Martin O’Halloran

Development of Clinically Informed 3-D Tumor Models for Microwave Imaging Applications

In this letter, a novel method for the generation of numerical 3-D clinically informed breast tumor models for microwave imaging applications is proposed, which greatly enhances flexibility in creating clinically realistic models. The proposed method was clinically validated through collaboration with breast cancer clinicians and conforms to the BI-RADS labeling standards. Moreover, the issue of clinically accurate tumor positioning within existing breast models is also addressed.

Evaluation of Image Reconstruction Algorithms for Confocal Microwave Imaging: Application to Patient Data

Confocal Microwave Imaging (CMI) for the early detection of breast cancer has been under development for over two decades and is currently going through early-phase clinical evaluation. The image reconstruction algorithm is a key signal processing component of any CMI-based breast imaging system and impacts the efficacy of CMI in detecting breast cancer. Several image reconstruction algorithms for CMI have been developed since its inception. These image reconstruction algorithms have been previously evaluated and compared, using both numerical and physical breast models, and healthy volunteer data. However, no study has been performed to evaluate the performance of image reconstruction algorithms using clinical patient data. In this study, a variety of imaging algorithms, including both data-independent and data-adaptive algorithms, were evaluated using data obtained from a small-scale patient study conducted at the University of Calgary. Six imaging algorithms were applied to reconstruct 3D images of five clinical patients. Reconstructed images for each algorithm and each patient were compared to the available clinical reports, in terms of abnormality detection and localisation. The imaging quality of each algorithm was evaluated using appropriate quality metrics. The results of the conventional Delay-and-Sum algorithm and the Delay-Multiply-and-Sum (DMAS) algorithm were found to be consistent with the clinical information, with DMAS producing better quality images compared to all other algorithms.

Open-Ended Coaxial Probe Technique for Dielectric Measurement of Biological Tissues: Challenges and Common Practices

Electromagnetic (EM) medical technologies are rapidly expanding worldwide for both diagnostics and therapeutics. As these technologies are low-cost and minimally invasive, they have been the focus of significant research efforts in recent years. Such technologies are often based on the assumption that there is a contrast in the dielectric properties of different tissue types or that the properties of particular tissues fall within a defined range. Thus, accurate knowledge of the dielectric properties of biological tissues is fundamental to EM medical technologies. Over the past decades, numerous studies were conducted to expand the dielectric repository of biological tissues. However, dielectric data is not yet available for every tissue type and at every temperature and frequency. For this reason, dielectric measurements may be performed by researchers who are not specialists in the acquisition of tissue dielectric properties. To this end, this paper reviews the tissue dielectric measurement process performed with an open-ended coaxial probe. Given the high number of factors, including equipment- and tissue-related confounders, that can increase the measurement uncertainty or introduce errors into the tissue dielectric data, this work discusses each step of the coaxial probe measurement procedure, highlighting common practices, challenges, and techniques for controlling and compensating for confounders.

A multistage selective weighting method for improved microwave breast tomography.

Microwave tomography has shown potential to successfully reconstruct the dielectric properties of the human breast, thereby providing an alternative to other imaging modalities used in breast imaging applications. Considering the costly forward solution and complex iterative algorithms, computational complexity becomes a major bottleneck in practical applications of microwave tomography. In addition, the natural tendency of microwave inversion algorithms to reward high contrast breast tissue boundaries, such as the skin-adipose interface, usually leads to a very slow reconstruction of the internal tissue structure of human breast. This paper presents a multistage selective weighting method to improve the reconstruction quality of breast dielectric properties and minimize the computational cost of microwave breast tomography. In the proposed two stage approach, the skin layer is approximated using scaled microwave measurements in the first pass of the inversion algorithm; a numerical skin model is then constructed based on the estimated skin layer and the assumed dielectric properties of the skin tissue. In the second stage of the algorithm, the skin model is used as a priori information to reconstruct the internal tissue structure of the breast using a set of temporal scaling functions. The proposed method is evaluated on anatomically accurate MRI-derived breast phantoms and a comparison with the standard single-stage technique is presented.

Supervised Learning Classifiers for Electrical Impedance-based Bladder State Detection

Urinary Incontinence affects over 200 million people worldwide, severely impacting the quality of life of individuals. Bladder state detection technology has the potential to improve the lives of people with urinary incontinence by alerting the user before voiding occurs. To this end, the objective of this study is to investigate the feasibility of using supervised machine learning classifiers to determine the bladder state of ‘full’ or ‘not full’ from electrical impedance measurements. Electrical impedance data was obtained from computational models and a realistic experimental pelvic phantom. Multiple datasets with increasing complexity were formed for varying noise levels in simulation. 10-Fold testing was performed on each dataset to classify ‘full’ and ‘not full’ bladder states, including phantom measurement data. Support vector machines and k-Nearest-Neighbours classifiers were compared in terms of accuracy, sensitivity, and specificity. The minimum and maximum accuracies across all datasets were 73.16% and 100%, respectively. Factors that contributed most to misclassification were the noise level and bladder volumes near the threshold of ‘full’ or ‘not full’. This paper represents the first study to use machine learning for bladder state detection with electrical impedance measurements. The results show promise for impedance-based bladder state detection to support those living with urinary incontinence.

Brain haemorrhage detection using a SVM classifier with electrical impedance tomography measurement frames

Brain haemorrhages often require urgent treatment with a consequent need for quick and accurate diagnosis. Therefore, in this study, we investigate Support Vector Machine (SVM) classifiers for detecting brain haemorrhages using Electrical Impedance Tomography (EIT) measurement frames. A 2-layer model of the head, along with a series of haemorrhages, is designed as both numerical models and physical phantoms. EIT measurement frames, taken from an electrode array placed on the head surface, are used to train and test linear SVM classifiers. Various scenarios are implemented on both platforms to examine the impact of variables such as noise level, lesion location, lesion size, variation in electrode positioning, and variation in anatomy, on the classifier performance. The classifier performed well in numerical models (sensitivity and specificity of 90%+) with signal-to-noise ratios of 60 dB+, was independent of lesion location, and could detect lesions reliably down to the tested minimum volume of 5 ml. Slight variations in electrode layout did not affect performance. Performance was affected by variations in anatomy however, emphasising the need for large training sets covering different anatomies. The phantom models proved more challenging, with maximal sensitivity and specificity of 75% when used with the linear SVM. Finally, the performance of two more complex classifiers is briefly examined and compared to the linear SVM classifier. These results demonstrate that a radial basis function (RBF) SVM classifier and a neural network classifier can improve detection accuracy. Classifiers applied to EIT measurement frames is a novel approach for lesion detection and may offer an effective diagnostic tool clinically. A challenge is to translate the strong results from numerical models into real world phantoms and ultimately human patients, as well as the selection and development of optimal classifiers for this application.

Dielectric properties of bones for the monitoring of osteoporosis

AbstractOsteoporosis is one of the most common diseases that leads to bone fractures. Dual-energy X-ray absorptiometry is currently employed to measure the bone mineral density and to diagnose osteoporosis. Alternatively, the dielectric properties of bones are found to be influenced by bone mineral density; hence, dielectric properties of bones may potentially be used to diagnose osteoporosis. Microwave tomographic imaging is currently in development to potentially measure in vivo dielectric properties of bone. Therefore, the foci of this work are to summarize all available dielectric data of bone in the microwave frequency range and to analyze the confounders that may have resulted in variations in reported data. This study also compares the relationship between the dielectric properties and bone quality reported across different studies. The review suggests that variations exist in the dielectric properties of bone and the relationship between bone volume fraction and dielectric properties is in agreement across all studies. Conversely, the evidence of a relationship between bone mineral density and dielectric properties is inconsistent across the studies. This summary of dielectric data of bone along with a comparison of the relationship between the dielectric properties and bone quality will accelerate the development of microwave tomographic imaging devices for the monitoring of osteoporosis. Graphical abstractᅟ

Diagnosing Breast Cancer with Microwave Technology: Remaining Challenges and Potential Solutions with Machine Learning

Currently, breast cancer often requires invasive biopsies for diagnosis, motivating researchers to design and develop non-invasive and automated diagnosis systems. Recent microwave breast imaging studies have shown how backscattered signals carry relevant information about the shape of a tumour, and tumour shape is often used with current imaging modalities to assess malignancy. This paper presents a comprehensive analysis of microwave breast diagnosis systems which use machine learning to learn characteristics of benign and malignant tumours. The state-of-the-art, the main challenges still to overcome and potential solutions are outlined. Specifically, this work investigates the benefit of signal pre-processing on diagnostic performance, and proposes a new set of extracted features that capture the tumour shape information embedded in a signal. This work also investigates if a relationship exists between the antenna topology in a microwave system and diagnostic performance. Finally, a careful machine learning validation methodology is implemented to guarantee the robustness of the results and the accuracy of performance evaluation.

Parameter Search Algorithms for Microwave Radar-Based Breast Imaging: Focal Quality Metrics as Fitness Functions

Inaccurate estimation of average dielectric properties can have a tangible impact on microwave radar-based breast images. Despite this, recent patient imaging studies have used a fixed estimate although this is known to vary from patient to patient. Parameter search algorithms are a promising technique for estimating the average dielectric properties from the reconstructed microwave images themselves without additional hardware. In this work, qualities of accurately reconstructed images are identified from point spread functions. As the qualities of accurately reconstructed microwave images are similar to the qualities of focused microscopic and photographic images, this work proposes the use of focal quality metrics for average dielectric property estimation. The robustness of the parameter search is evaluated using experimental dielectrically heterogeneous phantoms on the three-dimensional volumetric image. Based on a very broad initial estimate of the average dielectric properties, this paper shows how these metrics can be used as suitable fitness functions in parameter search algorithms to reconstruct clear and focused microwave radar images.

Performance of leading artifact removal algorithms assessed across microwave breast imaging prototype scan configurations

Microwave imaging is a promising imaging modality for the detection of early-stage breast cancer. One of the most important signal processing components of microwave radar-based breast imaging is early-stage artifact removal. Several artifact removal algorithms have been reported in the literature. However, the neighbourhood-based skin subtraction and hybrid artifact removal algorithms have shown particularly promising results in different realistic 3D breast phantoms. For the first time in this paper, both algorithms have been evaluated and compared using the scan approaches of the most common microwave breast imaging prototype systems. The tests include 3D numerical as well as experimental breast phantoms scanned with hemispherical, cylindrical and adaptive scanning patterns. The efficacy of both algorithms has been evaluated across a range of appropriate performance metrics.