Dun Li

A clinical prototype for active microwave imaging of the breast

Despite its recognized value in detecting and characterizing breast disease, X-ray mammography has important limitations that motivate the quest for alternatives to augment the diagnostic tools that are currently available to the radiologist. The rationale for pursuing electromagnetic methods is strong given the data in the literature, which show that the electromagnetic properties of breast malignancy are significantly different than normal in the high megahertz to low gigahertz spectral range, microwave illumination can effectively penetrate the breast at these frequencies, and the breast is a small readily accessible tissue volume, making it an ideal site for deploying advanced near-field imaging concepts that exploit model-based image reconstruction methodology. In this paper a clinical prototype of a microwave imaging system, which actively illuminates the breast with a 16-element transceiving monopole antenna array in the 300-1000 MHz range, is reported. Microwave exams have been delivered to five women through a water-coupled interface to the pendant breast with the participant positioned prone on an examination table. This configuration has been found to be a practical, comfortable approach to microwave breast imaging. Sessions lasted 10-15 min per breast and included full tomographic data acquisition at seven different array heights beginning at the chest wall and moving anteriorly toward the nipple for seven different frequencies at each array position. This clinical experience appears to be the first report of active near-field microwave imaging of the breast and is certainly the first attempt to exploit model-based image reconstructions from in vivo breast data in order to convert the measured microwave signals into spatial maps of electrical permittivity and conductivity. While clearly preliminary, the results are encouraging and have supplied some interesting findings. Specifically, it appears that the average relative permittivity of the breast as a whole correlates with radiologic breast density categorization and may be considerably higher than previously published values, which have been based on ex vivo tissue specimens.

Conformal microwave imaging for breast cancer detection

An important feature of our Gauss-Newton iterative scheme for microwave breast image reconstruction is that only the heterogeneous target zone within the antenna array is represented using the finite-element method, while the surrounding homogeneous coupling medium is modeled with the boundary-element (BE) method. The interface between these two zones may be arbitrary in shape and position with the restriction that the BE region contains only the homogeneous coupling liquid. In this paper, we demonstrate through simulation and phantom experiments, as well as in patient examinations, that the detection of tumor inclusions can be enhanced as the target zone approaches the exact breast perimeter. It is also shown that central artifacts that often appear in the reconstructed images that may potentially confound the ability to distinguish benign and malignant conditions have also been reduced.

Parallel-detection microwave spectroscopy system for breast imaging

A liquid-coupled, noncontacting, broadband microwave imaging system has been designed and fabricated. Extension of the operating bandwidth allows us to exploit the potential of new clinical information for breast cancer diagnosis at frequencies higher than previously achieved. The new system design implements a parallel-detection scheme that allows signals to be simultaneously sampled at multiple receiving antenna sites (in 8 s for a single tomographic slice at a single frequency). It also has important features such as high cross-channel isolation (>120 dB), smooth broad bandwidth receiver response, and adjustable intermediate frequency signal amplification factors of 1 to 2000 to ensure successful realization of a large linear dynamic range which is especially important to counteract the increased signal loss at the higher operating frequencies. The new system is capable of recovering dielectric properties of breastlike phantoms with tumor inclusions over the frequency range from 0.5 to 2.1 GHz when embedded in an 87%/13% glycerin/water background. Errors in the measurement data are less than 0.5% in signal amplitude and 1° in phase, on average.

A two-stage microwave image reconstruction procedure for improved internal feature extraction.

We have developed a two-stage Gauss-Newton reconstruction process with an automatic procedure for determining the regularization parameter. The combination is utilized by our microwave imaging system and has facilitated recovery of quantitatively improved images. The first stage employs a Levenberg-Marquardt regularization along with a spatial filtering technique for a few iterations to produce an intermediate image. In effect, the first set of iterative image reconstruction steps synthesizes a priori information from the measurement data versus actually requiring physical prior information on the interrogated object. Because of the interaction of the Levenberg-Marquardt regularization and spatial filtering at each iteration, the intermediate image produced from the first reconstruction stage represents an improvement in terms of the least squared error over the initial uniform guess; however, it has not completely converged in a least squared sense. The second stage involves using this distribution as a priori information in an iteratively regularized Gauss-Newton reconstruction with a weighted Euclidean distance penalty term. The penalized term restricts the final image to a vicinity (determined by the scale of the weighting parameter) about the intermediate image while allowing more flexibility in extracting internal object structures. The second stage makes use of an empirical Bayesian/random effects model that enables an optimal determination of the weighting parameter of the penalized term. The new approach demonstrates quantifiably improved images in simulation, phantom and in vivo experiments with particularly striking improvements with respect to the recovery of heterogeneities internal to large, high contrast scatterers such as encountered when imaging the human breast in a water-coupled configuration.

Comparisons of three alternative breast modalities in a common phantom imaging experiment.

Four model-based imaging systems are currently being developed for breast cancer detection at Dartmouth College. A potential advantage of multimodality imaging is the prospect of combining information collected from each system to provide a more complete diagnostic tool that covers the full range of the patient and pathology spectra. In this paper it is shown through common phantom experiments on three of these imaging systems that it was possible to correlate different types of image information to potentially improve the reliability of tumor detection. Imaging experiments were conducted with common phantoms which mimic both dielectric and optical properties of the human breast. Cross modality comparison was investigated through a statistical study based on the repeated data sets of reconstructed parameters for each modality. The system standard error between all methods was generally less than 10% and the correlation coefficient across modalities ranged from 0.68 to 0.91. Future work includes the minimization of bias (artifacts) on the periphery of electrical impedance spectroscopy images to improve cross modality correlation and implementation of the multimodality diagnosis for breast cancer detection.

Microwave thermal imaging: initial in vivo experience with a single heating zone

The deployment of hyperthermia as a routine adjuvant to radiation or chemotherapy is limited largely by the inability to devise treatment plans which can be monitored through temperature distribution feedback during therapy. A non-invasive microwave tomographic thermal imaging system is currently being developed which has previously exhibited excellent correlation between the recovered electrical conductivity of a heated zone and its actual temperature change during phantom studies. To extend the validation of this approach in vivo, the imaging system has been re-configured for small animal experiments to operate within the bore of a CT scanner for anatomical and thermometry registration. A series of 5–7 day old pigs have been imaged during hyperthermia with a monopole antenna array submerged in a saline tank where a small plastic tube surgically inserted the length of the abdomen has been used to create a zone of heated saline at pre-selected temperatures. Tomographic microwave data over the frequency range of 300–1000 MHz of the pig abdomen in the plane perpendicular to the torso is collected at regular intervals after the tube saline temperatures have settled to the desired settings. Images are reconstructed over a range of operating frequencies. The tube location is clearly visible and the recovered saline conductivity varies linearly with the controlled temperature values. Difference images utilizing the baseline state prior to heating reinforces the linear relationship between temperature and imaged saline conductivity. Demonstration of in vivo temperature recovery and correlation with an independent monitoring device is an important milestone prior to clinical integration of this non-invasive imaging system with a thermal therapy device.

Image accuracy improvements in microwave tomographic thermometry: phantom experience

Evaluation of a laboratory-scale microwave imaging system for non-invasive temperature monitoring has previously been reported with good results in terms of both spatial and temperature resolution. However, a new formulation of the reconstruction algorithm in terms of the log-magnitude and phase of the electric fields has dramatically improved the ability of the system to track the temperature-dependent electrical conductivity distribution. This algorithmic enhancement was originally implemented as a way of improving overall imaging capability in cases of large, high contrast permittivity scatterers, but has also proved to be sensitive to subtle conductivity changes as required in thermal imaging. Additional refinements in the regularization procedure have strengthened the reliability and robustness of image convergence. Imaging experiments were performed for a single heated target consisting of a 5.1 cm diameter PVC tube located within 15 and 25 cm diameter monopole antenna arrays, respectively. The performance of both log-magnitude/phase and complex-valued reconstructions when subjected to four different regularization schemes has been compared based on this experimental data. The results demonstrate a significant accuracy improvement (to 0.2° C as compared with 1.6° C for the previously published approach) in tracking thermal changes in phantoms where electrical properties vary linearly with temperature over a range relevant to hyperthermia cancer therapy.

A broadband microwave breast imaging system

A liquid-coupled, non-contacting, broadband (0.5 - 3 GHz) microwave imaging system has been designed and fabricated to exploit new clinical information for breast cancer diagnosis. Ultimately this system may complement conventional X-ray mammography by reducing the number of false negatives and false positives, especially in challenging cases such as patients with radiographically dense breast tissue.

Conformal imaging with a non-contacting microwave antenna array

A model-based, non-contacting microwave imaging system, which allows for an arbitrarily shaped imaging domain to exist within the antenna array, has been developed using our Gauss-Newton iterative image reconstruction approach based on a hybrid of the finite element (FE) and boundary element (BE) methods. A new feature has been introduced to conform to the reconstructed field of view exactly to the coupling medium/object interface. This facilitates deployment of the reconstruction parameters solely to the zone occupied by the object, potentially improving resolution. Enhancements using this technique have been demonstrated in both simulations and phantom experiments.

Spectrum analysis of microwave breast examination data and reconstructed images

A broadband microwave tomographic imaging system is being developed for breast cancer detection which exploits the significant electrical property contrast between normal and malignant tissue. The improved image reconstruction algorithm directly processes the log-magnitude and phase values of the measured fields with the frequency rich data set being utilized to acquire the absolute scattered phase terms, especially when multiple wrapping has occurred. In addition, a lower contrast coupling medium has been selected which alleviates certain problems associated with the phase unwrapping task but also appears to reduce artifacts associated with higher permittivity contrasts between the surrounding coupling medium and the breast. Sequences of multi-plane image sets from a low contrast medium examination of a 53 year old woman over a broad frequency range are presented to illustrate the general consistency of the images over the range and to demonstrate the improved resolution at the higher frequencies.