Coherent change detection microwave imaging system for medical applications

Abstract

A medical imaging system and method for detecting temporal changes in biological tissue through coherent change detection of complex-valued microwave signals. Unlike conventional microwave imaging systems that acquire single-timepoint data, the disclosed system maintains phase coherence between temporally separated measurements using dedicated real-time calibration channels that continuously monitor and correct phase drift in the signal path, as well as physical and algorithmic registration and compensation methods. The system acquires magnitude and phase backscatter data from tissue at multiple timepoints, applies spatial registration to compensate for patient repositioning, and computes coherent differences from the registered complex data to detect changes in tissue dielectric and conductivity properties below the spatial resolution limit of single-timepoint imaging. The coherent change detection method enables identification of tissue changes and growth with sensitivity exceeding conventional temporal subtraction techniques. Applications include breast cancer screening and monitoring, intracranial hemorrhage detection, and lung cancer screening and monitoring.

Description

COHERENT CHANGE DETECTION MICROWAVE IMAGING SYSTEM FOR MEDICAL APPLICATIONSCROSS-REFERENCE TO RELATED APPLICATIONSThis nonprovisional patent application claims priority to and the benefit of U.S. Provisional Patent Application No. 63 / 706,688, filed October 13, 2024, entitled "Coherent Change Detection Microwave Imaging System for Medical Applications," which is incorporated herein by reference in its entirety.TECHNICAL FIELD

[0001] This invention relates generally to medical imaging systems and methods, and more particularly to coherent change detection microwave imaging systems for detecting temporal changes in biological tissue to identify pathological conditions such as cancer in breast, lung, brain, and other organ tissues, and for monitoring treatment efficacy.BACKGROUND

[0002] Medical imaging plays a crucial role in early detection and monitoring of pathological conditions, particularly cancer. Current imaging modalities including X- ray mammography, magnetic resonance imaging (MRI), and ultrasound have known limitations. For example, X-ray mammography exhibits reduced sensitivity in dense breast tissue, involves ionizing radiation, and requires painful breast compression. Dense breast tissue, classified as BI-RADS C (heterogeneously dense) or BI-RADS D (extremely dense) under the American College of Radiology Breast Imaging Reporting and Data System, affects approximately 50% of screening-age women and represents a critical challenge for mammography-based detection. In dense breast tissue, mammography sensitivity drops to 62-68% compared to 87-97% in fatty breast tissue, resulting in delayed detection of malignant tumors. The resulting reduction in mortality based on annual mammogram-based screening is 10% for the roughly half of women with dense breast tissue and 40% for the half of women with fatty breast tissue. The mammography-associated ionizing radiation limits screening frequency, particularly in younger women and high-risk populations. MRI systems require significant capital investment and have high associated examination costs,may involve contrast agents, and typically have long examination times. Microwave imaging has emerged as a promising non-ionizing alternative, focusing on the ability to detect the increased dielectric constant and conductivity associated with malignant tissue. However, conventional microwave imaging has faced challenges related to detecting small dielectric contrasts, achieving high spatial resolution, and maintaining performance consistency. Published studies have shown that while malignant breast tissue exhibits dielectric properties approximately 3-10 times greater than adipose tissue at microwave frequencies (0.3-10 GHz), the contrast between malignant and normal glandular or fibroconnective tissues is no more than 10% (Lazebnik et al., Physics in Medicine & Biology, Vol. 52, p. 6093, 2007), presenting significant detection challenges due in part to the significant measurement variability of the imaging techniques to date. Conventional microwave imaging systems have not achieved the signal-to-clutter ratios necessary for the reliable detection in this challenging tissue environment. Additionally, there is a conflict between obtaining the resolution required to detect malignant tumors (ideally less than 5mm), which require high frequencies (e.g. 5-10GHz), and penetrating deep enough into the tissue to measure the tumor reflections (5-10cm in some cases), which require lower frequencies (e.g. 100MHz - 2 GHz).

[0003] Coherent change detection (CCD) techniques have been developed for other fields, such as ground-penetrating radar for detecting subsurface features. Such systems are typically designed for detecting high-contrast, discrete changes in relatively rigid and static geological media. Because the media is primarily static, registration and maintaining coherence is simpler to implement between scans. These applications do not address the unique challenges of microwave medical imaging, which involves detecting subtle, low-contrast changes within soft, deformable, and physiologically active biological tissue that is subject to repositioning and growth variations between scans. Additionally, biological tissue requires specialized impedance matching at the tissue interface, respiratory motion compensation, and detection algorithms specific to vascularization patterns and angiogenesis - which are not present in geological GPR applications. Therefore, a need exists for a medical imaging system that adapts coherent change detection techniques specifically for monitoring biological tissue using microwave imaging.SUMMARY

[0004] The present invention provides a coherent change detection microwave imaging system specifically designed for medical applications that overcomes the limitations of prior art. Biological tissue changes associated with malignant growth include increased vascularization (angiogenesis) and cellular proliferation. These physiological changes manifest as alterations in tissue dielectric properties due to increased blood perfusion, higher cellular density, and modified water content. The present invention detects these changes through coherent analysis of complexvalued microwave measurements, enabling identification of pathological tissue development between temporally separated imaging sessions.Primary Embodiments

[0005] In one aspect, the invention comprises a coherent change detection microwave imaging system for detecting temporal changes in biological tissue. The system includes at least one microwave electronics module configured to generate and receive ultra-wideband microwave signals. The system further includes a plurality of antenna elements operatively connected to the microwave electronics module. The system also includes at least one calibration channel configured to compensate for system variations in real-time. The system additionally includes at least one control system configured to illuminate a target tissue region with microwave signals at multiple timepoints, measure amplitude and phase characteristics of signals from the target tissue region, maintain phase coherence between measurements taken at different timepoints, and compare complex-valued measurements between timepoints to detect dielectric property changes in the tissue.

[0006] Some embodiments of this system include wherein the ultra-wideband microwave signals span a frequency range of 0.3-10 GHz. In other embodiments, a plurality of antenna elements is shielded to reduce mutual coupling below -20dB to - 30dB across the operating bandwidth, achieved through individual element shielding and enclosures. Some embodiments further comprise a positioning mechanism configured to move the antenna elements relative to the target tissue region,wherein the antenna elements are housed within an enclosure having inner surfaces covered with radio frequency (RF) absorbing material to suppress cavity modes and reduce electromagnetic interference. Additional embodiments further comprise an image reconstruction module configured to generate tissue images based on the detected dielectric changes using hybrid radar-tomographic reconstruction algorithms.

[0007] In certain embodiments, the microwave electronics module comprises a direct digital synthesis (DDS) module for frequency generation with low phase noise to ensure stability, a heterodyne receiver architecture with intermediate frequency processing typically having a bandwidth of approximately 50 MHz or lower, analog- to-digital converters with at least 14-bit resolution, preferably 16-bit resolution, and sampling rates of 100-500 MSPS (mega-samples per second), or field- programmable gate arrays (FPGAs), custom chips, DSPs, processors, or GPUs for signal processing. Some embodiments include antenna elements arranged with element spacing of less than or equal to A / 10 of the ultra-wideband microwave signals, where A is the wavelength at the center frequency. Other embodiments further comprise one or more temperature sensors integrated with the microwave electronics module, wherein the control system or processing module is configured to use data from the temperature sensors to perform predictive thermal compensation and correct for phase drift caused by thermal variations.

[0008] In an alternative embodiment employing direct sampling architecture, the microwave electronics module comprises high-speed analog-to-digital converters configured to directly digitize the amplified radio frequency signals at sampling rates exceeding 1 giga-sample per second (GSPS), in some embodiments in the range of 10-20 GSPS or higher. The digitized signals are then processed in the digital domain using field-programmable gate arrays (FPGAs) or application-specific integrated circuits (ASICs) that perform digital filtering, down-conversion, and signal conditioning operations. This direct sampling approach eliminates analog mixing stages and enables flexible software-defined receiver configurations.

[0009] In some embodiments, the system further comprises a machine learning module configured to classify tissue changes based on training data including frequency-dependent response signatures of malignant tissue, adapt detection thresholds based on patient-specific characteristics and tissue density classifications, and improve performance through population data analysis. Some embodiments include a control system configured to enable performance improvement without hardware modification by implementing receiving algorithm updates through secure remote updates, adapting operating parameters based on accumulated population data, or maintaining backward compatibility with previously acquired measurement data.

[0010] Additional embodiments include an antenna array adaptable to different anatomical geometries, wherein the control system is configured to adjust operating parameters including frequency range, antenna positioning, and signal processing algorithms based on target tissue type, thereby enabling configuration for multiple anatomical applications including breast, lung, and brain imaging. In some embodiments, antenna elements may be adapted to use higher dielectric constant material inserts between 1 .25 to approximately 60 to reduce the size of the antenna, improve impedance matching with the skin, or change the radiation pattern for certain frequency ranges. In some embodiments, resistively loaded antennas such as resistively loaded dipole antennas, are used to reduce ringing and cross-coupling between elements. In certain configurations and embodiments, the system is configured for breast imaging and comprises impedance-matched coupling interfaces with relative permittivity approximately 10-20 to minimize skin reflections and operating parameters optimized for breast tissue dielectric properties including a frequency range of 1 -7 GHz for optimal penetration through dense tissue. In some embodiments the impedance matching is accomplished through liquid medium such as glycerin and saline water. In other embodiments, Delrin or plastics filled with higher impedance material fragments such as composite ceramics or polymers can be used to enable impedance matching. In other embodiments, polylactic acid may be used for impedance matching. In other configurations, the system is configured for lung imaging and comprises respiratory gating synchronization to mitigate motion artifacts from breathing and penetration depth optimization for thoracic imaging using a frequency range of 1-8 GHz. In still other configurations, the system is configured for brain imaging and comprises frequency selection optimized for skullpenetration using a frequency range of 1-6 GHz and algorithms for detecting ischemic changes associated with stroke or tumor growth.

[0011] In certain embodiments specifically configured for dense breast tissue screening, the system is optimized for detecting malignant tumors in breasts classified as BIRADS C (heterogeneously dense) or BI-RADS D (extremely dense) density. Such embodiments operate within a frequency range of 1-10 GHz optimized for penetration through fibroglandular tissue, and employ signal processing algorithms configured to achieve signal-to-clutter ratios exceeding 10 dB for tumors as small as 2mm diameter at depths up to 5cm within dense breast tissue. The coherent change detection processing in these embodiments is configured to detect dielectric property changes of 5-10% between sequential measurements, wherein such changes correspond to early-stage malignant tumor development including angiogenesis and cellular proliferation in dense breast tissue where conventional mammography sensitivity falls below 68%.

[0012] Some embodiments configured for breast imaging further comprise a breast positioning and coupling system that includes a breast cup fabricated from impedance-matched material having relative permittivity between 10-20 to minimize electromagnetic reflections at the skin-breast cup interface. The breast cup includes positioning features configured to provide repeatable breast positioning between temporally separated sequential imaging sessions, achieving positional repeatability within 10mm prior to algorithmic registration. The breast cup geometry is configured to maintain consistent tissue compression and breast shape across multiple imaging sessions to facilitate temporal image registration and coherent change detection. In such embodiments, the breast cup is configured to reduce skin reflection artifacts when combined with the calibration channel and skin subtraction algorithms, enabling detection of sub-resolution tissue changes in fibroglandular breast tissue.

[0013] Some embodiments further comprise a local acquisition unit containing the microwave electronics module and the plurality of antenna elements, a remote processing infrastructure configured to receive measurement data from one or more acquisition units via secure communication protocols, perform image reconstruction using distributed computing resources including GPU clusters, and maintain a temporal database of measurement data organized chronologically to enable longitudinal analysis, and a secure communication protocol ensuring regulatory-compliant data transmission between local and remote components using AES-256 encryption and HIPAA-compliant data handling procedures.Method Embodiments

[0014] In another aspect, the invention provides a method for coherent change detection microwave imaging of biological tissue. The method comprises acquiring first measurement data of a target tissue region at a first timepoint by illuminating the tissue with ultra-wideband microwave signals (in some embodiments spanning 0.3- 10 GHz) and measuring amplitude and phase of resulting signals across multiple frequency points. The method further comprises acquiring second measurement data of the target tissue region at a second timepoint separated from the first timepoint by days, weeks, months, or years or after a tissue change, using substantially the same acquisition process including the same antenna configuration, antenna positioning protocol, frequency range, and calibration procedures to maintain measurement consistency. The method additionally comprises processing the first and second measurement data to account for spatial differences caused by patient repositioning or tissue deformation. The method also comprises computing coherent differences between the processed measurement data while preserving both magnitude and phase information. The method further comprises identifying tissue regions exhibiting dielectric or conductivity property changes based on the coherent differences, wherein the identified changes indicate potential pathological conditions.

[0015] Alternative method embodiments include wherein processing the measurement data comprises registering the first and second measurement data using structural landmark detection to identify invariant anatomical features and elastic deformation algorithms to account for tissue deformation, achieving positioning accuracy of approximately 1 mm compared to 14.7 mm unregistered accuracy. Some embodiments further comprise classifying detected changes using frequencydependent response signatures that distinguish malignant from benign tissue based on characteristic dielectric property variations across the 0.3-10 GHz spectrum.

[0016] In certain embodiments, computing coherent differences comprises complex subtraction of registered signals maintaining phase information through S-parameterprocessing, frequency-dependent weighting based on tissue-specific absorption characteristics with attenuation rates of approximately 4 dB / cm at 4 GHz center frequency, or statistical thresholding to distinguish true changes from measurement noise while achieving signal-to-clutter ratios exceeding 10dB. Some method embodiments include wherein the identified tissue regions include sub-resolution tissue changes smaller than a spatial resolution defined by the acquisition process, wherein said sub-resolution tissue changes as small as 2 mm diameter at 5 cm depth are detected through phase-coherent processing of the coherent differences despite centimeter-level nominal spatial resolution.Calibration System Embodiments

[0017] In still another aspect, the invention provides a calibration system for maintaining phase coherence in microwave imaging. The system comprises a primary signal path including antenna elements for tissue interrogation. The system further comprises a parallel calibration path configured to experience and characterize system variations experienced by part or all of the primary signal path, wherein the parallel calibration path bypasses the antenna elements to isolate system-induced variations from tissue response. The system additionally comprises a compensation processor configured to measure signal characteristics for the calibration path at rates between 0.01 Hz and 1000 Hz, compute correction factors based on the measured characteristics to account for thermal drift and cable flexure, and apply correction factors to maintain phase coherence across sequential measurements separated by extended time periods, reducing total path-induced phase shift, ideally to less than 0.1 radians. The parallel calibration path is thermally coupled to the primary signal path by routing calibration cables or traces alongside or bundled with primary signal cables, ensuring that both paths experience substantially identical thermal environments. This thermal coupling enables the calibration path to accurately characterize temperature-induced phase variations affecting the primary measurement path.

[0018] Alternative calibration system embodiments include further comprising a switching network configured to route signals through the primary signal path and the parallel calibration path using automated switching at the scan rate or higher frequencies, wherein the parallel calibration path bypasses the antenna elementswhile maintaining equivalent cable lengths and thermal exposure. In some embodiments, the compensation processor is further configured to measure signal characteristics at rates between 0.01 Hz and 1000 Hz with typical rates of approximately 100 Hz matching antenna scan rates, store calibration history for drift trend analysis to enable predictive compensation, implement predictive compensation based on temperature gradients measured by integrated thermal sensors, and alert operators when calibration parameters exceed acceptable ranges indicating potential system malfunction.System Integration Embodiments

[0019] Various embodiments provide integrated systems incorporating multiple aspects of the invention. For example, embodiments include systems for evaluating breast tissue changes over time, comprising the coherent change detection microwave imaging system as described above, a database for storing breast images from multiple timepoints with patient identification and chronological organization, and an analysis module configured to compare breast images from different timepoints to identify temporal changes in breast tissue properties including size, vascularization, and dielectric characteristics. In certain embodiments, the analysis module is further configured to differentiate between benign and malignant lesions based on temporal changes in dielectric properties and frequency-dependent response signatures characteristic of malignant tissue.

[0020] The system architecture described herein enables multiple clinical applications including breast cancer screening with scan times of under 3 minutes, lung cancer monitoring with respiratory-gated acquisition, brain tumor detection using transcranial imaging, and therapeutic monitoring during treatment protocols including assessment of response to neoadjuvant chemotherapy. Various embodiments achieve performance metrics representing a breakthrough for breast cancer screening, targeting sensitivity and specificity exceeding 90% even in the highest density breast tissue (BI-RADS C / D classifications). By leveraging coherent processing of both magnitude and phase data between scans separated by days, weeks, months or years, the system achieves Signal-to-Clutter Ratios exceeding 10 dB and Signal-to-Noise Ratios exceeding 10 dB as required for high sensitivity and specificity performance. In some embodiments, the technology enables detection of2 mm malignant tumors at depths of up to 5 cm through sub-resolution detection capabilities enabled by phase-coherent temporal analysis.BRIEF DESCRIPTION OF THE DRAWINGS

[0021] FIG. 1 is a block diagram illustrating the overall system architecture, including the local acquisition unit with microwave electronics module, antenna array, calibration system, and control system, connected via secure communication protocols to the remote processing infrastructure with image registration, coherent change detection, and machine learning classification modules.

[0022] FIG. 2 illustrates the antenna array configuration in top-down view, showing multiple ultra-wideband antenna elements arranged in three mobile pairs at 0- degree, 120-degree, and 240-degree positions around a central breast cup with individual RF shielding and impedance-matched material, housed within an RF- absorbing enclosure with precision positioning mechanisms.

[0023] FIG. 3 shows the signal processing flow diagram including coherent change detection, regional landmark alignment, and tissue classification algorithms.

[0024] FIG. 4 depicts the calibration system architecture with primary and parallel signal paths.

[0025] FIG. 5 shows a three-dimensional coherent change detection microwave image simulated by a full wave microwave simulation tool on virtual phantoms derived from human MRI exams. The resulting change detection image highlights breast tissue with a localized region of dielectric property changes corresponding to malignant tumor tissue.

[0026] FIG. 6 shows a two-dimensional coherent change detection microwave image slice simulated by a full wave microwave simulation tool on virtual phantoms derived from human MRI exams. The resulting change detection image highlights breast tissue with a localized region of dielectric property changes corresponding to malignant tumor tissue.

[0027] FIG. 7A illustrates a top-down view of an exam table with breast imaging area showing breast cup positioning.

[0028] FIG. 7B illustrates a side cross-sectional view of the exam table and antenna array configuration without protective shell, showing antenna positioning relative to the breast cup.

[0029] FIG. 7C illustrates a side view of the exam table with a protective shell enclosure around the imaging hardware.

[0030] FIG. 7D illustrates a three-dimensional perspective view of the impedance- matched breast cup.DETAILED DESCRIPTIONSystem Architecture Overview

[0031] Referring to FIG. 1 , the coherent change detection microwave imaging system 100 comprises a local acquisition unit 110 and a remote processing infrastructure 150 connected via secure communication protocols 140. The local acquisition unit 110 contains the core hardware components for signal generation, transmission, reception, and initial processing, while the remote infrastructure 150 provides the substantial computational resources required for complex image reconstruction, machine learning classification, and temporal database management. This distributed architecture allows the local unit to be compact and cost-effective while leveraging powerful cloud-based computing for advanced analysis. In alternative embodiments, the remote processing infrastructure 150 may be replaced with local processing infrastructure when network connectivity is unavailable or when local processing is preferred for security or latency requirements.Local Acquisition Unit 110

[0032] The microwave electronics module 112, contained within the local acquisition unit 110, is responsible for signal generation and reception. For signal generation, the module comprises a Direct-Digital-Synthesis (DDS) module 114 configured to generate precise digital waveforms, which are then converted into analog signals for transmission by the antenna array.

[0033] For signal reception, the microwave electronics module 112 utilizes a Heterodyne Receiver architecture 1 16. This receiver is configured to process the attenuated microwave signals received by the antenna elements after interacting with the target tissue. The incoming high-frequency signal (fRF) is first amplified by a low-noise amplifier (LNA) to improve the signal-to-noise ratio. The amplified signal is then mixed with a stable, locally generated reference signal (fio) from a local oscillator, producing a lower, fixed intermediate frequency (fir) signal, where fiF=l fRF - to I.This down-conversion step allows the signal at the intermediate frequency to be amplified and filtered with precision, low component count, and high stability prior to being digitized by the analog-to-digital converters (ADCs) 118.

[0034] In an alternative embodiment, the Heterodyne Receiver is replaced with a direct sampling architecture, wherein high-speed ADCs digitize the amplified radio frequency signal directly, and subsequent processing, including filtering and downconversion, are performed in the digital domain.

[0035] A switch matrix 130 is operatively connected between the microwave electronics module 1 12 and the antenna array 200. The switch matrix 130 is configured to selectively route signals between individual antenna elements and the microwave electronics, enabling sequential or parallel antenna operation modes and providing flexibility in signal path configuration. In some embodiments, the switch matrix is replaced by additional independent transmit and / or receive channels.

[0036] A control system 125 coordinates the operation of all components within the local acquisition unit 110, including timing control, position management and tracking, sequence management, and system synchronization. The control system 125 interfaces with the FPGA / DSP / Processing module 115 to execute measurement protocols and manage data flow throughout the system. In some embodiments, motion tracking sensors are incorporated and used to enable timed signal transmission and reception to reduce motion blur.

[0037] The local acquisition unit 110 also includes an FPGA / DSP / Processing module 115 configured to provide high-speed digital signal processing capabilities 120 and real-time control functions. The FPGA / DSP / Processing module 1 15 coordinates with the microwave electronics module 112 to enable rapid data acquisition and preliminary signal conditioning.

[0038] A data recording and storage unit 145 provides local storage capabilities for measurement data, enabling offline operation and serving as backup storage in case of communication interruptions. The data recording and storage unit 145 can store raw measurement data, processed signals, and system configuration parameters.

[0039] A communications module 135 within the local acquisition unit 110 manages data transmission to the remote processing infrastructure 150 via the secure communication protocols 140. The communications module 135 handles dataformatting, compression, and transmission scheduling to optimize bandwidth utilization and ensure reliable data transfer.Safety and Regulatory Compliance

[0040] The system implements multiple safety features ensuring patient safety during microwave tissue illumination. The transmitted power is limited to levels well below established exposure guidelines. In one embodiment, average transmitted power does not exceed 10 mW, with peak power below 100mW. In some embodiments, Specific Absorption Rate (SAR) calculations confirm tissue heating remains below 1°C even for continuous 10-minute exposures, well within safety standards.

[0041] In some embodiments, examination duration is limited to reduce any potential exposure, with typical breast scans completing within 1-5 minutes per breast. In some embodiments, emergency stop mechanisms immediately cease all radio frequency transmission.

[0042] In some embodiments, patient data transmission via secure communication protocols 140 implements AES-256 encryption compliant with HIPAA requirements. In some embodiments, data storage maintains audit trails and access controls meeting 21 CFR Part 11 electronic record requirements.Remote Processing Infrastructure

[0043] The remote processing infrastructure 150 includes a communications module 1050 configured to receive data from the local acquisition unit 1 10 and manage bidirectional communication for system control and data transfer.

[0044] A data reception and storage module 1010 is configured to securely receive, validate, and store measurement data transmitted from local acquisition units. The data reception and storage module 1010 maintains a temporal database that organizes patient measurement data by patient and chronologically to enable longitudinal analysis and comparison across multiple imaging sessions. The module includes data integrity verification, redundant storage systems, and automated backup protocols to ensure reliable data preservation for coherent change detection processing.

[0045] An image registration module 1060 is configured to align measurement data from different timepoints by identifying structural landmarks and applying geometric transformations to compensate for patient positioning variations. The image registration module 1060 employs algorithms including elastic deformation techniques and feature-based alignment to achieve spatial correspondence between sequential measurements.

[0046] A coherent change detection module 1070 processes the registered measurement data to identify and quantify changes in tissue properties between timepoints. The coherent change detection module 1070 computes complex-valued differences while preserving phase information and applies statistical analysis to distinguish true biological changes from system noise.

[0047] The machine learning classification module 1080 analyzes the output from the coherent change detection module 1070 to classify detected changes as benign, malignant, or requiring further investigation. The machine learning classification module 1080 incorporates tissue-specific algorithms trained on frequency-dependent response signatures to improve diagnostic accuracy.

[0048] A security protocols module 1040 is configured to implement cloud storage, processing, and communications data protection and regulatory compliance measures throughout the remote processing infrastructure. The security protocols module 1040 manages end-to-end encryption for data transmission and storage, implements access control and user authentication systems, and maintains audit trails for all data processing activities. The module ensures compliance with healthcare data protection regulations including HIPAA requirements and provides secure multi-tenant data isolation when serving multiple healthcare facilities.Antenna Array and Positioning

[0049] Referring to FIG. 2, the antenna array configuration 200 is shown in a top-down view optimized for breast imaging applications. Referring to FIG. 7A, the measurement system and exam table is shown from a top down with a breast cup 750 in the middle. Referring to FIG. 7B, a side view of the exam table 740 and the antenna array configuration 200, is shown without the protective shell 730. Referring to FIG. 7C, a side view of the exam table 740 and the protective shell 730 is shown. Referring to FIG. 7D, a 3D side view of the impedance matched breast cup 750 isshown. The system comprises multiple ultra-wideband (UWB) antenna elements 210 arranged in pairs around a central breast cup 750. Each antenna element 210 is positioned in close proximity to an impedance-matched interface of the breast cup to ensure high signal coupling efficiency into the tissue.

[0050] The antenna elements 210, in some embodiments shown in FIG. 2, are arranged in three mobile pairs positioned around the breast cup 750. The first and second antenna element 210 positioned at a tangential offset from the first to provide spatial diversity and improved signal-to-clutter ratio. In some embodiments, the antennas are arranged with element spacing of less than or equal to A / 10 at a center frequency of the ultra-wideband microwave signals. The tangential spacing between antenna elements 210 in each pair ensures that their individual shielding enclosures 250 do not overlap, reducing electromagnetic interference between adjacent elements. In other embodiments, a plurality of antennas is mounted at one or more of the mount points to provide small angular difference measurements from a fixed relative position from the transmitting antenna. The antennas may be arranged in 2D or 3D configurations so as to obtain angular and / or polarization diversity at a small relative angle to the transmitting antenna(s).

[0051] Each antenna element 210 is individually surrounded by RF shielding and breastimpedance matched material 250 to minimize mutual coupling between adjacent elements and to maintain electromagnetic isolation between the paired antenna elements. Shielding does not surround the direction of the breast where signals are transmitted and received.

[0052] The entire antenna array 200 is housed within an RF-absorbing and shielding enclosure 220 that provides electromagnetic isolation from external interference sources. In some embodiments, the enclosure 220 includes inner surfaces covered with RF attenuating material to suppress cavity modes and maintain consistent antenna performance across the operating frequency range. The enclosure 220 and individual rf shielding and impedance matched material 210 are uniform to reduce any and all variability between elements and between positions within the cavity.

[0053] To ensure measurement consistency and phase stability across all antenna positions, the antenna array 200 implements strict uniformity requirements for both antenna elements and their surrounding electromagnetic environment. Each antenna element 210 is fabricated using uniform materials, dimensions, andconstruction techniques to maintain consistent electrical performance. In some embodiments, antenna elements that are specifically used for transmission may have a differing or nonuniform design relative to receive-only antenna elements and associated cavities, such as to improve signal transmission gain relative to resistively loaded receive antennas.

[0054] In some embodiments, the individual RF shielding enclosures 250 surrounding each antenna element are manufactured to uniform specifications with controlled dimensional tolerances of ±0.5 mm or better. In some embodiments, the shielding material composition, thickness, and electromagnetic properties are uniform across all elements. In some embodiments, the impedance-matching material integrated with each shielding enclosure 250 has uniform dielectric constant and loss tangent specifications to ensure consistent coupling to tissue across all measurement positions.

[0055] The individual RF shielding enclosures 250 housing the antenna elements along with the RF-absorbing and shielding enclosure 220 provide a uniform electromagnetic environment by maintaining consistent absorber material properties, thickness, and placement throughout the interior cavity. This uniformity eliminates position-dependent cavity resonances and reflections that could introduce measurement variability. In some embodiments, the absorber material is selected to provide consistent attenuation characteristics across the 0.3-10 GHz operating bandwidth, with absorption performance varying by less than ±5 dB across different locations within the enclosure.

[0056] Element and cavity uniformity is critical for coherent change detection because any systematic differences between antenna positions or between measurement sessions would introduce artifacts in the temporal subtraction process. By minimizing signal variability between elements and positions within the cavity to levels substantially below the expected tissue change signals, the system achieves the phase stability and measurement consistency required for high-sensitivity detection of subtle dielectric property changes.

[0057] Precision positioning mechanisms 230 enable controlled relative rotational movement of the antenna elements around the breast cup 750. The positioning mechanism 230 allows for data acquisition from many angular positions around the breast and breast cup 750, substantially improving spatial sampling resolution andresulting signal quality. In some embodiments, the antennas are adjusted in the axial direction as well so as to enable minimizing reflection from the breast and accounting for differences in shape of the breast inside of the breast cup 750. In some embodiments, shaped, padded, or molded features surrounding the exam table aperture 770 on top of the exam table are used to repeatably align the breast in the breast cup to reduce positioning errors. In some embodiments, these features are made of medical-grade silicone, medical-grade polyurethane foam, medicalgrade thermoplastic elastomers, medical-grade polyetheretherketone, or medicalgrade polycarbonate with a preference towards materials that are substantially transparent to electromagnetic radiation in the 1 -10GHz range. In some embodiments, the antennas are also adjusted in the vertical direction relative to the breast cup, enabling measurement in multiple planar or cross-planar orientations. The precise position information for each antenna at each time point is recorded to enable image reconstruction.

[0058] The breast cup 750 provides consistent patient positioning and uniform impedance matching to minimize skin reflections during the imaging process. The active radial positioning of all antenna elements 210, combined with their tight tangential spacing and radial orientation pointing toward the breast, ensures consistent electromagnetic field distribution and penetration depth throughout the breast volume while maintaining optimal phase coherence for the coherent change detection algorithms.

[0059] This antenna configuration enables the system to maintain phase coherence and signal-to-clutter ratios necessary for coherent change detection between sequential imaging sessions, while the individual shielding 250 around each antenna element 210 and the tangential spacing between paired elements substantially reduces cross-coupling that could degrade measurement accuracy. In some embodiments, the controlled tangential separation ensures consistent signal propagation characteristics and electromagnetic isolation across all antenna pairs.Calibration System

[0060] Referring to FIG. 4, the calibration system 400 is essential for maintaining phase coherence across sequential measurements. The design includes parallel channels, in some embodiments these are cables or traces, that run along the same pathwayas the main signal path 410, but only simulate the inclusion of antennas. In some embodiments, the simulation of the antenna in the calibration channels is through attenuators of similar amplitude as the loss from the antennas. It actively compensates for system variations caused by thermal drift, cable flexure, and component aging using a parallel calibration path 420 and a switching network 430. This system is designed to correct for major sources of phase error, such as cable flexure and temperature swings that can induce substantial uncompensated phase errors. One embodiment of the compensation processor 440 measures these drifts at a rate between 0.01 Hz-1000 Hz and applies correction factors, reducing the total path-induced error to substantially less than A / 4, thereby maintaining the high phase coherence required for effective CCD. In some embodiments, thermal sensors 122 are used to estimate changes in the path and cable lengths for phase compensation purposes.Coherent Change Detection Imaging Results

[0061] Referring to FIG. 5, a three-dimensional coherent change detection microwave image demonstrates the system's capability to isolate tissue changes. The image shows breast tissue 510 with a localized region 550 exhibiting dielectric property changes corresponding to malignant breast tumor tissue and associated vascularization. This image was generated using MRI-derived virtual phantoms converted to electromagnetic tissue models, followed by full-wave finite-difference time-domain (FDTD) simulation of the microwave imaging system response. The simulated data from sequential virtual imaging sessions were registered and subjected to coherent change detection processing to isolate regions of temporal change while suppressing static background tissue.

[0062] Referring to FIG. 6, a two-dimensional slice through a coherent change detection microwave image shows a cross-sectional view of breast tissue 610. The highlighted region 650 indicates an area of detected dielectric property change identified through the coherent change detection process. This image similarly represents results from MRI-derived virtual phantom simulations processed through full-wave FDTD electromagnetic modeling, followed by image registration and coherent subtraction of sequential measurements. The detection of region 650 demonstratesthe system's ability to identify localized tissue changes that may indicate malignant growth, even when such changes represent subtle variations in dielectric properties.

[0063] The imaging results shown in FIGS. 5 and 6 validate the technical approach described herein, demonstrating that coherent change detection processing effectively isolates regions of tissue change from background clutter and static anatomical features. The simulation methodology employed - converting human exam MRI data to dielectric property maps or virtual phantoms, performing full-wave electromagnetic simulation, registering and applying coherent change detection algorithms - provides proof of the system's malignant tissue detection capabilities.Physical Implementation Details

[0064] Referring to FIGS. 7A-7D, the physical configuration of the imaging system for breast cancer screening applications is illustrated. FIG. 7A shows a top-down view of an examination table 740 configured for prone-position patient imaging. The examination table 740 includes an aperture through which the breast cup 750 is positioned, allowing the patient's breast to naturally pendant into the imaging volume aligned by the breast cup 750 during examination while the patient lies face-down on the padded surface of the table.

[0065] FIG. 7B illustrates a side cross-sectional view revealing the internal configuration of the antenna array 200 positioned beneath the examination table 740. The antenna array 200 surrounds and underlies the breast cup 750, with individual antenna elements 210 and their associated shielding enclosures 250 visible. Linear positioning stages 760 provide structural support and enable precise vertical and horizontal positioning of the antenna array components. The RF-absorbing enclosure 220 contains the antenna array assembly and provides electromagnetic isolation from external interference. In this view, the protective shell 730 has been removed to show internal components. In some embodiments as shown, independent rotational positioning mechanisms 230 for each antenna element set are implemented. Individual antenna elements 210 with shielding 250, and RF- absorbing enclosure 220 are shown. Additional linear stages 760 provide structural support and vertical or horizontal positioning for the antenna array 200.

[0066] FIG. 7C depicts the examination table 740 with the protective shell 730 installed, providing physical protection for the imaging system components, processing, and,in some embodiments, providing additional electromagnetic shielding. The protective shell 730 encloses the antenna array 200 and associated positioning mechanisms while maintaining patient safety and system durability during clinical operation.

[0067] FIG. 7D shows a three-dimensional perspective view of the breast cup 750, illustrating its impedance-matched construction. The breast cup 750 is fabricated from materials having dielectric properties selected to minimize electromagnetic reflections at the skin-cup interface, in some embodiments having relative permittivity in the range of 10-20 to provide impedance matching between air and biological tissue. The breast cup 750 is available in multiple sizes to accommodate different breast volumes and ensure consistent positioning across diverse patient populations.Multi-Stage Processing Architecture

[0068] The system operates by executing a multi-stage process to detect and classify temporal changes in tissue. The process includes the following stages:1 . Data Acquisition and Calibration

[0069] At a first timepoint, the control system 125 directs the microwave electronics module 1 12 to sequentially transmit and receive signals through the antenna array 200. During acquisition of the microwave signals 310 and associated measurement data 314, the calibration system 400 concurrently measures and corrects for system drifts and path or cable motion to maintain phase stability by division or subtraction of the parallel channel measurements in the calibration and filtering algorithms 320. In some embodiments, prior calibration measurement data 312 from the system with and without calibration test objects may be subtracted from or divided into the microwave signals 310. In some embodiments, filtering may be applied 320 to reduce measurement noise. In some embodiments, this process may be repeated at subsequent positions around the breast and at different coronal planes. In some embodiments, motion blur compensation algorithms 324 may be applied to compensate for phase change due to measured signal path motion.

[0070] At a second time point 316, a second set of images is acquired using the same process. The temporal separation between first and second timepoints is typically selected based on the expected rate of biological tissue changes. For cancer screening applications, typical separation intervals range from 3 months to 2 years,corresponding to standard screening intervals. For therapeutic monitoring applications, intervals may be shorter (weeks to months) to assess treatment response. The system maintains calibration data and patient positioning references to enable coherent comparison across these extended timeframes.2. Image Reconstruction

[0071] The image reconstruction module 1020 is built with multi-frequency reconstruction algorithms 322 that use the reflected and backscattered microwave phase and magnitude data to generate a 3D image of the breast or tissue that is being imaged. In some implementations, the reconstruction algorithm 322 is a hybrid radar-tomographic approach where the image is converged upon based on an initial coarse estimate of the image and increasingly finer and more accurate iterations using a cost-based improvement estimation process. In some embodiments, the initial coarse estimate of the image may be derived from prior images of the same modality or different modalities such as MRI, CT, or mammogram. In some embodiments, this step may be skipped and the acquired data is used directly in the registration or machine learning classification steps.3. Spatial Registration

[0072] The image registration module 1060 employs a multi-stage registration process adapted for deformable biological tissue. In some embodiments, the landmark and segmentation algorithms 340 employ a tiered approach. In a first coarse registration stage, anatomical landmarks including chest wall interface, nipple location, major tissue interfaces, and other typically stable and repeatable radio frequency features are identified using edge detection and segmentation algorithms. In some embodiments, these landmarks provide initial spatial correspondence between measurement sessions with typical accuracy of 5-10mm. In a second fine registration stage, region and subregion alignment 344, elastic deformation algorithms model tissue compression and repositioning. In one embodiment, finite element modeling of soft tissue biomechanics estimates deformation fields. In another embodiment, B-spline based non-rigid registration aligns tissue regions using mutual information or cross-correlation metrics. In some embodiments,regional and subregional registration achieves positioning accuracy of 1-2mm, compared to 10-15mm accuracy without registration. The prior landmarks and segmentation data 342 from the prior screening data 316 is used to process the prior images in a consistent manner.

[0073] The registration process accounts for natural tissue deformation caused by gravitational loading, breathing, positioning changes, tissue compression, and cardiac motion. In some embodiments, registration parameters are initially selected through training on hundreds or more of labeled microwave imaging exams or analysis by comparing aligned structural features between sequential measurements.4. Coherent Difference Computation

[0074] Following registration, the coherent change detection module 1070 computes the difference between the two complex-valued datasets (S1 and S2). In some embodiments, the coherent change detection algorithm 350 is a complex subtraction (AS=S2-S1), which preserves both magnitude and phase differences. This coherent process is sensitive enough to detect permittivity and conductivity changes caused by growth and vascularization of malignant tissue, that may be smaller than the nominal spatial resolution of the system. Statistical thresholding may then be applied to the resultant difference data (AS) to distinguish significant changes from system noise. The resulting outputs are combined with a CCD image assembly 352 to create the microwave CCD image 372.5. Machine Learning Classification

[0075] Finally, the machine learning classification module 1080 analyzes the coherent change detection image and data to isolate regions that have significantly changed. The reconstructed microwave image 370 or raw data is then used, in addition to the coherent change detection data, to determine if there is malignant tissue in the image. Feature extraction algorithms identify distinguishing characteristics, such as frequency-dependent response signatures of malignant tissue that have been shown in literature. These features and data are input to a classifier 360, such as a Support Vector Machine (SVM), which has been trained on a curated database to classify thechanges as benign, malignant, or requiring follow-up as an output 374. In some embodiments, alternatives to machine learning classifiers, such as matched filters or heuristic based approaches are implemented to achieve similar results.Clinical Application Configurations

[0076] An embodiment of the system is adaptable to multiple clinical applications through configurable operating parameters. For breast imaging applications, the system is configured for prone-position imaging using a frequency range of 0.3-10 GHz with impedance-matched breast cups to minimize skin reflections. For thoracic imaging applications, one embodiment of the system uses a 1-8 GHz frequency range and incorporates respiratory gating synchronization to mitigate motion artifacts from breathing. For brain imaging applications, one embodiment of the system is optimized for transcranial imaging using a lower frequency range of 1-6 GHz for improved skull penetration, with algorithms specifically tuned to detect dielectric changes associated with ischemic stroke or tumor growth.

[0077] In some embodiments, the system is designed towards a mobile application or a portable application with motion stages and antenna arrays mounted with robust structures, self-calibration techniques to remove shock and vibration derived misalignment, lightweight PCB based antenna and associated light weight shielding system, and reduced form factor. In some embodiments, a mobile exam table is used to enable operations in the back of a mobile cancer screening van or vehicle. In other embodiments, the medical imaging system is portable and does not include an exam table mount. In some embodiments, the mechanical alignment is accomplished using small, shaped and molded structures mounted on top of the measurement system.

[0078] This detailed description provides an overview of the system, its components, and its methods of operation, demonstrating how the invention overcomes the limitations of prior art to provide a highly sensitive, non-ionizing imaging modality for a variety of critical medical applications.

Claims

CLAIMS:Claim 1 : A medical coherent change detection microwave imaging system for detecting temporal changes in biological tissue, comprising: at least one microwave electronics module configured to generate and receive ultra-wideband microwave signals; a plurality of antenna elements operatively connected to said microwave electronics module; at least one calibration channel configured to measure and compensate for system signal path length variations due to path thermal changes or cable motion; an image registration module configured to spatially align measurement data acquired at different timepoints by identifying structural landmarks in said biological tissue and applying elastic deformation algorithms or regional matching algorithms to compensate for tissue repositioning and deformation; and at least one control system configured to illuminate a target tissue region with said microwave signals at multiple timepoints, measure amplitude and phase characteristics of signals from said target tissue region at each timepoint, maintain phase coherence between measurements taken at different timepoints through said calibration channel, compensate for tissue deformation between timepoints using said image registration module, and perform phase-coherent processing of complex-valued measurements to detect dielectric property changes in said tissue caused by biological tissue changes including vascularization and growth.Claim 2: A method for coherent change detection microwave imaging of biological tissue, comprising: acquiring first measurement data of a target tissue region at a first timepoint by illuminating said tissue with ultra-wideband microwave signals spanning 0.3-10 GHz and measuring amplitude and phase of resulting signals including calibration channel data to maintain measurement consistency;acquiring second measurement data of said target tissue region at a second timepoint, using the same antenna configuration and frequency range as well as calibration channels to maintain measurement consistency; processing said first and second measurement data to create phase coherent reconstructed images of the biological tissue by applying hybrid or coherent microwave image reconstruction algorithms; processing said first and second measurement data to account for patient repositioning and tissue deformation by applying spatial registration algorithms; computing coherent differences between said processed measurement data while preserving both magnitude and phase information; and identifying tissue regions exhibiting dielectric or conductivity property changes based on said coherent differences, wherein said changes indicate potential pathological conditions.Claim 3: A calibration system for maintaining phase coherence in microwave imaging, comprising: at least one primary signal path including antenna elements; at least one parallel calibration path configured to experience and characterize system variations experienced by part or all of said primary signal path, wherein said parallel calibration path comprises a parallel signal path having equivalent cable lengths to said primary signal path and traverses substantially the same physical pathway as said primary signal path, and wherein said parallel calibration path bypasses said antenna elements while maintaining thermal coupling to said primary signal path; a switching network configured to selectively route signals through said primary signal path and said parallel calibration path; and at least one compensation processor configured to measure signal characteristics for said calibration path at rates between 0.01 Hz and 1000 Hz, compute correction factors based on the measured characteristics, and apply said correction factors to maintain phase coherence across sequential measurements separated byextended time periods.Claim 4: Claim 4: The system of claim 1 , further comprising an antenna array with adjustable positioning mechanisms, wherein said control system is configured to adjust one or more of said frequency range, emitted power amplitude, antenna element spacing, antenna element location, antenna element orientation, and signal processing algorithms based on dielectric properties of said target tissue type, thereby enabling configuration for multiple anatomical applications including breast imaging, lung imaging, and brain imaging applications.Claim 5: The system of claim 1 , further comprising: a local acquisition unit containing the microwave electronics module and the plurality of antenna elements; a remote processing infrastructure configured to receive measurement data from one or more acquisition units, perform image reconstruction using computing resources, and maintain a temporal database of measurement data; and a secure communication protocol ensuring regulatory-compliant data transmission between local and remote components.Claim 6: The system of claim 1 , wherein said plurality of antenna elements comprises: ultra-wideband antenna elements arranged with element spacing of less than or equal to one-tenth wavelength at a center frequency of said ultra-wideband microwave signals; uniform elements and associated surrounding cavity to minimize differences between measurements between individual elements; individual radio frequency shielding enclosures surrounding each antenna element to reduce mutual coupling between adjacent elements below -20 dB; and resistively loaded antenna structures configured to reduce ringing and crosscoupling between elements.Claim 7: The system of claim 1 , further comprising a positioning mechanism configured to move said antenna elements in at least one of rotational, axial, radial, and lineardirections relative to said target tissue region, wherein said antenna elements are individually shielded with radio frequency shielding material to reduce mutual coupling across the operating bandwidth, and wherein said antenna elements are housed within an enclosure having inner surfaces covered with radio frequency absorbing material to suppress cavity modes and reduce electromagnetic interference.Claim 8: The system of claim 1 , further comprising an image reconstruction module configured to generate two- or three-dimensional tissue images using hybrid radar- tomographic reconstruction algorithms applied to said complex-valued measurements, wherein said reconstruction algorithms utilize both amplitude and phase information across multiple frequency points to resolve dielectric property distributions within said biological tissue.Claim 9: The system of claim 1 , wherein the microwave electronics module comprises: a digital synthesis module configured to generate said ultra-wideband microwave signals spanning a frequency range of 0.3-10 GHz; a heterodyne receiver architecture configured to down-convert received signals to an intermediate frequency for processing; one or more analog-to-digital converters with at least 14-bit resolution; and a field-programmable gate array or processor configured to perform real-time signal processing and system control.Claim 10: The system of claim 1 , further comprising one or more temperature sensors integrated with the microwave electronics module, wherein the control system is configured to use data from the temperature sensors to perform predictive thermal compensation.Claim 11 : The system of claim 1 , further comprising a machine learning module configured to classify tissue changes based on training data, wherein the machine learning module is further configured to enable performance improvement without hardware modification by receiving algorithm updates through secure remote communications.Claim 12: The system of claim 1 , wherein said microwave electronics module comprises: a low-noise amplifier configured to amplify received microwave signals; high-speed analog-to-digital converters configured to directly digitize said amplified microwave signals at sampling rates exceeding 1 giga-sample per second; and digital signal processing circuitry configured to perform filtering and downconversion of said digitized signals in the digital domain.Claim 13: The method of claim 2, wherein processing the measurement data comprises registering the first and second measurement data using structural landmark detection and / or elastic deformation algorithms.Claim 14: The method of claim 2, further comprising classifying detected changes using frequency-dependent response signatures that characterize differences in dielectric properties between malignant and benign tissue across said 0.3-10 GHz frequency range, wherein malignant tissue exhibits higher dielectric property variations compared to surrounding tissue.Claim 15: The method of claim 2, wherein computing coherent differences comprises: performing complex subtraction of registered signals while maintaining both magnitude and phase information; applying frequency-dependent weighting based on tissue-specific absorption characteristics; and applying statistical thresholding to distinguish true biological changes from measurement noise and system artifacts.Claim 16: The method of claim 2, wherein the identified tissue regions include subresolution tissue changes smaller than a spatial resolution defined by the acquisition process, wherein said sub-resolution tissue changes are detected through phase- coherent processing of the coherent differences.Claim 17: The system of claim 3, wherein the compensation processor further is configured to use one or more of:measure signal characteristics at a rate sufficient to track thermally-induced and motion-induced system variations between 0.01 Hz and 1000 Hz; store calibration history for drift trend analysis; implement predictive compensation based on temperature measurements; and alert operators when calibration parameters exceed acceptable ranges.Claim 18: The system of claim 4 or claim 1 , configured for breast imaging, comprising impedance-matched coupling interfaces and operating parameters optimized for breast tissue dielectric properties.Claim 19: A coherent change detection microwave imaging system specifically configured for detecting malignant tumors in dense breast tissue classified as BI-RADS C or BI-RADS D density, comprising: the system of claim 1 configured with operating frequency range of 1-10 GHz optimized for penetration through fibroglandular tissue; signal processing algorithms configured to achieve signal-to-clutter ratio exceeding 10 dB for 2mm malignant tumors at depths up to 5cm within dense breast tissue; and coherent change detection processing configured to detect dielectric property changes of 5-10% between sequential measurements, wherein said changes correspond to early-stage malignant tumor development including angiogenesis and cellular proliferation in dense breast tissue.Claim 20: The system of claim 1 , configured for breast imaging and further comprising a breast positioning and coupling system, said system comprising: a breast cup fabricated from impedance-matched material having relative permittivity between 10-20 to minimize electromagnetic reflections at the skin-breast cup interface; positioning features configured to provide repeatable breast positioning between temporally separated sequential imaging sessions, achieving positional repeatability within 10mm prior to algorithmic registration;a geometry configured to maintain consistent tissue compression and breast shape across multiple imaging sessions to facilitate temporal image registration and coherent change detection; wherein said breast cup is configured to reduce skin reflection artifacts when combined with calibration and skin subtraction algorithms, enabling detection of subresolution tissue changes in fibroglandular breast tissue.

Images