0.8 to 0.45 as the angle increased from 0 to 12 deg. Scores for experimental data varied from 0.74 at 0 deg to 0.62 at 23 deg; the lowest value was 0.1 at 7 deg. Accordingly, the SSA algorithm was successful in detecting the seed with clinically relevant angles between seed and beam.

Unlike many tumors, prostate cancer cannot be imaged reliably by any single, existing, conventional modality. This makes guiding biopsies, targeting therapy, and monitoring effects problematic. Therefore, our objective is to improve significantly the depiction of prostate cancers using multimodality, multifeature techniques combined with nonlinear classifiers such as artificial neural networks (ANNs) and support vector machines (SVMs). Our previous studies show encouraging potential for tissue typing and imaging based on ultrasound spectrum analysis using ANNs and SVMs that produce ROC‐curve areas of 0.84 to 0.88 compared to 0.64 for conventional ultrasound imaging. Similar studies by others have shown that magnetic‐resonance spectroscopy can produce comparable ROC‐curve areas based on the ratio of choline and creatine to citrate. Because these two modalities sense entirely distinct properties of tissue, i.e., microscopic mechanical properties by ultrasound and chemical properties by magnetic resonance, their combination should markedly improve classification performance in distinguishing between cancerous and noncancerous prostate tissue. We have undertaken preliminary studies incorporating these two modalities. These studies demonstrate the feasibility of coregistering features from each modality and generating new, combined‐modality images that exploit the attributes of both modalities.

Existing theoretical frameworks for ultrasound spectrum analysis relate properties of backscattered echo signals from tissues to scatterer properties. However, the reliance of these frameworks on numerous key assumptions may not be appropriate for evaluating tissue properties. The objective of our studies is to validate those assumptions or reduce dependence on them so that estimates of scatterer properties can be made with greater confidence. Ultrasound rf data were acquired from isolated, uveal melanoma cells suspended in liquid media using center frequencies of 10 and 35 MHz. Average cell size was 12 μm and concentrations ranged from 1.3×108 cells/ml to 2.0×106 cells/ml. Theoretical spectral‐parameter values predicted by the Lizzi framework were calculated using known values for size and concentration and published values for sound propagation velocity and acoustic impedance. Normalized spectral parameters were computed from the experimental rf data using standard Fourier methods. Predicted and experimental values for the slope, midband, and intercept parameters agreed to within 11%. Initial results tend to validate the theory at these scatterer concentrations. Future work will provide more insight into scattering behavior under a greater variety of conditions, e.g., scatterer sizes and concentrations.

Singular spectrum analysis (SSA) shows promise for reliably imaging radioactive seeds implanted in the prostate gland. Brachytherapy using small titanium‐shelled radioactive seeds is proving to be a well‐accepted means of treating prostate cancer. Unfortunately, implantation using needles inserted transperineally causes gland movement, and the resulting distortion may cause seed misplacement and dosimetry errors. If actual seed locations can be determined during the brachytherapy procedure, then dosimetry errors can be corrected immediately in the operating room. SSA utilizes pairs of eigenvalues derived from the autocorrelation matrix of envelope‐detected radiofrequency echo signals to identify seed‐specific signal repetitions. The power spectrum associated with a repetition signal is computed to derive a P‐value indicative of the likelihood of the presence of a seed at the location of that repetitive signal. Simulations to assess performance as a function of different levels of white and speckle noise and the presence of signals at repetition periods not associated with seeds. Experiments in an ideal scattering environment were performed, using seeds implanted in beef. Results show an encouraging potential of SSA for seed detection and imaging, and illustrate the basic trade‐off considerations for optimization of SSA for clinical applications. [Work supported by NIH Grant R01 CA098465.]

No reliable method of imaging prostate cancer currently exists. These prostate tissue‐typing studies aim to improve the effectiveness of biopsy guidance and treatment monitoring by developing better imaging methods for identifying cancerous prostate tissue. Success will reduce the false‐negative rate of biopsies and treatment side‐effects. Ultrasonic (US) radio‐frequency (rf) echo‐signal data acquired during biopsy examinations were used to compute spectral parameters. These spectral parameters along with clinical parameters, e.g., PSA, were used to train a neural network classifier using biopsy results as the gold standard. Cancer‐likelihood scores from a look‐up table were used to generate tissue‐type images (TTIs). The ROC‐curve area for US neural‐network‐based classification was 0.844±0.018 vs 0.638±0.031 for B‐mode‐based classification, and the sensitivity of neural‐network based classification was superior to that of B‐mode‐guided biopsies. These classification methods are being extended to include magnetic‐resonance spectral (MRS) techniques that use the choline to citrate ratio‐to‐distinguish cancerous from noncancerous prostate tissue. 3D renderings of prostatectomy histology, US, and MR images show encouraging correlations, and combining MRS parameters with US spectral parameters appears to have potential to further improve prostate‐cancer imaging. [Work supported in part by NIH Grant CA053561.]

Interstitially implanted radioactive seeds are becoming a popular and effective means of treating a wide range of diseases, primarily cancer. Brachytherapy of prostate cancer, for example, is proving to be as effective, in terms of 5‐year survival, as surgery for management of gland‐confined disease, and to many, the side effects of brachytherapy are less debilitating than those of surgery. While seeds are implanted under guidance from transrectal ultrasound, gland movement and distortion during implantation may lead to seed misplacement; implanted seeds are difficult to visualize ultrasonically after implantation for a variety of reasons. We are investigating a variety of ultrasonic techniques intended to overcome the limitations of conventional ultrasonic imaging, and the following three methods are showing encouraging promise: resonance/vibration imaging; modified elastography; and two‐dimensional correlation. The resonance/vibration method induces natural vibrations in seeds and, using power‐Doppler m...

Three‐dimensional (3‐D) depictions of ultrasonic scan data have been developed in our laboratories for visualizing tissue properties and structure in several medical applications including characterization of prostatic and ophthalmic tissues. Data for 3‐D depictions are acquired as a set of parallel ultrasonic scans and are displayed as volume or surface renderings or as interactively dissectable blocks. Acquired data consist of digitized radio‐frequency (rf) echo signals. Tissue‐typing information is based on parameters derived from linear‐regression analyses of normalized power spectra of the digitized rf echo‐signal data. Tissue properties, such as scatterer size, are computed from spectral parameter values. Tissue typing is performed by comparing computed parameter values from an unknown tissue with database values associated with known tissue types. Images portraying spectral parameter values, scatterer properties, or most‐likely tissue type have been developed. Three‐dimensional presentations are generated from sets of scan‐plane data using desktop computers or imaging workstations. Volumes can be examined in several ways, e.g., by dynamically presenting the entire 3‐D volume as a tilting or a rotating transparent rendering. Three‐dimensional images of the prostate gland facilitate visualization of cancerous regions warranting biopsy. [Research supported, in part, by the National Cancer Institute and the National Eye Institute.]

Ultrasonic tissue characterization based on spectrum analysis of radio‐frequency (rf) echo signals is proving to be clinically useful in ophthalmology and abdominal studies. Parameters computed using spectrum analysis can be theoretically related to scattering properties of tissue such as scatterer size and impedance and therefore to histology. Such parameters can also be correlated empirically with clinical features, histopathology, treatment history, etc. They appear to be useful for clinical evaluation of disease, particularly detecting, diagnosing, staging, and monitoring cancer. Although parameters can be presented graphically or numerically, two‐dimensional parameter images significantly facilitate the interpretation provided by tissue characterization. The portrayal of parameters in three dimensions further enhances the visualization of parameter distributions, and provides an even more powerful tool for tissue evaluation. In addition, three‐dimensional (3‐D) presentations of the surfaces of tissue...

An analytic model is described for application in ultrasonic tissue characterization. The model is applicable to clinical broadband pulse echo systems. It treats spectra derived from received echo signals and relates them to physical tissue properties. The model can be applied to deterministic tissue structures (e.g., retinal detachments, larger blood vessels, and surface layers of the kidney) and to stochastic tissue structures (e.g., various tumors). The beam patterns included in the model are those generated by focused transducers typically used in high-resolution clinical ultrasound. Appropriate calibration procedures are also treated; these are needed for interpretation of absolute spectral parameters. The results obtained with the analytic model have been used to design a digital processing system and the associated techniques which are now being applied during examinations of the eye and abdominal organs. The results have proven useful in interpreting data from various types of tissues. To illustrate the application of these results, representative clinical data, obtained from the digital system, are presented for two types of tissue architectures. The first case is a detached retina representing a deterministic structure characterized by well-defined thickness and reflection coefficients. The second case is asteroid hyalosis and represents a stochastic entity in which the positions of small scattering particles are best described in statistical terms, and characterization is accompanied by means of normalized power spectra.

Ultrasonic characterization, using digital techniques, is being applied to clinical data gathered in examination of the eye and orbit. Digital rf echo signals are obtained from structures situated within a 7.5 × 20 mm region of the eye. A dedicated minicomputer system is employed to select specific portions of these records for analysis. Using a simple instruction set, a variety of analysis techniques can be chosen to characterize specific types of structures. For example, small ocular tumors are examined with spectrum analysis techniques to delineate surface features, internal scattering characteristics, and shadowing of posterior scatterers such as orbital fat elements. The use of analysis procedures is guided by the display of reconstructed, gray‐scale B‐scans. Computational results are presented in a variety of display formats (including color‐coded and isometric modes) to facilitate recognition of fundamental tissue features. Data are catalogued for continuing evaluation of tissue characteristics tha...